Forensic Geotechnical Engineering n5x3v

[email protected]

Hydraulic fracture in the core was considered as most likely the key mechanism estimated by desk work with laboratory aids. Independent to Review Cause of Teton Dam Failure (1976) reached a conclusion as follows, The had quickly identified piping as the most probable cause of the failure, and then focused its efforts on determining how the piping started. Two mechanisms were possible. The first was the flow of water under highly erodible and unprotected fill, through ts in unsealed rock beneath the grout cap, and development of an erosion tunnel. The second was “cracking caused by differential strains or hydraulic fracturing of the core material.” The was unable to determine whether one or the other mechanism occurred, or a combination. If pore pressure sensors had been installed in the dam section, process of increasing pore pressures during filling water might have provided vital information of the failure process.

The referred on monitoring as “There were not enough instruments in the dam to provide adequate information about changing conditions of the embankment and abutments.” The debate on the cause of the failure still continues until recently (Solava and Delatte, 2003) and at present. If pore pressure monitored data had been provided, we could have reached the valuable understanding. 2.2

Collapse of Nicoll Highway

）

（

The Committee of The Collapse of Nicoll Highway Inquiry(2005) that took place on 21 April 2004 was caused by an accident of subway construction in Singapore. The Failed section in the subway construction near Nicoll Highway was a cut and cover excavation shown in Fig.1.

Fig.1 Geotechnical condition for design of the failed section

Fig.2 Deflection of retaining wall of the failed section

(Number with underline is displacement of JGP slab) The excavation is planned in a very deep soft marine clay layer

and some special design of JGP (Jet Grouting Pile) was considered. The slab near the bottom of the excavation was to take strong horizontal load from the soft clay layer. Among various instrumentations, the most important information was the inclination of the temporal retaining wall. 9/1

10/1

11/1

12/1

1/1

2/1

3/1

4/1

2nd

GL=102.9m

5/1

100 1st strut

H eight(m )

3rd

4th 5th 6th 7th8th Strut and Excavation Stages

90 80

lower JGP

60 W all Deflection(mm)

500

Maximum Deflection of Retaining Walls

400

South.I104

300

North.I65 Design Level

200

a

The

9th

upper JGP

70

The original design condition was assumed as the same soil layers at both sides. Boring study after the failure have revealed that the dense sand gravel layer below the soft marine clay is inclined tow rd south and the soft marine layer is deposited deeper than the original design as shown in Fig. 4. If the site engineer could evaluate the different response of the walls of south and north, he might proceed to perform boring study and could prevent the failure. The site engineer just agreed with the proposal of modifying design parameter by the contractor and allowed to increase the allowable design level without any geotechnical consideration as shown in Fig.3. JGP slab is rather stiff elements compared to the soft soil at the site. The horizontal compression displacement allowable in the slab might be controlled by failure strain level under compression of the order less than possibly a few percents. Since the horizontal length of the JGP slab is about 20m, 20 cm of the compression at 8th strut stage (Fig.2) corresponds to 1% of compression strain, which was the design allowalble strain. The instrumentation and monitored data really had provided not only before the failure but also after the failure to give constrains of the possible mechanism discussed in forensic engineering process.

100 0 9/1

10/1

2003

11/1

12/1

1/1

2/1

Time

2004

3/1

4/1

5/1

3 Fig.3 Maximum deflection of walls of the south and north side Though under emergency, the measurement was not made daily. However, the obtained data gives clear tendency of the deformation as shown in fig.2. Two things should be noted. One is the deflection of retaining wall. The other is the horizontal compression of JGP slab. It should be noticed not only the increament of the maximum deflection itself but also the difference between south and north walls. As shown in Fig.2 and 3, the maximum deflections at sixth and seventh excavation level show the same order of 160mm and 200mm for north and south sides. However, after the seventh excavation, the maximum deflection of the south side wall increased with the excavation depth. At the same time, the deflection of the north side was kept constant about 20cm and even decreasing. However, this difference of deflection needs some reasons of either weaker soil condition or smaller stiffness of retaining wall in south side.

Earthpresure by soft clay southside

Earthpresure by soft clay northside

Fig.4 Deflection of walls at the final step and the confirmed geotechnical condition by boring after the failure

INSTRUMENTATION FOR MONITORING

Recent technologies of monitoring displacements or strains have much advanced within the past decade. Among these several technologies, only two methods of GPS (Global Positioning System) and BOTDR (Brillouin Optical-fiber Time Domain Refletometer) are introduced here. However, conventional methods are still important and used as daily practice. Due to the limitation of the number of paper, the author will review only a trend of instrumentation for monitoring in geotechnical engineering in some aspect. Reader should refer an introductory book as well as containing professional level by John Dunnicliff (1993). 3.1

Mechanical Sensor

Photo.2 shows a simple and yet accurate monitoring device to measure gap movement for foundation of Hudson-Athens Lighthouse (U.S. National of Historic Places) in the Hudson River in the state of New York in the United States.

Photo.2 Mechanical type of gap sensor (http://en.wikipedia.org/wiki/Strain_gauge) The two halves of the device are rigidly attached to the foundation wall on opposite sides of the crack. The red reference lines are on the transparent half and the grid is on the opaque white half. Both vertical and horizontal movement can

be monitored over time. In this picture, the crack can be seen to have widened by approximately 0.3mm (and no vertical movement) since the gauge was installed. 3.2

Strain Gage Sensor

Electric sensor is also common for automatic monitoring. Photo.3 is strain gage based displacement sensor. The range of measurement is 5 to 10 mm with accuracy of 0.5-1%/Full Scale.

becomes available in the market. 4

In the past decade, sensor technology had entered another era of innovative principles like MEMS (Micro Electro Mechanical Systems). In the chapter, among these cutting edge technology, GPS and fiber optic sensors are introduced. 4.1

Photo-3 Strain gage based displacement sensor (Tokyo Sokki) Strain gage for steel plate was also developed based upon vibrating wire sensor as shown Photo-4.

RECENT DEVELOPMENT OF SENSORS

Carrier-phase tracking of GPS

Carrier-phase tracking signal of GPS has resulted in a revolution in land surveying. A line of sight along the ground is no longer necessary for precise positioning. Positions can be measured up to about 30 km from reference point without intermediate points. This use of GPS requires specially equipped carrier tracking receivers. Displacements of multiple points may be monitored by Carrier Phase Tracking GPS method that provides relative displacements of target monitoring points to the reference point.

Photo.4 VW spot-welded strain Gauge (Durham Geo Slope indicator Co.) Vibrating wire transducer measures change of natural frequency of a wire installed in a very small pipe on a steel plate. It is designed to set the sensor by spot-welding to steel surface. The accuracy of the vibrating wire sensor is about 0.05% of full scale of 2500µstrain. 3.3

(Difference in travel distance=n19+δ (cm)) Fig.5 Basic Concept of the Carrier Phase Tracking modified from (http://www.shamen-net.com/word/img/)

Displacement, Load cell and Pore pressure sensor

Conventional strain gauge and vibrating wire are basic transducer for not only strain but also other application of displacement, load, and pore pressure sensors. Other types of transducers like LVDT (Linear Voltage Differential Transducer) are also used for various types of sensors. The accuracy of LVDT sensor is around 0.1-.0.5%/F.S. 3.4

Inclinometer

Inclinometer sensor system that had been developed in 1970s was based upon force balance sensor. The force balance principle that is to measure how much force needed to get back the original zero position of the displacement sensor have resulted in the wider range to measure more than 30 degrees with the accuracy of 0.01-0.02%/FS. Recently, Micro-Electro-Mechanical Systems (MEMS) have realized much cost effective with the same performance

Photo 5 GPS receiver (http://www.furuno.co.jp/product/gps/terrain/dana2000.html) As shown in Fig.5, distance between the monitoring station and

GPS is divided into two segments of equi-distance and the difference. The difference in travel distance is obtained as to detect the time difference between the receiving times of the same phase of the signal carrier signals. The L1 and/or L2 carrier signals are used in carrier phase surveying. The wave length used for the system is about 19cm of the L1 carrier band. The difference in travel distance = n19 + δcm) <where n is an integer number and called integer bias. With accuracy of 1% of wave length in detecting the leading edge, the error might be as low as 2 millimeters.

space conditions like humidity, temperature, and air pressure. To avoid such errors, it is better to set monitoring position with relative position within some limited range of several hundred meters. The most time consuming process is to estimate the wave number n. Satellite is moving over the receivers from L1 to L2 along the known orbit. Using four GPS satellite signals, the phase difference as δ is obtained for every satellite. As shown in Fig.6, point P is the crossing point between ray path from satellite to the monitoring receiver and the equi-distance circle (Circle 1 in Fig.6). As the satellite moves from L1 to L2, the crossing point P1 moves to P2. Since the radius of the circle is very long compared with base line, the arc from P1 and P2 to the reference receiver (RR) is considered as a straight line. The locus of the points P1 and P2 forms an elliptical orbit. The wave number n is obtained as to give the minimum error to show the orbit as elliptical shape in 3D space. Vertical displacement(mm) 40

forced displacement

30 20 10 0 -10 5/10

5/15

5/20

5/25

5/30

6/4

Fig.8 Response of GPS in vertical component Fig.8 shows response of the vertical component against forced displacement that is shown in red bold line. It is found that the response of vertical component is not sharp to small displacement but becomes very clear when the input displacement is larger than the 5mm of the level of standard deviation.

Fig.6 Determining wave number of integer bias for Carrierphase tracking of GPS method GPS Monitoring at G-3, Mikage

20

20

UD

Displacement(mm)

10

10

0

0

-10

-10

-20 2008/2/1-3/26 2/1 2/8 2/15 2/22 2/29 10 5

-20

3/7

NS

SD=5.02mm (%) 10.0

20.0

10

30.0

NSav=0.43mm SD=2.06mm

5

0

0

-5

-5

-10 2008/2/1-3/26 2/1 2/8 2/15 2/22 2/29 10 5

3/14 3/21 3/28 0.0

UDav=0.017mm

3/7

EW

-10 3/14 3/21 3/28 0.0 10

(%) 10.0

20.0

EWav=0.31mm

Optical Fiber Sensors

Fiber optics, though used extensively in the modern world, is a fairly simple and old technology. Guiding of light by refraction, the principle that makes fiber optics possible, was first demonstrated in Paris in the early 1840s. Prof. Jun-ichi Nishizawa, a Japanese scientist at Tohoku University, was the first to propose the use of optical fiber for communications in 1963. Nishizawa invented other technologies that contributed to the development of optical fiber communications as well.

SD=1.86mm

5

0

30.0

4.2

0

-5

-5

-10 2008/2/1-3/26 2/1 2/8 2/15 2/22 2/29

3/7

-10 3/14 3/21 3/28 0.0

(%) 10.0

20.0

30.0

Number in total 1252

Fig.7 Scattering of monitored position by GPS carrier phase tracking (Iwasaki, 2008) Table-1 Errors of displacement obtained by GPS

Component Standard Deviation (mm) UD 5.02 NS 2.06 EW 1.86 Errors are caused by several factors. One of the common factors is the variation of wave velocity near the surface among the monitoring points. Wave velocity may be affected by such

http:/ / en.wikipedia.org/ wiki/ Optical_fiber Photo.6 Optical fiber

The RI of 1.0 means light travel in the medium with the speed in vacuum air. The RI greater than 1.0 means light travel in the medium with lower speed than in vacuum air. To confine the optical signal in the core, the refractive index of the core must be greater than that of the cladding. If the wave length of the light is longer than compared to diameter of the core cable, the light is confined within the core and can travel very long distance with little scattering. The boundary between the core and cladding may either be abrupt change of refractive index, in step-index fiber, or gradual, in graded-index fiber. Single mode optical cable(SM) is designed as abrupt change to show little loss with capability of travelling long distance of 10km. Fig.9 Structure of optical fiber

4.3

Figure 10. Refractive Index (A and B: refraction, C: full reflection) Optical fiber has been developed for sending images of medical use and for optical signal as communication cable as shown in Photo. 9. Fig.9 shows a typical structure of the optical fiber that composes from core and clad. An optical fiber consists of fiber of quartz or plastics in a cylindrical shape that works as waveguide for transmitting light along its axis, by the process of total internal reflection. The fiber consists of a core surrounded by a cladding layer. The refractive index (or index of refraction, see Fig.10) of a medium is a measure of how much the speed of light (or other waves such as sound waves) is reduced in the medium compared to a reference medium. In treating light, the reference medium is usually vacuum air.

BOTDR (Brillouin Refletometer)

Optical-fiber

Time

Domain

When a sharp and strong optical pulse is sent into an optical fiber cable, some weak optical signal was recognized as returned to the entered gate. The signals are originated from the scattering of the light at the propagating front. Three different kind of scattering are identified as Rayleigh, Raman, and Brillouin. The scattered signals of Brillouin are found to show special characteristics. The frequency of the scattered pulse is shifted about 10GHz from that of the input pulse and changes proportional to strain and temperature in the fiber as shown in Fig.11

Fig.11 Brillouin shift vs. Strains and Temperature

Fig.12 Basic Concept on BOTDR

As shown in Fig.12, when a signal is sent to an optical fiber, backward travelling scattered light continuously arrives at the front gate. The point where the scattered signal originated is estimated by the delayed arrival time from the time of sending pulse. Since the relationship between the strain and Brillouin frequency shift is known, the analyzed frequency shift is easily transformed to strains or temperature. When temperature effect should be known independently, another optical fiber cable with free strain may be installed. The accuracy of the measurement depends upon accuracy to obtain the number of the frequency of Brillouin shift. At present, the measurement is usually given for 1m segment of the cable. The error range of strain is about 1020µstrain and that of temperature is 0.5-1 degree in Celsius.

An optical fiber was fixed along a plastic pipe. The fiber cable was given some pretension strain. The pipe strain gauge of optical fiber gives continuous change of the strain compared to conventional pipe strain measurement that gives only discrete points. As shown in Fig.13, the strain shows continuous change from compression to extension that corresponds to bending deformation at the depth of 10 to 11m. This depth was considered as a sliding surface. 4

Pile 126: BOTDR Strain Measurement between 12/08/05 and 8/11/05

∆T [°C]

2 0

BOTDR as Pipe Strain Gauge

Nakano et al. (2003) reported a case history of the application of the BOTDR to identify sliding surface by inserting a pipe equipped with optical fiber in a slope as shown in Fig.13.

E leva tio n (m )

-2

4.4

-4

Exc. Lvl.

-6 -8

-10 Outerside of Excavation Nearside of Excavation Axial strain

-12 -14

-16 -0.04 -0.03 -0.02 -0.01 (a)

0

0.01 0.02

Strain (%) (b)

((a) strains at pipe edges) ((b) strain in a loosed tube cable) Mohamad et al.( 2009) Fig.15 Measured strains along edges and in a loosed tube +

6 4

1st Prop

2 Fig.13 Monitoring slope instability by vertical pipe with optic cable

2nd Prop 3rd Prop (26 Oct)

Elevation (m)

0 -2 -4 -6

Exc. Lvl.

Base Slab Floor

-8

-10 -12 -14

BOTDR 8 Nov 05 Incl 3 Nov 05 Incl 21 Nov 05 Incl Averaged 12 Nov 05

-16BOTDR vs. conventional inclinometer -2 0 2 4 6 8 10

12

Displacement (mm)

Fig.14 Lateral deformation estimated from equivalent curvature with no axial strain (Mohamad (2007, 2009))

Fig.16 Comparison of Lateral Displacement (Mohamad (2007, 2009))

14

4.5

BOTDR as Inclinometerequiped with Secant Pile Wall

Mohamad et al.(2007, 2009) show some BOTDR application to monitor axial strain of bored piles and lateral deflection of retaining wall of secant pile wall of a series of intersecting reinforced concrete piles. The diameter of the secant wall was 450mm and a pair of optical fiber cable was installed along the reinforced steel was installed along steel cage giving a pretension of 2000µstrain. Another single optic fiber loose tubed cable was equipped along the monitored pile. Fig.14 shows how to evaluate lateral deformation from the edge strain along a pile. Since the edge strains are induced by mix modes of axial compression and bending deformation. To obtain bending strains, axial compression should be separated. These axial components should be subtracted from the apparent edge strains as shown in Fig.14. Temperature is another factor that affects upon the apparent strain values. Fig.15 shows an example of the measured strains of BOTDR for retaining wall after excavation. The excavated level was about -4.5m. Fig.15 (a) shows two edge strains of outside and nearside of the pipe as well as axial strain of the averaged values. Fig.15 (b) shows rather constant temperature in the soil up to a level of -6.5m and gradually decreases to -13 degree at the top of the pipe. The air temperature was lower than the in the ground, temperature decreases from near the ground surface to the top of the pile. Fig.16 shows comparison between a lateral deformation curve obtained from BOTDR and two curves by conventional inclinometer measurements. It is noticed that the averaged curve of the two deflection curves from the conventional inclinometer corresponds well to the one from BOTDR. Compare to the conventional inclinometer of manual handling, inclinometer system by BOTDR provides automatic and realtime evaluation of deformation of retaining wall. Inclinometer measures inclination angle relative to gravity. BOTDR measures a pair of the edge strains under bending mode. It should be notice that the distance between two edges must be wide enough to keep acceptable accuracy of bending deformation. At present, the facility to send and receive the signal as well as analyzing the BOTDR is rather expensive, In the near future, the cost effective system is expected available and become useful especially for big and deep excavation works. 5

CONCLUSION

In the first part, two comparative case studies are overviewed. Teton Dam that was failed by piping. No further detailed mechanism is clarified due to not enough instrumentation. In the collapse of Nicoll Highway, though the data were not collected as it should be the available data of inclination of the subway construction nearby the Highway show clear fact of one sided deformation of the wall from which the geotechnical condition for the design was anticipated different from the design condition. If the construction had been modified, the collapse might well be avoided. After the failure, the results of instrumentation contributed to identify the process of to the failure from the geotechnical point of view. In the second part, review of sensors is made including some edge cutting technologies of two sensors of Carrier-phase tracking of GPS method and BOTDR. These sensing technologies in addition with other sensor like MEM (Micro-machine Electronic Machine) sensor are still on a way of development. Forensic engineers, especially in the field of geotechnical field

where most of the phenomena takes place in dark underground, should use and take advantages of installation and monitored data to reach the most likely mechanism among several possibilities. 6

REFERENCES

The Committee of Inquiry, 2005,”Report on the Incident at the MRT Circleline Worksite that led to the Collapse of the Nicoll Highway on 20 April 2004,” Ministry of Manpower, Government of Singapore Geotechnical Instrumentation for Dunnicliff J. 1988. Monitoring Field Performance John Wiley & Sons, New York (USA)

“

,”

Iwasaki, Y. 2008, “Basic Study of Carrier-phase tracking of GPS method at Mikage site,” GRI report Independent to Review Cause of Teton Dam Failure, 1976, “Report to the U.S. Department of the Interior and State of Idaho on Failure of Teton Dam. Idaho Falls,” Idaho. December 1976 Nakano, M., Yamazaki, H., and Okuno, M., 2003, “The earthquake disaster prevention monitoring system using optical fiber technology,” JSCE Journal of Earthquake Engineering, Vol.27, pp.1-4 Mohamad, H., Bennett, P.J., Soga, K., Klar, A. & Pellow, A.2007.” Distributed Optical Fibre Strain Sensing in a Secant Piled Wall,” Proceedings of the Seventh International Symposium on Field Measurements in Geomechanics (FMGM2007), ASCE Geotechnical Special Publication, No. 175 Mohamad, H., Soga,K., & Bennett, P. J.,2009, “Fibre optic installation techniques for pile instrumentation,” Proc.,17th ICSMGE(in print) Solava and Delatte, 2003, “Lessons from the Failure of the Teton Dam,” Proceedings of the 3rd ASCE Forensics Congress, October 19 - 21, 2003, San Diego, California

Chapter 12 LEGAL PROCESS AND JURISPRUDENCE Dhirendra S. Saxena (Sax), P.E., F NAFE Senior Principal and Chief Consultant, ASC geosciences, inc., Lakeland, Florida, USA

9.1

INTRODUCTION

For geotechnical engineers it is only a myth to believe that practicing perfect engineering, or conforming to normal standards of care, will provide immunity from civil liability. Unfortunately, when problems or failures occur, all parties including engineers get named in the lawsuit regardless of their innocence. In USA, attorneys sue everyone involved with a damaged project to secure compensation or claim to their liability insurance limit. A practicing geotechnical engineer cannot provide services without the fear of a lawsuit. Strategies for limiting liability range from assessing risk to securing professional liability insurance and including limitation of liability classes in consulting contracts between engineer and the client. Although experts retained by opposing parties generally disagree on issues resulting from differences in professional judgment, they are invaluable to the jurisprudence system in America. A new discipline known as forensic geotechnical engineering (FGE) has been created to deal with investigation of soil-interaction related failures of engineered facilities or structures. Services of forensic geotechnical engineers, experienced in the jurisprudence system, are generally commissioned to investigate such failures. They also prepare reports, and sometimes provide expert witness testimony, in an attempt to assist their client in understanding the cause and liability for the failure. Because of the adversarial nature of most failure investigations many forensic engineers are generally under pressure to give an opinion that benefits their client’s position. Accordingly, the technical facts are not presented, yielding to this kind of pressure sometime leads to inappropriate occurrences, practicing outside one’s competence, manipulation of facts, crafted testimony, inadequate or defective investigations, misrepresentation of standards of practice, or other ethical violations. A forensic geotechnical engineer looks at a pile of rubble as a problem that needs to be solved. Each broken piece in the pile tells a story about the forces that acted upon it and somewhere in the twisted mass of intertwined building components is the first facture surface that resulted in the progressive collapse of the structure. At the end of the failure analysis process, which may include laboratory testing of critical items and hours of analysis in the office, a clear picture of what happened begins to develop, and the forensic engineer must then be able to explain the mode of failure and the forces, action, or process that caused the failure. The explanation may be a simple letter, a detailed report, or expert testimony in the courtroom (Bell, 2007). “What happened?”, and “Why did it happen?” are usually the first two questions asked of the forensic engineer. Of course, these are often followed by, “How can it be fixed?”, and all too often, “Who’s fault is it?”, and “Who is going to pay?” (Bell, 2007) When something breaks, falls down, or otherwise fails to perform as it was intended, our society demands to know why.

This inevitably becomes the job of the forensic engineer, who must be equally comfortable

in the field, the office, and the courtroom while providing an objective analysis of the failure, whether the facts are helpful or harmful to the client’s interests. FGE prepares civil engineers to read, think, speak, and analyze like a lawyer. In addition, it familiarizes him with jurisprudence system so that he is much able to understand and deal with legal issues since he has to work closely with statutes and regulations may become involved in litigation, or who may serve as an expert witness. 9.1.1

Who Wants to Know What

A forensic geotechnical engineer needs to answer questions, such as: •

The building owner who wants to know why the building fell down and who is going to pay for it.

•

The insurance adjuster who needs to determine if the cause of the collapse is covered under the building owner’s insurance policy, and if it was the result of a defect that someone else might be liable for.

•

The contractor and subcontractors, who built the building and their liability carriers, want to know if the defect is construction-related and if so, which trade or trades are responsible.

•

The design professionals and their liability carriers want t o know if the defect is design-related.

Each of these entities may retain their own forensic engineers to assist them with understanding how they may or may not be responsible. There will be attorneys for each entity who will evaluate and argue how the law should be applied to the insurance policies and contracts. The forensic engineer’s opinion of what happened may be explored during a deposition in which the attorneys for all sides have the opportunity to probe the engineer as to the process used and the basis for the opinion. If the parties cannot reach a settlement, the forensic engineer may be called to provide expert testimony in the courtroom, with the goal of helping the judge and/or jury understand the facts and how they relate to the particular incident. Opposing attorneys normally attempt to discredit the engineer through cross-examination with the hopes of revealing some flaw in the forensic engineer’s thought process. The forensic engineer must be able to apply the art and science of engineering during the entire process, from the extreme conditions encountered in the field, through the intense pressure experienced in the courtroom, and never lose sight that seeking the truth is all that matters. By understanding what went wrong, the forensic engineer is able to add to the body of knowledge that hopefully will prevent a similar occurrence from happening again. 9.2

TECHNICAL ISSUES

Within the framework of dispute relating to construction, particularly within the jurisprudence system, forensic engineering may include investigation of the physical causes of accidents and other sources of claims and litigation, preparation of engineering reports, testimony at hearings

2

and trials in istrative or judicial proceedings, and the rendition of advisory opinions to assist in the resolution of disputes affecting life or property. Within the construction industry, disputes usually result from unfulfilled expectations of the project developers and purchasers, from errors or omissions of the design team, or from construction or material defects. Non-technical issues, principally time and cost, are frequently considerations and unanticipated influences such as weather and unforeseen conditions that may negatively influence the performance of construction. In the “perfectly” constructed project, the expectations of the owner/developer are completely understood by the design team and are converted to a clear, comprehensive set of drawings and specifications. After execution of an equitable contract outlining the and conditions, schedule, and compensation arrangements, the competent construction team executes the construction according to the design within budget and on schedule. And finally, the expectations of the owner/developer/occupant are met. Unfortunately, in our imperfect world, few projects are perfect. Expectations are not met, designs are inadequate, construction is improper, materials fail, and nature exerts its influence. Whether before, or after calling the lawyers, a forensic engineer is usually needed to provide the technical expertise to investigate and answer the questions. The process used by the forensic engineer is an adaptation of the scientific method, “the process by which scientists attempt to construct an accurate (that is, reliable, consistent and non-arbitrary) representation of the world.” The scientific method has four steps: 1. Observation and description of a phenomenon (an occurrence, observation, event, or trend) or group of phenomena i.e. identify the problem. 2. Formulation of a hypothesis (a theory or statement regarding cause and effect) to explain the phenomena, i.e. decide on the procedure. 3. Use of the hypothesis to predict quantitatively the results of new observations. 4. Performance of experimental tests of the predictions with properly performed experiments, i.e. collect the data and perform the analysis. It is the fourth step that often is overlooked. Because the experimental tests may lend either to confirmation of the hypothesis or to the ruling out of the hypothesis, the scientific method requires that a hypothesis be ruled out or modified if its predictions are clearly and repeatedly incompatible with experimental tests. Further, no matter how elegant or desirable a hypothesis or theory is its predictions must agree with experimental results if it is a valid description of nature. The scientific method attempts to minimize the influence of the forensic engineer’s bias or prejudice on the outcome of an experiment. That is, when testing a theory, the engineer may have a preference for one outcome or another, and it is important that this preference not influence the results or their interpretation. Another common mistake is to ignore or rule out data (measurements or information) which do not the engineer’s theory. Ideally, the engineer is open to the possibility that the theory is correct or incorrect. Sometimes, however, the engineer may have a strong belief that the hypothesis is true (or false), or feels internal or external pressure to get a specific result. In that case, there may be a psychological tendency to find “something wrong” with data which do not the engineer’s expectations, while data which do agree with those expectations may not be checked as carefully. 9.3

ETHICAL ISSUES

3

Ethics is defined as a discipline dealing with good and bad, right and wrong, moral duty and obligation. Why does a group need a code of ethics? One does not have to look any further than the daily newspaper to find that unethical behavior exists in the highest levels of government, in the clergy and in virtually every segment of society. Social scientists have studied and found that in any group a certain percentage will deviate from the norm. Therefore, there is the need to codify and regulate or influence the behavior of various professions and groups. But, when we speak of a code of ethics applicable to engineers do we mean merely the business practice with respect to another professional or a client, or are there higher goals and standards to which a professional engineer should aspire? The NSPE code of ethics states that in the fulfillment of their professional duties engineers shall: “1. Hold paramount the safety, health and welfare of the public in the performance of their professional duties; 2. Perform services only in the areas of their competence; 3. Issue public statements only in an objective and truthful manner; 4. Act in professional matters for each employer or client as faithful agents or trustees; 5. Avoid deceptive acts in the solicitation of professional employment.” Upon close examination of the Canons, it is seen that none are in conflict with the practice of forensic engineering. The forensic engineer should adhere to the SPE Cannons not only as a member of NSPE but also as an engineer (Dixon, 1992). Engineers are recognized as the most ethical professionals above doctors, ants, and lawyers in a survey of 200 CEO’s of national companies. While ethical core values for engineers are generally identified as integrity, honesty, faithfulness, charity, responsibility, and self discipline the classical seven deadly sins are described as pride, greed, anger, envy, gluttony, lechery, and sloth. It is clear that there are problems in ensuring the integrity of expert testimony. There are no minimum qualifications for becoming an expert witness, nor any means of regulating them. Code of ethics are voluntary and not often enforced (Grover et al., 2003). In summary, the middle of the road rule for an ethical engineer is to be honest, tell the truth no matter how bad it may seem, and let your conscience be your guide (Baker, 1997). 9.4

LEGAL ISSUES

The legal considerations of these forensic geotechnical engineering services illustrate the reality that the engineering investigation of a failure incident is a fact-finding mission that results in uncovering the probable causes of that failure. It concentrates on the identification of hidden clues. The procedures adopted for the analysis, testing, opinions, and written reports should be able to satisfy even legal scrutiny of their validity. 9.4.1

Role of Expert Witness

Experts are vitally important to universal jurisprudence. Expert witness can assist retained counsel in fact gathering, request for documents, and evaluation of evidence; fashioning precise questions for opposing experts and through candid discussions of strengths and weaknesses of the case; retained counsel to properly qualify and question an expert witness to maximize his

4

impact on the jury; and expansion of the traditional testifying roles in the litigation process. An expert in forensic engineering is not automatically dubbed an “expert witness”. An individual is not granted the privileges of an expert witness until a legal forum confers that recognition upon a properly qualified professional. Experts must possess specialized knowledge of science, profession, business or occupation that is beyond the experience of the lay juror and can assist the fact finder (Grover, 2003). Additionally, expert witnesses in the field of engineering are governed by their own code of ethics. Well over 90% of civil cases settle prior to expert witnesses being called for a trial. As such experts can expect that the majority of testimony that they give will be given at deposition. Accordingly, the experts need to excel during their depositions as they are key element of the discovery process. If the expert witness is a litigation consultant, he/she is not subject to discovery by the opposing counsel. If disclosed as an expert, then he/she is subject to discovery by the opposing counsel. Engineers should perform services only in areas of their competence and they should undertake assignments when qualified by education and/or experience in the specific fields involved. Direct examination is an expert’s opportunity to persuade the jury that his/her opinion should be believed. The goal during direct testimony is to persuade the jury to find one’s opinion creditable. To connect with the jury an expert needs to understand what jurors want and employ the best ways to communicate with them. Experienced expert prepares a well written, persuasive expert based upon reliable accepted methodology. Expert witness can express his/her opinion in an unambiguous and legally sufficient manner. Expert witness can dramatically improve their performance by being good teachers, both to the lawyers who hire them, and to jurors and other fact- finders. They must keep it simple with illustrations, explanation and non-technical language. The four key elements of a successful expert witness assignment are (SEAK, 2006): 9.4.1.1 Part I: Prevention The most valuable experts isolate themselves from attacks and deny opposing counsel ammunition to attack their credentials and credibility. It helps to identify a detailed checklist of potential areas of attack that experts may be subject to regarding their credentials and credibility including: every word on their CVs, past testimony, their image, controversial or political associations, missing credentials, fee schedules, fee agreements, marketing materials, web page, speeches, work on past case, apparent and actual conflicts of interest, non-related litigation, hobbies, professional complaints or discipline, presentations and writings. The best experts form and express airtight opinions and opinions that hold up under the most rigorous scrutiny and cross-examination. There are many ways in which opposing counsel is able to poke holes in an expert’s opinion. The expert must provide specific actions to bulletproof his opinion including, proper case and client selection, avoiding time crunches, using careful and confident language, not overstating or understand facts or opinions, consistency, dealing with the opinions of other experts, knowing exactly what needs to be proved, testing alternative theories, taking careful and precise measurements, being well-trained and wellversed in any computer program used, ing computer results, leaving no stone unturned, taking photographs, ing factual assumption, gaining as much first hand

5

knowledge as possible, thoroughly researching the issues at hand, obtaining and carefully reviewing all relevant documents, not sharing draft reports with counsel, and avoiding “junk science”. 9.4.1.2 Part II: Preparation Peak performance requires proper disciplined preparation done correctly prior to deposition and trial. Well-prepared experts are able to deliver confident testimony, deal with crossexamination far more effectively, and are in much better position to articulate and defend their opinions. It should include advanced techniques that can and should be used to prepare for depositions, direct examination and cross-examination. 9.4.1.3 Part III: Performance The best experts recognize that most cases are won and lost in the discovery phase and that the expert’s deposition is a crucial, often outcome determinative, component of the case. In order to excel at the highest level during a deposition, experts need to be able to recognize and defeat opposing counsel’s deposition tactics and recognize how these tactics differ from those used during trial. The best experts deliver powerful and understandable direct testimony. They also learn to explain and demonstrate numerous advanced techniques for delivering captivating, memorable and persuasive direct expert testimony. These advance techniques include, showing – not telling, getting to the point up front and explaining later, being well-prepared and wellorganized, making the complex simple, entertaining, being likeable, highlighting your most relevant qualifications, working on a smooth flow and style, getting out of the jury box early and often, using visual aids that work, aggressively self-editing, employing powerful, memorable analogies, showing your human side and bonding with the jury, using precise language, using confident language, employing short preview and review summaries, using numbered lists, citing references, speaking conversationally, conforming your testimony to the theme of the case, and reading and reacting to the jury. 9.4.1.4 Part IV: Practice The best experts skillfully answer cross-examination questions and explain how the question could have been avoided, how they could have and should have prepared to answer the question, identifying the tactic that counsel is using and delivering a response that defeats the tactic and/or allows the expert to go on the offensive. 9.4.2 Legal Considerations In order to fully evaluate the forensic expert’s role in the legal process, the expert should understand the definitions and concepts of a dispute resolution. The professional services provided in the investigative process and the duties expected of the expert witness are generally the same no matter what method of formal dispute resolution is selected by the client. Following is an o v ervi ew o f the v arious methods of dispute resolution that may be encountered (Task Committee on Guidelines for Failure Investigation, 1989). 9.4.2.1

Civil Litigation

This process occurs in a court of justice having jurisdiction in the dispute. The court is presided over by a judge who decides on the applications of the law and procedure. These

6

decisions usually are determined by a jury of lay-persons or by the judge in the absence of a jury. The decision of this court and jury may be subjected to appeals and retrials. 9.4.2.2 Arbitration Instead of a judge and/or jury, an arbitrator or a of arbitrators are used in a private proceeding. Unlike litigation, arbitration eliminates or significantly relaxes the discovery process and is usually limited to the production of documents to an extent determined by the arbitrators. 9.4.2.3 istrative Hearings or Special Investigations Hearings are held by governmental authorities to resolve disputes or to gather information of public opinion prior to making an istrative decision, or for inquiries into the integrity of the structure or the cause of a failure of an engineering facility. 9.4.2.4 Private Litigation This form of dispute resolution requires both parties to agree to resolve their differences in a private court that is created, and paid for, by both parties. A mutually respected individual, such as a retired jurist, is chosen as a judge, and civil litigation rules are followed. 9.4.2.5 Mediation This method of “non binding” dispute resolution is used to resolve certain construction-oriented disputes. A mediator acceptable to both sides is selected and probes the disputants to determine their position relative to “minimum” settlement. He or she then strives to attain a middle ground that both can accept comfortably. 9.5

OVERVIEW

A well organized, methodical approach to failure investigation is needed because of the complexities of modern engineered facilities and the vast variety of possible failure causes. The investigative process should n o t employ a rigid “cookbook” approach, but there are certain steps that are common to an effective failure investigation. An emphasis on an open-minded acceptance of all pertinent data, while recognizing the relevant information, is the fundamental basis of a failure investigation. As various hypotheses develop and change, continuous review of previously rejected data is required. There is seldom a single cause of a failure but rather a complex interaction of components and forces. As a result, the outcome of a failure investigation seldom leads to absolutely irrefutable results but rather to a most probably cause of failure. While the collection of data and facts surrounding the failure are specific, expert opinions arising from these data may differ widely.

The most accepted failure investigation findings will be the one employing a qualified investigation team that presents the most plausible failure scenario, based on well documented data. (Task Committee on Guidelines for Failure Investigation, 1989). A close coordination between the forensic geotechnical engineer, his client, and client’s

7

attorney is important in order to understand the facts as they relate to a particular incident. This teamwork normally lends credence to their case and is instrumental in making it a winning combination. Two case histories that include all elements of forensic geotechnical engineering as well as legal issues of jurisprudence system are included with this chapter. They include situations where forensic geotechnical engineering was effectively utilized to identify, investigate, remediate, litigate, and resolute legal issues. 9.6

REFERENCES

ASC geosciences, inc. (2002) – Report of Field Investigation, Data Evaluation, and Engineering Consultation Services, ASC geosciences, inc., Report No. 02L2016. Baker, P.E., John A. (1997) – Professional Engineering Ethics versus Expert Witness Requirements, Los Angeles Section ASCE, Forensic Engineering Technical Group. Bell, P.E., John T. (2007) – What is Forensic Engineering?, Florida Engineering Society Journal, May 2007, pp.8-9. Dixon, E. Joyce (1992) – The NSPE Code of Ethics Applied to Forensic Engineering, Journal of the National Academy of Forensic Engineers, Vol. IX No. 1. Grover, J.D., P.E., J.L. (2003) Ethical Considerations for Expert Witnesses in Forensic Engineering, Ethical Dilemmas of Technical Forensic Practice, pp. 441-52. Saxena, D.S., P.E. (2005) – "Forensic (Geo-technical and Foundation) Engineering Case History", National Academy of Forensic Engineers (NAFE) Seminar, Chicago, Illinois, 10 July 2005. Saxena, D.S., P.E. (2007) – "Case Studies in Forensic Geotechnical and Foundation Engineering", 6th International Conference on Case Histories in Geotechnical Engineering, Arlington, VA, Paper No. OSP-2. SEAK inc. (2006) – Advanced Testifying Skills for Experts – The Master’s Program, pp. 4-11. Task Committee on Guidelines for Failure Investigation (1989) – Guidelines for Failure Investigation, American Society of Civil Engineer (ASCE).

B.

PART 2

9.7

Appendices

9.7.1

Case History One

Project specifics and forensic facts summary; •

The backyard slope of a one-story dwelling subsided abruptly, failed, and

8

slid into the lake along with the rip-rap from lake edge. It also extended into the neighboring property on the south side. •

The lakefront backyard of a residence was damaged and stability of the structure was threatened, and removal/replacement of the completed structure was considered a viable yet costly option.

•

The owner filed a claim against the developer/builder for negligence and for not informing them of a potentially unstable pre-existing condition.

•

The owner secured the services of a Forensic Geotechnical Engineer (FGE) to investigate, remediate, and assist in litigation as well as serve as an expert witness during the resolution of dispute and legal issues.

•

The developer/builder offered a band- aid solution that the owner then rejected upon advice of the FGE.

Figures 1-4 illustrate the salient features of this case history.

Fig. 1 Close-up view of the subsided backyard

9

Fig. 2 View of subsided slope of ading house

Fig. 3 View of sodded and finished backyard

10

Fig. 4 Reoccurrence of slope failure in finished backyard two weeks after completion 9.7.1.1 Forensic Field Exploration and Subsurface Condition Evaluation Subsurface stratigraphy beneath the site, as illustrated in Figure 7, consisted of very soft, loose low strength subsoils to 15.0 ft depth. These low strength subsoils consisting of highly plastic clays (mc 113%, LL-107, PI-89) which were unstable under the weight of the additional fill material. Slope stability analyses were performed for the original subsurface profile as well as the contractor/developer proposed restructured slope that yielded Factors of Safety (F.S.) of 0.87 and 0.93, respectively, for the two conditions. 9.7.1.2 Observations and Findings The expert’s investigation at the project development site revealed that; •

Slope failure was triggered by the presence of unconsolidated sediment layers under the site fill.

•

Unstable sediments to continue to consolidate resulting in continued movement of the back of the house and the backyard.

•

A retaining wall should have been constructed prior to the house being built to contain and stabilize the soils beneath the house foundation.

11

Fig. 7 Subsurface stratigraphy identifying delineation of compressible phosphatic clay Upon the advice of their legal counsel the owner put the developer on notice, and elected to proceed with the proper fix pending resolution of litigation and the claim. 9.7.1.3 Remediation Based on findings and observations from the FGE investigation, a two part remediation program was recommended and consisted of: 1. Installation of helical anchor piers under the exterior wall footing, along the back of the house, for the underpinning stabilization. 2. Installation of a helical anchor bulkhead to stabilize the backyard along the lake. 3. Based upon field monitoring of pier installation operations, anchored timber bulkhead, post stabilization walk through, and review/evaluation of the field stabilization data, it was concluded that the methods and procedures utilized and installed by the specialty contractor were effective, satisfactory, and acceptable. Slope stability analyses of this remediated structure indicated a Factor of Safety (F.S.) of 1.57.

Figures 6-9 illustrate various sequences of the underpinning and bulkhead stabilization program.

12

Fig. 6 View of cap and bracket assembly over helical anchor under the footing

Fig. 7 Close-up of geotextile lay down and wrap up at helical anchor tie back on timber bulkhead

13

Fig. 8 Additional fill being dozed into the anchored geotextile

Fig. 9 View of completed and restored backyard 9.7.1.4 Concluding Remarks 1. Following the close coordination between the forensic geotechnical engineer, the owner, and the remediation contractor and based upon results of remediation monitoring it was determined that the repair of the residence structure had been achieved satisfactorily. 2. The final repair resulted in restoration of the residential structure and backyard to its originally planned and constructed stage, as illustrated in Figure 9. 3. As a result of the monitored and satisfactory remediation program, the consultant recommended acceptance of the restored structure. 4. The property owner retained an attorney who sent a notice of claim, and remediation

14

cost summary, to the developer/builder. Following litigation, arbitration the property owner, through their legal counsel, was able to recover these remediation costs along with attorney fees. 9.7.2

Case History Two

This case history identifies a request received from a building owner for a forensic engineering review to investigate a situation where a severe site specific soil-structure deficiency occurred and caused post construction damage to an office building in west central Florida. Evaluations were also made to see if repairing or replacement of the structure was, in fact, necessary. 9.7.2.1

Project Description

Construction of an office building was completed in June 2006 as per design plans, permitting, and pre-construction technical from the civil engineer. Foundation included installation of 12 inch diameter timber piling, driven to 24 tons capacity (as per the pile driving records prepared by a geotechnical engineering representative at the site) along the exterior perimeter grade beam ing the load bearing walls. The interior slab was a 4 inch thick fiber mesh concrete ed on compacted soil for the entire building footprint. Soon after occupying the building, the owner(s) observed certain impacts on their property. Concerns identified by the owner(s) had included: •

noticeable settlement and deflection of the interior soil ed floor slab.

•

visible cracks throughout the interior of the building.

•

differential settlement of building foundations (exterior piles and interior floor slab) resulting in misalignment of the doors.

•

possible damage to the below grade utility lines.

•

substantial deflection distress and development of cracks throughout the interior and foundation floor slab within the office building as a consequence of differential settlement due to the consolidation of existing peat material within the building footprint area. Additionally, total and differential settlement of the foundation system and resulting distress appeared to be of a continuing and progressive nature.

Figures 1 – 8 illustrate the salient features of the case history and owner’s justifiable concern.

15

Fig. 1 Two close up views of sinking floor at conference room door

Fig. 2 Close up view of interior partition wall separation from sinking floor slab

Fig. 4 Close up view of wall crack showing new crack and additional movement

16

Fig. 8 Separation and sinking of slab along building exterior A 2-stage program was carried out by the forensic geotechnical consultant in conjunction with filing of a claim by the building owner. 9.7.2.2 Stage 1 Investigation Stage 1 consisted of a review of the field inspection of the structure, review of the project foundation drawings, and pre-design geotechnical exploration report.

The findings, comments, and conclusions derived from the Stage 1 investigation at the project site revealed that: •

there were wide spread separations between the walls and the floors, and between partition walls and ceilings. The separations appeared to be greatest toward the center of the building.

•

a floor elevation survey of the building floor slab conducted in March 2007 indicated partial settlement of 7.5 cms (3 inches) at the center of the building, as illustrated in Fig. 9.

•

subsoil test borings performed by the pre-design geotechnical engineer were 12.0 m (40 ft) deep and consisted of organic compressive soils. Pressure treated timber piles with an embedment depth of 20 to 25 ft (for a 12 inch diameter) were utilized and a net allowable capacity of 12 tons was recommended.

•

during construction pile lengths of 16.0 m to 18.0 m (50 to 55 ft) were used. Furthermore, 30.0 cm (12 inch) square prestressed precast concrete (PCC) piles were substituted in lieu of pressure treated timber piles. In addition, the contractor elected to use an unusually high energy hammer in lieu of the recommended hammer of low to moderate energy for the timber pile.

17

•

All three of the borings had significant weight-of-rod (WOR) and weightof-hammer (WOH) zones throughout the in filled soils. In those cases where underlying sandy soils were encountered, it appeared that the deep sands were in a more stable condition. This profile indicated an ongoing raveling associated with the relic sinkhole, or an indication that the in filled organic soils were weak and very loose.

•

the interior floor slab was only soil ed and not structural (with interior grade beam ed).

•

proper pile capacity determination, its installation, and structural grade beam ed floor slab was not performed.

9.7.2.3 Stage 2 Investigation In an effort to properly address, investigate, and evaluate the alleged subsoil deficiency a detailed subsoil exploration program was undertaken by advancing test borings to depths ranging from 40.0 m (130 ft) to 60.0 m (190 ft). Test boring stratigraphy showing internal erosion and solution features is illustrated in Fig. 5. It consisted of elements listed below.

•

the site was underlain by very poor soil conditions not capable of providing for any type of construction that assumes bearing to be developed from the underlying soils.

•

upper fill soils were mixed with debris underlain by predominately organic soils with organic content ranging from 5% (at 80 ft depth) to 18.5% (at 60 ft depth) and 85% (at 135 ft depth).

•

no limestone was encountered to the termination depth of 43.0 m (135 ft) to 60.0 m (190 ft). Project site stratigraphy showing internal erosion and solution features is illustrated in Fig. 5.

18

Fig. 5 Project site stratigraphy showing internal erosion and solution features •

damage to the building was the result of factors including consolidation of significant factors including consolidation of significant depth of very loose organic soils (located within a large relic sinkhole) along with the consolidation of buried debris located within or just below the develop fill soils and loading street zone.

•

the old aerial photographs from 1941 through 2005 showed the wetland feature and confirmed that the building was constructed in a wetland drainage basin area over deposits of very deep, very loose organic soils.

9.7.2.4

Concluding Remarks •

the owner reported that the damage had been ongoing since the building was constructed, which was consistent with a concrete slab ed by very loose, compressible, debris laden organic soils. The pile ed perimeter walls ed by the timber pile showed very little damage, indicating that no significant differential movement of the perimeter wall had yet occurred.

•

it was determined that deleterious soil condition was known to exist prior to construction of the building and is the reason that a pile foundation was recommended by the project geotechnical engineer. However, the foundation design only ed for of the perimeter walls and did not address the need for of the slab. It was concluded that this design flaw/oversight was the primary reason that the building had sustained substantial interior damage.

•

it was also concluded that no feasible and effective options were

19

available to remedy the relic sinkhole condition and that the wetland site over deposit of very deep organic soil was not a suitable site for this building construction. •

additionally, as a direct result of structural and geotechnical engineer’s severe design deficiency, as well as contractor’s failure to exercise, due care the building had sustained substantial irreversible damage.

•

the property owner’s claim to the developer/builder was successful and the owner was able to recover a major portion of the damage cost resulting from design oversight and relic sinkhole related damage.

20

Chapter 13. POTENTIAL ROLE OF RELIABILITY IN FORENSIC GEOTECHNICAL ENGINEERING K K Phoon Department of Civil Engineering, National University of Singapore, Singapore 117576 E-mail [email protected] G L Sivakumar Babu Department of Civil Engineering Indian Institute of Science, Bangalore, India, 560012 E-mail: [email protected] M Uzielli Georisk Engineering S.r.l., Florence, Italy E-mail : [email protected]

ABSTRACT This chapter discusses some possible roles for reliability and risk in forensic geotechnical engineering. A preliminary statistical framework is presented to quantify the difference between expected and observed performance in the presence of unavoidable and potentially significant geotechnical variabilities. The probable ranges of geotechnical variabilities are reviewed. Other potentially useful results in the recent reliability and risk literature are highlighted. The intention of this chapter is to stimulate further discussions and research in this important but somewhat overlooked area.

1

1. INTRODUCTION Forensic geotechnical engineering (FGE) is a relatively new field. ISSMGE TC40 was formed only in 2005 under Dr VVS Rao as chair. Subcommittee 6 was convened under Dr KK Phoon to explore the potential role of reliability in FGE. A cursory Google search using keywords “reliability, forensic, and geotechnical” shows that this concept is entirely new. This chapter seeks to explore if reliability/risk concepts are potentially useful to forensic geotechnical engineering. This study is very preliminary given the dearth of previous literature. Brown (2006) noted that: “The concept of reliability as ‘the likelihood that a system will perform in an acceptable manner’ (Bea, 2006) is important in forensic geotechnical engineering.” No details were discussed. Engineering with natural materials such as soils and rocks is difficult and challenging owing to the variability and the uncertainties associated with engineering decisions. In fact, one key aspect that distinguishes geotechnical engineering from structural engineering is the large variability (natural/intrinsic variability, testing errors, and transformation uncertainties introduced when measured parameters are converted to engineering parameters) related to naturally occurring geomaterials. In the case of quality of concrete, variability (in of coefficient of variation or COV) can be classified as follows (Rétháti 1988 citing 1965 specification of the American Concrete Institute): COV < 10% COV = 10 – 15% COV = 15 – 20% COV > 20%

excellent good satisfactory bad

The COVs related to the natural variability of geomaterials as well as those pertaining to measurements errors and transformation uncertainties can be much larger and do not fall within a narrow range. Phoon & Kulhawy (1999a; 1999b) compiled an extensive database of soil statistics for calibration of simplified reliability-based foundation design equations. Details are given in Section 3. One key observation that emerged from their calibration studies is that geotechnical variability of common design parameters can be broadly grouped using the following qualitative categories: Geotechnical parameter Undrained shear strength

Effective stress friction angle

Horizontal stress coefficient

Property variability Low Medium High Low Medium High Low Medium High

COV (%) 10 – 30 30 – 50 50 – 70 5 – 10 10 – 15 15 – 20 30 - 50 50 – 70 70 - 90

These observations are not entirely surprising given that the volume of geo-materials investigated by direct or indirect means is extremely small in comparison to the volume of interest. Chiles and Delfiner (1999) cited volume fractions investigated at Brent field, North sea, to be 10-9 for cores and cuttings and 10-6 for logging. The immediate impact of these observations is that Leonards (1982) definition of “failure” - unacceptable difference between expected and observed performance – cannot be evaluated in a meaningful way using deterministic methods. To elaborate, “expected performance” must vary given the backdrop of potentially significant geotechnical variability. We can argue that this variability is “inherent”, in the sense of being irreducible, in geotechnical engineering based on the prevailing standards of best practice. Some reduction in some of the components of variability is always possible by carrying out more tests, but it is generally not realistic to expect orders of magnitudes reduction. In view of this widely acknowledged and widely established reality in geotechnical engineering, it is more realistic and perhaps more credible to quantify “unacceptable difference” in a statistical sense. In broad , forensic engineering is related to the investigation of failures with the view of rendering an opinion regarding responsibility. Hence, an objective statistical measure of “unacceptable difference” (specifically, a difference not explainable by underlying variability) should provide useful additional information in the formulation of such an opinion. It should be emphasized, however, that the statistical criteria tentatively proposed in the following sections should never replace – nor downplay the importance of - preliminary qualitative assessments of the quality of design based on geotechnical knowledge

2

and experience. Rather, they should serve as a complementary tool that could improve the degree of objectivity in forensic geotechnical assessments. 2. STATISTICAL FRAMEWORK 2.1 Rejection criteria for factor of safety A preliminary statistical framework can be constructed by using the widely used global factor of safety (FS) as an indicator of performance. It is plausible to imagine FS as a lognormal random variable with mean = μ, standard deviation = σ and coefficient of variation = σ/μ = θ. Silva et al. (2008) provided some empirical evidence to this lognormal assumption for slope problems. The lognormal distribution satisfies the most basic constraint that most load and strength parameters are non-negative. It is implicit in this model that one speaks of the “likelihood” of failure, rather than failure in our usual absolute deterministic sense. Fig. 1 illustrates a lognormal distribution for μ = 3 and θ = 0.3. It can be shown that ln(FS) is normally distributed with mean (λ) and standard deviation (ξ) given by: λ = ln(μ) – 0.5ξ2 and

ξ2 = ln(1+θ2).

(1)

The equivalent normal distribution for Fig. 1 is shown in Fig. 2 with λ = 1.056 and ξ = 0.29. Note that the standard deviation of ln(FS) (ξ) is approximately equal to the COV of FS (θ) up to θ of about 0.5.

Probability density

0.6 Fail 0.4

0.2

0 0

2

4

6

FS Fig.1 Log-normal distribution of FS

3

1.5

Probability density

Fail 1.0

0.5

0.0 -1

0

1

2

3

ln(FS) Fig. 2 Equivalent normal distribution The observed performance, FS , can be viewed as a sample mean (of sample size 1) from a population of factors of safety such as that shown in Fig. 1. It is obvious that a low value of FS near to 1 is not impossible, although it is “unlikely” if the population follows a lognormal distribution with μ = 3 and θ = 0.3. A naïve interpretation is that “failure” (as in FS < 1) is always possible and geotechnical variability is responsible, rather than human errors. Sowers (1993) noted that the majority of foundation failures were due to human shortcomings. This naïve interpretation essentially misses the key principle of hypothesis testing described below. Our proposed approach is to assume that: (a) “expected performance” can be described by the reliability index and (b) a conventional hypothesis test can be performed to ascertain if the target reliability index had been achieved in the original design based on the observed sample mean, FS . To do this, it is necessary to define “expected performance”. If FS follows a log-normal population, the reliability index is given by:

β=

E[ln(FS)] λ = σ[ln(FS)] ξ

(2)

Numerous reliability calibration studies (e.g., Phoon et al. 1995) have shown that existing foundations are typically designed to achieve a target reliability index (β) of about 3 (corresponding to a probability of failure of about one in a thousand). Target reliability indices for other types of geotechnical design are not as well established presently. In Fig. 2, the reliability index is β = 3.64 or probability of failure = Φ-1(-β) = 0.000136, where Φ is the standard normal cumulative distribution function. In EXCEL, Φ(.) and Φ-1(.) are computed using the NORMSDIST(.) and NORMSINV(.) functions, respectively. Based on the definition of the reliability index and β = 3, the following null and alternate hypotheses on the population mean [actually, the mean of ln(FS)] can be formulated as: H0: λ = 3ξ H1: λ < 3ξ

4

Assuming that ξ is known and a sample size of 1, the null hypothesis is rejected at the customary 5% level of significance if:

( )

( )

ln FS − λ ln FS − 3ξ = < Φ −1 (0.05) = −1.645 ξ ξ

[

(

FS < exp(1.355ξ) = exp 1.355 ln 1 + θ 2

(3)

)]

(4)

The above rejection criterion provides an example of a simple numerical yardstick to evaluate “unacceptable difference between expected and observed performance” in the presence of potentially significant geotechnical variability. A rejection means that the observed factor of safety does not the claim that β = 3 at the 5% level of significance. Different values of the threshold factor of safety can be computed for different significance levels. It is also possible for a rejection to arise because the underlying “true” geotechnical variability was grossly over-estimated. A “do not reject” scenario means that the observed factor of safety is not unreasonably “low” and failure may be caused by geologic “surprises”, limitations in the existing factor of safety, critical failure mechanism not identified, etc. Again, it is possible for “do not reject” to arise because we have grossly under-estimated the underlying geotechnical variability. The above framework is by no means perfect and comprehensive, given the diversity and complexities of actual failures. However, it does provide an example of an objective framework to perform an initial evaluation of the observed factor of safety, particularly to eliminate the more obvious claim that it is an “unfortunate” realization caused by geotechnical variability. Fig. 3 presents the critical values of factor of safety below which the difference between expected and observed performance is “unacceptable”. For example, if the expected performance corresponds to β = 3 and the underlying variability corresponds to θ = 0.3, an observed factor of safety of 1.4 is too “low” and not explainable as a random outcome from a population of FS with β = 3.0 and θ = 0.3. It will be useful to interpret actual case studies using this chart to evaluate its usefulness in forensic geotechnical engineering and to finetune it if necessary. In the general case where we observe n factors of safety, the rejection criterion for β will become:

(

FS < exp βξ − 1.645 ξ

[

(

n

)

= exp ln(1 + θ 2 ) β − 1.645

n

)]

(5)

The rejection curves for β = 3.0 corresponding to various sample sizes (n) are shown in Fig. 4. If this statistical framework proves to be useful, more sophisticated rejection criteria can be developed based on a sample estimate of ξ (rather than the population version used in the above equations). It suffices to note here that they do not follow the t-distribution in standard statistical texts.

5

3.0 β = 3.5

Critical factor of safety

2.8 2.6 2.4 2.2

β = 3.0

2.0 1.8 β = 2.5

1.6 1.4 1.2 1.0 0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

Coefficient of variation Fig. 3 Rejection criteria for different expected performance levels - a possible role for risk in FGE based on one observed factor of safety

3.0

Critical factor of safety

2.8

n=5

2.6 2.4 n=2

2.2 2.0 1.8

n=1

1.6 1.4 1.2 1.0 0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

Coefficient of variation Fig. 4 Rejection criteria for different sample sizes of observed factors of safety based an expected performance of β = 3. 2.2 Example An example of the use of Fig. 3 can be illustrated with reference to the results presented in Duncan (2000) wherein a case study of underwater slope failure was reported. The failure took place entirely within San Francisco Bay mud, a normally consolidated, slightly organic clayey silt or silty clay of marine origin. Previous experience in the area indicated that 1(H):1(V) slopes with a factor of safety of 1.25 were satisfactory. To reduce the quantity of excavation, slopes of 0.875(H): 1(V) having a factor of safety of 1.17 were excavated which failed subsequently. A risk-based back-analysis indicated a probability of failure of 18% which is

6

unacceptable though the factor of safety is 1.17. Duncan (2000) mentioned that the coefficient of variation of shear strength parameters was high and hence the probability of failure was high. This case study can be plotted in Fig. 3 as an open circle. It is clear that an observed FS = 1.17 cannot the claim that β = 2.5 for any COV larger than about 0.2. Following the same argument, we can conclude that β = 3 is not ed for any COV larger than about 0.1. Because COV of 0.1 is the lower bound for most geotechnical problems, it is reasonable to say that β = 3 was not achieved in the original design with a fair degree of confidence. 3. GEOTECHNICAL UNCERTAINTIES The evaluation of geomaterial (soil and rock) parameters is one of the key design aspects that distinguish geotechnical from structural engineering. The basic premise underlying Section 2 is that the factor of safety (FS) is a lognormal random variable. It is possible to infer the underlying coefficients of variation of FS from the empirical evidence presented in Silva et al. (2008). This approach of lumping all sources of geotechnical uncertainties into a COV of FS is quite crude, but it does allow practitioners to apply the statistical framework in Section 2 rapidly to test the claim that failure is an unfortunate “roll of the dice” caused by geotechnical uncertainties. Some important COV thresholds for FS are given in Section 3.1. It is preferable to distinguish between two main sources of geotechnical uncertainties explicitly. The first arises from the evaluation of design soil properties, such as undrained shear strength and effective stress friction angle. This source of geotechnical uncertainty is complex and depends on inherent soil variability, degree of equipment and procedural control maintained during site investigation, and precision of the correlation model used to relate field measurement with design soil property. A total variability analysis that lumps all these components produces only site-specific statistics which should not be generalized to other sites. Realistic statistical estimates of the variability of design soil properties have been established by Phoon & Kulhawy (1999a, 1999b) and presented in Section 3.2. The second source arises from geotechnical calculation models. Although many geotechnical calculation models are “simple”, reasonable predictions of fairly complex soilstructure interaction behavior still can be achieved through empirical calibrations. Model factors, defined as the ratio of the measured response to the calculated response, usually are used to correct for simplifications in the calculation models. Recent literature includes estimation of model statistics for the calibration of deep foundation resistance factors for AASHTO [American Association of State Highway and Transportation Officials] (Paikowsky 2002). None of these studies addresses the applicability of model statistics beyond the conditions implied in the database. This question mirrors the same concern expressed previously on the possible site-specific nature of soil variabilities. More rigorous model statistics for both ultimate and serviceability limit states are presented in Section 3.3. It is crucial to highlight that the COV of FS is indeed exceedingly crude, because it lumps parametric and model uncertainties into a single number. 3.1 Coefficient of variation of factor of safety Silva et al. (2008) proposed several relationships between the annual probability of failure and the factor of safety based on 75 projects (zoned and homogeneous earth dams, tailings dams, natural and cut slopes, and some earth retaining structures) and expert judgment. As shown in Fig. 5, there are 4 categories of earth structures. They are defined by Silva et al. (2008) as: ƒ Category I—facilities designed, built, and operated with state-of-the-practice engineering. Generally these facilities have high failure consequences; ƒ Category II—facilities designed, built, and operated using standard engineering practice. Many ordinary facilities fall into this category; ƒ Category III—facilities without site-specific design and substandard construction or operation. Temporary facilities and those with low failure consequences often fall into this category; ƒ Category IV—facilities with little or no engineering. Silva et al. (2008) compared the empirical data in Fig. 5 with some theoretical curves, but did not provide any mathematical details. Their theoretical curves can be easily reproduced using the following procedure: 1. Assume that FS is lognormally distributed with parameters λ and ξ. 2. The horizontal axis of Fig. 5 is the mean factor of safety, μ. 3. If the COV of FS (θ) is sufficiently small, λ ≈ ln(μ) and ξ ≈ θ. 4. The vertical axis of Fig. 5 is the probability of failure, given by:

7

⎛0−λ⎞ ⎡ − ln (FS) ⎤ ⎟⎟ = Φ ⎢ p f = Pr ob(FS < 1) = Pr ob[ln(FS) < 0] = Φ⎜⎜ ⎣ θ ⎥⎦ ⎝ ξ ⎠

Fig. 5 Annual probability of failure versus factor of safety for earth structures (Silva et al. 2008) The theoretical lognormal curves from Silva et al. (2008) are reproduced in Fig. 6 using the above procedure with θ = 0.072 (Category I), 0.109 (Category II), 0.174 (Category III), 0.316 (Category IV). Based on Silva et al. (2008), it would appear that the COV of FS for standard earth structures is between 0.1 and 0.2. These COV thresholds are important for FGE and they are summarized in Table 1. Lognormal probability curves for a more complete and more systematic range of COVs are shown in Fig. 7.

8

1.E-08

Annual probability of failure

1.E-07 1.E-06 1.E-05 1.E-04 1.E-03 I (COV = 0.072) 1.E-02

II (COV = 0.109)

1.E-01

III (COV = 0.174) IV (COV = 0.316)

1.E+00 0.5

1.5

2.5 3.5 Mean factor of safety

4.5

Fig. 6 Lognormal probability curves back-calculated from Silva et al. (2008)

1.E-08

Annual probability of failure

1.E-07 1.E-06 1.E-05 1.E-04

Coefficient of variation (COV)

0.05

1.E-03

0.10 1.E-02

0.20 0.30

1.E-01

0.40 0.50

1.E+00 0.5

1.5

2.5 Mean factor of safety

Fig. 7 Lognormal probability curves for higher COVs of FS

9

3.5

4.5

Table 1. COV thresholds for factors of safety corresponding to different categories of earth structures (back-calculated from Silva et al., 2008)

Categories of earth structures I II III IV

COV(FS) 0.072 0.109 0.174 0.316

3.2 Parametric uncertainties There are three primary sources of geotechnical uncertainties as illustrated in Fig. 8: (a) inherent variability, (b) measurement error, and (c) transformation uncertainty (Phoon and Kulhawy 1999a). A comprehensive effort was undertaken to provide realistic geomaterial statistics of sufficient generality to underpin current and future developments of practical reliability-based design codes (Phoon and Kulhawy, 1999a, 1999b; Kulhawy et al., 2000). The results of this compilation effort are also useful for FGE and are summarized below. Others are reported by Jones et al. (2002) and Uzielli et al. (2006). 3.2.1 Inherent or natural variability Inherent variability results primarily from the “real” heterogeneity which is inherent to geomaterials. Such heterogeneity stems from natural geologic and physical/chemical/biological processes that produced and continually modify the soil/rock mass in-situ. Uzielli et al. (2006) provided a state-of-the-art review of methods for quantifying inherent variability in geotechnical engineering analyses and design. Approximate guidelines for inherent soil variability are given in Table 2. A more detailed presentation containing the mean, standard deviation (S.D.), and range of the COV are shown in Table 3 with the total number of data groups per test (m). The statistics for rock are given for comparison.

Fig. 8 Sources of uncertainties contributing to overall uncertainty in design soil parameter. Table 2. Approximate guidelines for ranges of mean values and COVs of inherent soil variability (Phoon and Kulhawy 1999a). Test type Lab strength

Parametera su(UC)

Soil type Clay

Mean 2 10-400 kN/m

10

COV(%) 20-55

T

VST SPT DMT

su(UU) su(CIUC) φ qT qc

10-350 kN/m 2 150-700 kN/m o 20-40 2 0.5-2.5 MN/m 2 0.5-2.0 MN/m 2 0.5-30.0 MN/m 2 5-400 kN/m 10-70 blows/ft 2 100-450 kN/m 2 60-1300 kN/m 2 500-880 kN/m 2 350-2400 kN/m 1-8 2-30 2 10-50 MN/m

10-30 20-40 5-15 < 20 20-40 20-60 10-40 25-50 10-35 20-50 10-35 20-50 20-60 20-60 15-65

2

EPMT

Clay Sand Sand

400-2800 kN/m 2 1600-3500 kN/m 2 5-15 MN/m

10-35 20-50 15-65

wn wL wP PI LI γ, γd Dr

Clay & silt Clay & silt Clay & silt Clay & silt Clay & silt Clay & silt Sand

13-100 % 30-90 % 15-25 % 10-40 % 10 % 3 13-20 kN/m 30-70 %

su(VST) N A reading B reading ID KD ED

PMT

Lab index

a-

bcd-

2

Clay Clay Clay & sand Clay Clay Sand Clay Clay & sand Clay Sand Clay Sand Sand Sand Sand

pL

8-30 6-30 6-30 b b < 10 c 10-40 d 50-70 su = undrained shear strength; UC = unconfined compression test; UU = unconsolidated-undrained triaxial compression test; CIUC = consolidated isotropic undrained triaxial compression test; φ = effective stress friction angle; qT = corrected cone tip resistance; qc = cone tip resistance; VST = vane shear test; N = standard penetration test blow count; A & B readings, ID, KD, & ED = dilatometer A & B readings, material index, horizontal stress index, & modulus; pL & EPMT = pressuremeter limit stress & modulus; wn = natural water content; wL = liquid limit; wP = plastic limit; PI = plasticity index; LI = liquidity index; γ & γd = total & dry unit weights; Dr = relative density COV = (3-12%) / mean total variability for direct method of determination total variability for indirect determination using SPT values

Table 3. Inherent soil variability for soil and rock (Kulhawy et al. 2000). Test type Index

Strength

Parameter γ, γd wn wP wL PI - all data - ≤ 20% - > 20% γ, γd n φ, tan φ

Material type fine-grained fine-grained fine-grained fine-grained fine-grained fine-grained fine-grained rock rock sand, clay

m 14 40 23 38 33 13 20 42 25 48

11

Coefficient of variation (%) mean S.D. 7.8 5.8 18.1 7.9 15.7 6.0 18.1 7.1 29.5 10.8 35.0 11.4 26.0 9.0 0.9 0.7 25.9 19.4 13.9 10.4

range 2-20 7-46 6-34 7-39 9-57 16-57 9-40 0.1-3 3-71 4-50

Stiffness T

su qu qt-brazilian Et-50 qc

VST SPT DMT

qT su(VST) N A,B

sand clay clay rock rock rock sand, clay sand clay clay clay sand, clay sand, clay sand clay sand

ID - all data - w/o outliers KD - all data - w/o outliers ED - all data - w/o outliers

sand sand

32 16 100 184 74 32 65 54 11 9 26 23 56 30 26 30 29 31 29 31 30

9.0 23.5 31.5 14.2 16.6 30.7 36.6 38.2 28.4 7.9 25.3 38.0 27.9 34.8 19.9 40.7 37.7 41.2 37.6 42.7 41.1

3.0 13.0 14.2 11.7 10.4 15.0 15.5 16.3 6.8 4.9 6.5 10.8 11.9 11.3 6.2 21.6 14.2 19.2 13.5 19.6 17.6

4-15 10-50 6-80 0.3-61 2-58 7-63 10-81 10-81 16-40 2-17 13-36 19-62 12-59 13-59 12-38 8-130 8-66 15-99 15-67 7-92 7-69

3.2.2 Measurement error Inherent variability is caused primarily by the natural geologic processes that are involved in soil formation. Measurement error, on the other hand, arises from equipment, procedural/ operator, and random testing effects. A summary of total measurement error is given in Table 4 for laboratory tests and in Table 5 for field tests.

Table 4. Summary of total measurement error of some laboratory tests (Phoon & Kulhawy 1999a). Parametera

su(TC) su(DS) su(LV) φ(TC) φ(DS) tan φ(TC) tan φ (DS) wn wL wP PI γ

Soil type Clay, silt Clay, silt Clay Clay, silt Clay, silt Sand Sand, silt Clay Fine-grained Fine-grained Fine-grained Fine-grained Fine-grained

No. data group 11 2 15 4 5 2 6 2 3 26 26 10 3

No. tests/group Mean 13-17 9-13 9-13 26 82-88 41-89 41-89 41-89 82-88

Range 13 15 10 11 26 85 64 62 61 85

Property value (unitsb) Mean Range 7-407 125 108-130 119 4-123 29 2-27 o 19.1o 33.3o 24-40o o 30-35 32.7o 16-21 18 17-113 36 12-35 21 4-44 23 16-17 17.0

Property COV (%) Mean 8-38 19-20 5-37 7-56 3-29 13-14 2-22 6-22 6-12 3-11 7-18 5-51 1-2

Range 19 20 13 24 13 14 8 14 8 7 10 24 1

a - su = undrained shear strength; φ = effective stress friction angle; TC = triaxial compression test; UC = unconfined compression test; DS = direct shear test; LV = laboratory vane shear test; wn = natural water content; wL = liquid limit; wP = plastic limit; PI = plasticity index; γ = total unit weight b - units of su = kN/m2; units of wn, wL, wP, and PI = %; units of γ = kN/m3

12

Table 5. Summary of measurement error of common in-situ tests (Kulhawy & Trautmann 1996). Coefficient of variation, COV (%) Procedure Random Totala 5c - 75d 12 - 15 14c - 100d

Test Equipment 5c - 75d

Rangeb 15 - 45

Standard penetration test (SPT) Mechanical cone penetration test 10e -15f 15e -22f 15 - 25 5 10e -15f (MT) e f e f Electric cone penetration test (ET) 3 5 5 -10 7 - 12 5 - 15 Vane shear test (VST) 5 8 10 14 10 - 20 Dilatometer test (DMT) 5 5 8 11 5 - 15 Pressuremeter test, pre-bored (PMT) 5 12 10 16 10 - 20g Self-boring pressuremeter test 8 15 8 19 15 - 25g (SBPMT) 2 2 2 0.5 a - COV(Total) = [COV(Equipment) + COV(Procedure) + COV(Random) ] b - Because of limited data and judgment involved in estimating COVs, ranges represent probable magnitudes of field test measurement error c, d - Best to worst case scenarios, respectively, for SPT e, f - Tip and side resistances, respectively, for T g - It is likely that results may differ for po, pf, and pL, but the data are insufficient to clarify this issue

3.2.3 Transformation uncertainty The third component of uncertainty is introduced when field or laboratory measurements are transformed into design parameters using empirical or other correlation models (e.g., correlating the standard penetration test N value with the undrained shear strength) as shown in Fig. 9. Obviously, the relative contribution of these components to the overall uncertainty in the design parameter depends on the site conditions, degree of equipment and procedural control, and quality of the correlation model. Therefore, geomaterial statistics that are determined from total variability analyses only can be applied to the specific set of circumstances (site conditions, measurement techniques, correlation models) for which the design parameters were derived. In other words, the COV of geomaterials cannot be viewed as an intrinsic statistical property. For each combination of soil type, measurement technique, and correlation model, the uncertainty in the design soil property is evaluated systematically by combining the appropriate component uncertainties using a simple second-moment probabilistic approach: 2

2

2

⎛ ∂T ⎞ 2 ⎛ ∂T ⎞ 2 ⎛ ∂T ⎞ 2 s ξ2d ≈ ⎜ ⎟ s w + ⎜ ⎟ se + ⎜ ⎟ sε ⎝ ∂w ⎠ ⎝ ∂e ⎠ ⎝ ∂ε ⎠

(6)

in which ξd = T(ξm, ε), T(⋅) = correlation function between test measurement (ξm) and design parameter (ξd), ε = transformation uncertainty, w = inherent variability, e = measurement error, and s2 = variance.

13

Figure 9. Transformation uncertainty (Phoon & Kulhawy 1999b). The above equation refers to the variance of the design parameter at a point in the soil mass. For foundation design, it is not uncommon to evaluate the spatial average of the design parameter over some depth interval, rather than using the value of the design parameter at a point. The spatial average of ξd is defined as:

ξa =

1 La

zb

∫z

ξ d (z) dz

(7)

t

in which ξa = spatial average, zt and zb = top and bottom coordinates of a depth interval, respectively, and La = zb - zt = averaging length. If t and ∂T/∂w are constants, it can be shown that the variance of the spatial average (sξa2) is given by (Phoon & Kulhawy 1999b): 2

2

2

⎛ ∂T ⎞ 2 ⎛ ∂T ⎞ 2 ⎛ ∂T ⎞ 2 2 s ξ2a ≈ ⎜ ⎟ Γ (L a )s w + ⎜ ⎟ s e + ⎜ ⎟ s ε ⎝ ∂w ⎠ ⎝ ∂e ⎠ ⎝ ∂ε ⎠

(8)

in which Γ2(⋅) = variance reduction function, which depends on the length of the averaging interval, La. The following approximate variance reduction function is proposed for practical application (Vanmarcke 1983): Γ2(La) = 1 Γ2(La) = δv/La

for La ≤ δv for La > δv

in which δv = vertical scale of fluctuation. It can be seen that the variance reduction function decreases as the length of the averaging interval increases. Therefore, the effect of averaging is to reduce the uncertainty associated with inherent variability (sw2). The scale of fluctuation quantifies the spatial extension in which a property of interest can be considered significantly autocorrelated. Within separation distances smaller than the scale of fluctuation, the deviations from the trend function are expected to show relatively strong correlation. When the separation distance between two sample points exceeds the scale of fluctuation, it can be assumed that little correlation exists between the fluctuations in the measurements. From a geological and geomorphological perspective, it is intuitive that soil formation and modification processes, as well as factors contributing to the definition of the

14

in-situ state (e.g. stress) would result in a greater heterogeneity of soil properties in the vertical direction and, hence, in a weaker spatial correlation. Hence, the scale of fluctuation of a given soil property in the vertical direction is generally much smaller than in the horizontal direction. The scale of fluctuation is not an inherent property of a soil parameter. Numerical values of the scale of fluctuation depend at least on: (a) the spatial direction [e.g. horizontal, vertical]; (b) the measurement interval in the source data; (c) the type of trend which is removed during decomposition; (d) the method of estimation of the scale of fluctuation from residuals; and (e) modelling options from the specific estimation method. Uzielli et al. (2006) discussed these issues in detail. Vanmarcke (1977) proposed the following approximate relationship for evaluating the scale of fluctuation:

δ≈

2 Δ π

(9)

where

1 Δ= nc

nc

∑Δ i =1

(10)

i

is the average distance between the intersections of the fluctuating component and the trend of a given profile (see Fig.10).

Figure 10. Simplified estimation of the scale of fluctuation as proposed by Vanmarcke (1983)

Other methods to estimate the scale of fluctuation have been implemented in the geotechnical literature. Uzielli et al. (2006) provided an overview of such methods. Uzielli (2004) compared the estimates of the scale of fluctuation of profiles of normalized cone tip resistance using different estimation methods. None of the models were shown to rank consistently above or below the others, though however, in some cases the scatter among

15

estimates from different methods is significant. Ranges of in Table 6.

Table 6. Literature values for the horizontal ( parameters (from Uzielli et al. 2006) Property* Soil type su clay su clay sand, clay qc qc offshore soils 1/qc alluvial deposits clay qt qc1N cohesive-behaviour soils qc1N intermediate-behaviour soils cohesionless-behaviour soils qc1N fs sand fs deltaic soils cohesive-behaviour soils FR FR intermediate-behaviour soils FR cohesionless-behaviour soils Ic cohesive-behaviour soils Ic intermediate-behaviour soils Ic cohesionless-behaviour soils N sand w clay, loam wL clay, loam ´ clay clay, loam e organic silty clay ´p organic silty clay KS dry sand fill ln(DR) sand n sand *

**

h)

v

for a number of geotechnical parameters are given

and vertical (

v)

scale of fluctuation of geotechnical

Testing method** lab. testing VST T T T TU T T T T T T T T T T T SPT lab. testing lab. testing lab. testing lab. testing lab. testing lab. testing PLT SPT -

h (m) 46.0-60.0 3.0-80.0 14-38 23.0-66.0 170.0 180.0 0.3 67.0 3.3

v (m) 0.8-8.6 2.0-6.2 0.1-3.0 0.3-0.4 0.1-2.6 0.2-0.5 0.1-0.6 0.3-1.0 0.4-1.1 1.3 0.3-0.4 0.1-0.5 0.1-0.6 0.2-0.6 0.2-0.5 0.6 0.3-1.2 2.4 1.6-12.7 1.6-8.7 1.6 2.4-7.9 3.0 0.6 3.7 6.5

su=undrained shear strength; qc=cone tip resistance; qt=corrected cone tip resistance; qc1N=dimensionless, stress-normalised cone tip resistance; fs=sleeve friction; FR=stressnormalised friction ratio; Ic=T soil behaviour classification index; N=SPT blow count; w=water content; wL=liquid limit; ´=submerged unit weight; =unit weight; e=void ratio; ´p=preconsolidation pressure; KS=subgrade modulus; DR=relative density; n=porosity VST=vane shear testing; T=cone penetration testing; TU=piezocone testing; SPT=standard penetration testing; PLT=plate load testing

Useful guidelines on typical “total” coefficients of variation of many common design soil strength properties have been summarized by Phoon and Kulhawy (1999b) and are given in Table 7 for reference. Once again, tabulated values should not be applied uncritically, i.e. if it has not been assessed that the variability related to site conditions, testing methods and transformation models are significantly similar to those used to obtain specific literature values. Table 7. Approximate guidelines for design soil parameter variability (Phoon & Kulhawy (1999b). Design a parameter su(UC) su(UU)

b

Test

Direct (lab) Direct (lab)

Point COV (%) 20-55 10-35

Soil type Clay Clay

16

Spatial avg. c COV (%) 10-40 7-25

Correlation Equationf -

su(CIUC) Direct (lab) Clay 20-45 10-30 su(field) VST Clay 15-50 15-50 e e su(UU) qT Clay 30-35 30-40 e e su(CIUC) qT Clay 35-40 35-50 su(UU) N Clay 40-60 40-55 d KD su Clay 30-55 30-55 e su(field) PI Clay 30-55 Direct (lab) Clay, sand 7-20 6-20 φ e e qT Sand φ (TC) 10 10-15 e e φcv PI Clay 15-20 15-20 Ko Direct (SBPMT) Clay 20-45 15-45 Ko Direct (SBPMT) Sand 25-55 20-55 e e Ko KD Clay 35-50 35-50 e Ko N Clay 40-75 EPMT Direct (PMT) Sand 20-70 15-70 ED Direct (DMT) Sand 15-70 10-70 EPMT N Clay 85-95 85-95 ED N Silt 40-60 35-55 a - su = undrained shear strength; UU = unconsolidated-undrained triaxial compression test; UC = unconfined compression test; CIUC = consolidated isotropic undrained triaxial compression test; su(field) = corrected su from vane shear test; φ = effective stress friction angle; TC = triaxial compression; φcv = constant volume φ; Ko = in-situ horizontal stress coefficient; EPMT = pressuremeter modulus; ED = dilatometer modulus b - VST = vane shear test; qT = corrected cone tip resistance; N = standard penetration test blow count; KD = dilatometer horizontal stress index; PI = plasticity index c - averaging over 5 meters d - mixture of su from UU, UC, and VST e - COV is a function of the mean; refer to COV equations in Phoon & Kulhawy (1999b) for details f - Quality of correlation affects the COV of design parameters. The correlation models used are referenced by the equation numbers in Phoon & Kulhawy (1999b)

14 18 18 23 29 32 38 43 49 54 61 64

3.3 Model uncertainties 3.3.1. Ultimate limit state Robust model statistics can only be evaluated using: (1) realistically large scale prototype tests, (2) a sufficiently large and representative database, and (3) reasonably high quality testing where extraneous uncertainties are well controlled. It is common to correct for simplifications in the calculation model using the following multiplicative form (as exemplified by laterally loaded drilled shafts): Hm = M·Hu

(11)

in which Hm = “measured” lateral capacity (more precisely, capacity interpreted from load test), Hu = ultimate lateral capacity computed using limit equilibrium analysis, and M = model factor, typically assumed to be an independent log-normal random variable. It is well known that many different models exist for the computation of Hu. The distributions of the model factors for Hm determined using 2 different criteria (HL or Hh) and Hu computed from 4 different lateral soil stress models are shown in Fig. 11. Note that M < 1 implies that the calculated capacity is larger than the measured capacity, which is unconservative. If Hm is defined as the hyperbolic capacity (Hh), M < 1 is most likely unsafe as well since there is no reserved capacity beyond Hh and it is mobilized at very large displacements.

Capacity model (soil type)

Lateral or moment limit (HL)

17

Hyperbolic capacity (Hh)

Mean = S.D. = COV = n= pAD =

0.3 0.2

0.92 0.27 0.29 72 0.633

0.1 0

0.3 0.2

3.2


1.49 0.57 0.38 72 0.122

0.1

Relative Frequency

0.8 1.6 2.4 HL/Hu(Broms)

3.2

0.4

0.2

0.3 0.2


0.85 0.24 0.28 72 0.555

0.1 0

0.1

0.4 0.3 0.2

0.4


2.28 0.85 0.37 74 0.149

1.2 2.0 2.8 Hh/Hu(Broms)

3.6


1.32 0.38 0.29 74 0.270

0.4 0.3 0.2 0.1


0.88 0.36 0.41 75 0.736

0.1 0 0.0 0.8 1.6 2.4 3.2 H L/H u(simplified Broms)

0.4 1.2 2.0 2.8 3.6 H h/H u(Randolph & Houlsby)

Relative Frequency

Relative Frequency

0.2

3.6

0

0.4 0.3

1.2 2.0 2.8 H h/Hu(Reese)

0.1

0.0 0.8 1.6 2.4 3.2 H L/H u(Randolph & Houlsby)

Broms (1964b) (sand)

1.42 0.41 0.29 74 0.186

0 0.0

0.3

Mean = S.D. = COV = n= p AD =

0.4 Relative Frequency

Relative Frequency

0.4

0.8 1.6 2.4 HL/Hu(Reese)

0

Randolph & Houlsby (1984) (clay)

0.4

0 0.0

Broms (1964a) (clay)

Relative Frequency

0.4

Relative Frequency

Relative Frequency

Reese (1958) (clay)

0.4 0.3 0.2


1.30 0.49 0.38 77 0.141

0.1 0 0.4 1.2 2.0 2.8 3.6 Hh/Hu(simplified Broms)

Figure 11. Distribution of model factors (Phoon & Kulhawy 2005).

3.3.2. Serviceability limit state Phoon et al. (2006) reported a probabilistic characterization of load-displacement curves using an augered castin-place (ACIP) pile load test database. The normalized hyperbolic curve considered in their study is expressed as:

18

Q Q STC

=

y a + by

(12)

in which Q = applied load, QSTC = failure load interpreted using the slope tangent method, “a” and “b” = curvefitting parameters, and y = pile butt displacement. Note that the curve-fitting parameters are physically meaningful – the reciprocals of “a” and “b” are equal to the initial slope and asymptotic value, respectively. Each continuous load-displacement curve can be reduced to two curve-fitting parameters (plotted as a single point in Fig. 12). The scatter in the load-displacement curves is captured by the scatter between “a” and “b”. If the values of “a” are plotted as a histogram, a non-uniform distribution will be obtained (Fig. 13). For example, values close to 5 mm are more frequently encountered. The standard approach is to fit one of the many classical distributions (e.g. lognormal distribution) to the histogram. The crucial point to be emphasized here is that such an approach implicitly assumes that “a” and “b” are statistically independent random variables. However, Fig. 12 clearly shows that the variation of “a” (variation along y-axis) and the variation of “b” (variation along xaxis) are coupled. In other words, it is incorrect to assume that “a” can vary independently of “b”. The correct probabilistic model in this case is not independent random variables but a bivariate random vector. Computational details on the construction of this bivariate random vector are given elsewhere (Phoon and Kulhawy, 2008).

20 a parameter (mm)

a parameter (mm)

Spread foundation (uplift)

Augered cast-in-place pile (compression)

20 16 12 8 4

16 12 8 4 0

0 0.2

0.4

0.6

0.8

1.0

0.2

1.2

b

No. of tests = Mean = Standard deviation = Coefficient of variation = Mean = Standard deviation = Coefficient of variation = Correlation =

0.4

0.6

0.8

1.0

1.2

b parameter

b parameter

a

clay sand

40 5.15 mm 3.07 mm 0.60 0.62 0.16 0.26

a

b

-0.67

No. of tests = Mean = Standard deviation = Coefficient of variation = Mean = Standard deviation = Coefficient of variation = Correlation =

19

85 7.13 mm 4.66 mm 0.65 0.75 0.14 0.18 -0.24

Pressure-injected footing (uplift)

Drilled shaft (uplift)

5

clay sand

4

a parameter (mm)

a parameter (mm)

5

3 2 1

4 3 2 1 0

0 0.4

0.6

0.8

1.0

1.2

0.4

1.4

0.6

No. of tests = Mean = Standard deviation = Coefficient of variation = Mean = Standard deviation = Coefficient of variation =

b

Correlation =

1.0

1.2

1.4

b parameter

b parameter

a

0.8

48 1.34 mm 0.73 mm 0.54 0.89 0.063 0.07

No. of tests = Mean = Standard deviation = Coefficient of variation = Mean = Standard deviation = Coefficient of variation =

a

b

-0.59

Correlation =

25 1.38 mm 0.95 mm 0.68 0.77 0.21 0.27 -0.73

0.3


0.2

5.15 mm 3.07 mm 0.60 40 0.183

Relative Frequency

Relative Frequency

Figure 12. Correlation between hyperbolic parameters (Phoon et al. 2006; 2007)

0.1 0

0.3


0.2

0.62 0.16 0.26 40 0.887

0.1 0

0

5 10 15 a parameter (mm)

20

0.0

0.5 1.0 1.5 b parameter

2.0

Figure 13. Marginal distributions of hyperbolic parameters (Phoon et al. 2006).

4. OTHER POTENTIAL ROLES OF RISK-BASED ANALYSIS One of the important requirements in forensic geotechnical engineering is the identification of failure mechanisms. A possible role for risk-based analysis can be discerned from the simulation studies detailed below. Popescu et. al (1997) studied the effects of spatial variability of soil properties on soil liquefaction for a saturated soil deposit subjected to seismic excitation. They compared standard results derived from deterministic inputs and probabilistic results derived from stochastic finite element analyses. They concluded that both the pattern and the amount of dynamically induced pore water pressure buildup are strongly influenced by the spatial variability of soil parameters. For the same average values of soil parameters, more pore water

20

pressure build up is predicted by the stochastic model than by the deterministic model, which is attributed to a water injection phenomenon triggered by the presence of loose pockets in the spatially variable soil deposit. Griffiths et. al (2002) demonstrated that failure mechanisms of a strip footing are considerably different in the case of uniform soils and spatially variable soils. The spatial correlation structure of the foundation soil in of autocorrelation structure has an important effect on the failure mechanism. More extensive studies are presented by Kim (2005). Goldsworthy (2007) presented a fairly comprehensive risk-based study of the effect of site investigations on foundation failures. One conclusion is that an optimum site investigation minimizing the risk of failure can be planned based on the variability and autocorrelation structure of the foundation soil. The above studies and other similar efforts demonstrate that failures can be studied in a more realistic way using spatially variable soils, rather than the traditional uniform or layered soil profiles. Notwithstanding this, a significant number of failures does arise from human failures and a lack of proper quality assurance and control plans. Bea (2006) suggested that the failure development process can be categorized into three phases viz, initiating, contributing and propagating causes and a proper risk management strategy is necessary using observational methods. A risk-based approach to forensic geotechnical engineering may include elements such as: i.

Detailed soil investigations in the area in the form of vertical and horizontal soil profiles to formulate plausible hypotheses concerning prevalent/relevant failure mechanisms. The number and spatial location of boreholes and spacing should be adequate to provide a proper estimate of the mean, variance and autocorrelation properties of in-situ soils.

ii. Analysis of the actual failure mechanism in relation to results from probabilistic studies. iii. Reanalysis of different loading and the associated variability. iv. Re-examination of quality assurance plans which include specifications on quality control of materials and construction, construction sequences proposed and adopted, observations of performance depending on the nature of project such as deformations, pore pressures etc. which are expected in important projects; and v.

Reliability-based computational back-analysis.

5. CONCLUDING REMARKS This chapter seeks to explore if reliability/risk concepts are potentially useful to forensic geotechnical engineering. This study is very preliminary given the dearth of previous literature. One key aspect that distinguishes geotechnical engineering from structural engineering is the natural variability of geo-materials. Within this context, Leonards (1982) definition of “failure” – “unacceptable difference between expected and observed performance” – cannot be evaluated in a meaningful way using deterministic methods. In broad , forensic engineering is related to the investigation of failures with the view of rendering an opinion regarding responsibility. Hence, a statistical measure of “unacceptable difference” (specifically, a difference not explainable by underlying variability) should provide useful additional information in the formulation of such an opinion. A preliminary statistical framework is presented to quantify the difference between expected and observed performance in the presence of unavoidable and potentially significant geotechnical variabilities. It should be emphasized that the quality of the statistical analysis can only be as precise, accurate and meaningful as the engineer’s characterization of uncertainties. Other potentially useful results in the recent reliability and risk literature are highlighted. The intention of this chapter is to stimulate further discussions and research in this important but somewhat overlooked area. REFERENCES Bea, R. (2006). Reliability and human factors in geotechnical engineering. Journal of Geotechnical and Geoenvironmental Engineering, ASCE, 132(5): 631-643.

21

Brown E. T. (2006). Forensic engineering for underground construction. Proceedings of the ISRM International Symposium 2006 and the 4th Asian Rock Mechanics Symposium, Singapore 8 - 10 November 2006, Chapter 1, 1-16. Chilès, J-P. and Delfiner, P. (1999). Geostatistics – Modeling Spatial Uncertainty. John Wiley and Sons, New York. Duncan, J.M. (2000). Factors of safety and reliability in geotechnical engineering, Journal of Geotechnical and Geoenvironmental Engineering, ASCE, Vol.126, No.4, pp. 307-316. Goldsworthy, J. S. (2007). Quantifying the risk of geotechnical site investigations. PhD Thesis, University of Adelaide. Griffiths, D V, Fenton, G. A and Manoharan, N. (2002). Bearing capacity of rough rigid strip footing on cohesive soil: probabilistic study. Journal of Geotechnical and Geoenvironmental Engineering, Vol. 128(9), 73-755. Jones, A.L., Kramer, S.L. & Arduino, P. (2002). Estimation of uncertainty in geotechnical properties for performance-based earthquake engineering. PEER Report 2002/16, Pacific Earthquake Engineering Research Center, University of California, Berkeley. Kim, H. (2005). Spatial variability in soils: stiffness and strength, PhD Thesis, Georgia Institute of Technology, 2005. Kulhawy, F.H. & Trautmann, C.H. (1996). Estimation of in-situ test uncertainty. Uncertainty in the Geologic Environment - From Theory to Practice (GSP 58), ASCE, New York, 269 – 286. Kulhawy, F.H., Phoon, K.K. & Prakoso, W.A. (2000). Uncertainty in the basic properties of natural geomaterials. Proc. 1st International Conference on Geotechnical Engineering Education and Training, Sinaia, Romania, 297-302. Leonards, G. A. (1982). Investigation of failures. Journal of the Geotechnical Engineering Division, ASCE, 108(GT2): 187-246. Paikowsky, S.G. (2002). Load and resistance factor design (LRFD) for deep foundations. Proc. International Workshop on Foundation Design Codes and Soil Investigation in view of International Harmonization and Performance Based Design, Tokyo, Japan, 59 – 94. Balkema, Netherlands. Phoon, K.K, Kulhawy, F. H & Grigoriu, M. D. (1995). RBD of Foundations for Transmission Line Structures. Report TR-105000, Electric Power Research Institute (EPRI), Palo Alto. Phoon, K. K. & Kulhawy, F. H. (1999a). Characterization of geotechnical variability. Canadian Geotechnical Journal, 36(4):612-624. Phoon, K. K. & Kulhawy, F. H. (1999b). Evaluation of geotechnical property variability. Canadian Geotechnical Journal, 36(4):625-639. Phoon K.K. and Kulhawy, F.H. (2005). Characterization of model uncertainties for laterally loaded rigid drilled shafts. Geotechnique, 55(1), 45-54. Phoon, K. K. & Kulhawy, F. H. (2008), Serviceability limit state reliability-based design, Chapter 9, ReliabilityBased Design in Geotechnical Engineering: Computations and Applications, Taylor & Francis, April 2008, 344-383. Phoon, K.K., Chen, J.-R. & Kulhawy, F.H. (2006). Characterization of model uncertainties for augered cast-inplace (ACIP) piles under axial compression. Foundation Analysis & Design: Innovative Methods (GSP 153), ASCE, Reston, 82-89

22

Phoon, K. K., Chen, J. R. & Kulhawy, F. H. (2007), Probabilistic hyperbolic models for foundation uplift movement, Probabilistic Applications in Geotechnical Engineering (GSP 170), ASCE, Reston, CDROM. Popescu, R. Prevost, J. H and Deodatis, G. (1997). Effects of spatial variability on soil liquefaction: some design recommendations. Geotechnique, 47(5):1019-1036. Rétháti, L. (1988). Probabilistic solutions in geotechnics. Elsevier, New York. Silva, F., Lambe, T. W. and Marr, W. A. (2008), Probability and risk of slope failure. Journal of Geotechnical and Geoenvironmental Engineering, ASCE, 134(12): 1691-1699. Sowers, G. F. (1993). Human factors in civil and geotechnical engineering failures. Journal of Geotechnical Engineering, ASCE, 119(2): 238-256. Uzielli, M. (2004). Variability of stress-normalized T measurements and application to seismic liquefaction initiation assessment, University of Florence (Italy). Thesis able at: http://www.georisk.eu. Uzielli, M., Lacasse, S., Nadim, F. & Phoon, K.K. (2006). Soil variability analysis for geotechnical practice. Proc., Second International Workshop on Characterisation and Engineering Properties of Natural Soils, Singapore. Balkema, Netherlands, 1653 – 1752. Vanmarcke, E.H. (1983). Random Field: Analysis and Synthesis, MIT Press, Cambridge, Massachusetts.

23

-1-

Case Studies 1

TURNING HINDSIGHT INTO FORESIGHT (LEADERSHIP LESSONS FROM A LANDSLIDE) Mike Marley BE, MEngSC, FIEAust, RPEQ, ACIS Principal – Golder Associates Pty Ltd 611 Coronation Drive, Toowong, Qld [email protected]

INTRODUCTION Tom Peters (one of the great business thinkers of our time) proclaims in his “Leadership Essentials” that “great leaders are great story tellers”. He states” “We need stories, Riveting Tales that fire the imagination of ---- as-yet-reluctant heroes-inwaiting”. This is the story of a forensic engineering study to determine the cause of a major landslide in P.N.G., of the combination of good luck and good management that led to a successful outcome, and of the lessons the author took from the study which have informed his views on project management during the years since the study. Drawing on Peters’ “Leadership Essentials” (the ability to inspire, liberate, achieve) some lessons in leadership are highlighted, and the importance of populating the leadership literature with similar stories of real projects, as a means of recruiting and nurturing the next generation of inspirational leaders, is emphasised.

THE STORY Background The Ok Tedi gold and copper mine was developed in the remote Star Mountains in the Western Province of P.N.G., close to the Irian Jaya border (Figure 1). A tailings storage dam was included in the mine development plan, and construction of this dam was on the critical path for the mine construction programme to ensure that key milestones could be met for the project’s financial model.

T OK O OK T TE E D D K O OK OO KT EK DIK I OK TK O OK T T E ED D DIIII M M S E MI IMI S T IE IM N NE E ES E E MN IN NE SIIT IT ES TE E MIMIMIN M IM S SIIIT TE IIM

ii Ta Ta b b bu u b b iia iiiilllllgen Tb a TT ab au aM T auan T a llll T iiiT T a Ta aoT b b b u ub b b b tu H n een g aa H M M o o o un u n u Ha H age gen n n n n ntu un g H u n u n u nntttttt H Ha ge eg ee b u b bM tt n M Mo o o ou un u H Ha Ha aa ge gge egg n n Ta TT a ab bu ub

T A B U lll IBIL T A B U B bi iiiIB T T A BU BU BU IBss I ee La a L aL aL e A B U BU BU IB BrIB L L oun est B UB BU LBU La en e n aL aL aL a ae T T A B BUBIBW IIIIL W W s n ntttt L L set rW eBI nest WeW rn We estM e tee n Mo ost unuu nn urrru n tteW W We e st stn un su tee n n ttt M M o o o eu nrrrn H Ha H a a gege gen n n cree PP inrP vvP ro P Ha eg ee H a gge egg n n n orP P Pr oH iin n n n c c e e o o v viia c ePr r ro incc v rro PrP riin Pr o n nH c ca e e ge

OkkkT T O T O eee ddd ii

T BU BU BU m rru iA idi i di NNN T A A B U B U B U B U B UBI B IBIBIBIBIBIBIBIL L Luu m u n n g g e u m m m uiNi nNi gN e NB iNU u rriN Da Da Darrru iNiNin n ng g ge errrrru um m m DD aa rDu y orreessb a Da Da Da Da u urtM i di didiiiiiNi byy Po Mo orresb rrttM edd TT s Po M o esb e esb W We tte n Po o M o orrrrrresb esb eT ee WW e estestestestestests n no s P orrrrtrtrtrtttt M M Mo o esb e esb W We sP ee ePrrrrrro n IIIIIG B GE

Ok OO kO kO kO kO kk kO kO

ba y y y BIBIBIBIB IBL G E E sb IB IBL G La a L a e e b y y y LIB L aL a aIG aE L L a IIN W A N G E U E W A PAPU AN NE GU NE E U E WG A IIN PrPrPrPrPPr r Pr

KI UN G UN A A IKN KIKIKIKIUN UN N U GA KIIK IU G A IKI KI UN UN A A PU ro o o v vN iiin n nUN ccee G

Pa ap pu ua a P a p u a P

r Pr ooovviiinnncceee e e iv 0 50 100 NiN0iNiNiNiNiNiNiNin n n ge u m m g g errrrrru um ididididididDa n n ng e u u m m m R idDa Da Da u u Da killometres Da Da Da Darrrrrru u u kkkeee L e L Laaa ake ake Mu ay y P Pdo M uarke ra L ake ake ay ke ake ake ake M urrrra y ay akilometres y orrtrtrtrtt M M Mo o orr Po Po Po esb esb esb esb e esb esb sby y y n a an d TT M eT ee M daa nd Maan l l a kIBIBIBIBIBIBIG M a M an BIB G E E M a n nd d daa GE Ma Ma a n nd d da ckIB i Ok OO kO kMA O kAn O k kaaA kO kIIIO O kK M B tr AM B B AK A K K AM AM AI AIIAM SAIAIAIAIAIA AM B AK t B APU K AM AM A M B PA PA PA PA PB A PA PAK U P PA PA PAPU PU PUPU S pua pua pua pua pua pua P o W est er n P Prrro ovvviin in nccceee

W est er n

PUA A A G U U N E E A A A N N E W W PU A AN G GU U UIIIIIN N NE EA A N N NE E E EW WG a Wee ste sro ts er rn n pte W e ree aa a Papua a Pr oP v vapua iin ce ee ststsW PrPrPapua rPro nW c W ter rn nePr Pst rPW osts v v ien n c e pie a ap ac P KIKIKIKIKIKIKIKIKIUN UN UN A A UN UN UN GA UNG UN UN G A A

r e Wa s p e e WW a Wa WW a W a as sa a ap p a Wa a W s a p pe e eaa a iv 0a 50 100 a ssu W aa W uaa su F R lyW Is sllllla a a an n n nd d d d StStSS S t StStu u u rrrrrrtttttt IIIs t St s a a n n d d St t u u u R d n a l i s w a I d au aaiiiiIIs ivey L ake M M rararaa ay y Kiw d rra e b laan A t ike ske Lake a ke ke ke ake akee rd a y akilometres y d aa d nd iwd IIslla lan d n ee L e wu a d rttiiis n saIs D'''A 'Aa lb Isllla a r i la l lb Wa W W asu a su s u Wa asu asu WW a W asu asu su a aa a e A a TAPIIIL IL LA A ia iinn n u KiKiKiKiKiKiKiw aiii w wa a Le L eee w w wa ad d da abu aa r im eLw S S l IIs s sllla an nd d d a n u bu imaab w errriim l a Seew Ki K iw wa ai i I s l la a n nd d M M ae n kIs i a M a n nrrd d a Ma M M n n nd d da a a sllla a an n nd d d D'D 'A A rd ttta is is IiIs Ik D D'a 'Da A lllb b b e iis isc Suiii am aa D A R i tdr D A R U ma D A RU U Paraam u s G D D'D' A A Alllb b beerrrtttis iis s Is IIsllla a an n nd d LL e L eL eAM e Lew w w a a d d dM a a a B B AK K AM AI d n laB IIs SAIAIAIAIAIALe t s Aa M A B B AK AK IIAM B K AM AM M d n dA nAK G IIsslllaa S pua pua pua pua pua pua uaaN pu wG uii Paap New Gu P a Neew inea apu P Gu

nesiiaa nd on In do do neessiia IIn

d lan t urrtt Is S S d nd urrtt IIs Stttu lan Isllla

G

G ne ne nn eee a nn ee ne ne Tst AA TA A P IIILA L L Wee eTA rn n A A Ps IP W e st s tsv eA ree TA Pt LtA TT T Iro P apua w Pr rP oP v vw inW ce ee ststsW LA P rP rPrno c LA L Aapua W ter rn nTA Pr P rP o v iLer in n c c e Papua e in do do do dd oo dodd oo e do N N In I a a

Sew Se w b bu u u Sew Sew SS eS ee Se Sew we e errriim m maab Wa s sa a ap p e ea a a SSW S W SSs ip iie W Wa SWa

-2-

d lan r u ss d nd Darrru u IIIslla lan

Figure 1 A fast track approach was adopted, requiring investigation, and design to proceed concurrently with construction. This approach increases the overall risk associated with the work, and requires the design programme to be sufficiently flexible to accommodate design changes where investigations identify conditions significantly different from assumptions adopted in the preliminary design. The site is located in one of the most remote and challenging environments in the world: the terrain is rugged; the rainfall extreme (up to 11m per annum); and the jungle is almost impenetrable. Lesson 1:

Understand and respect the environment in which you are operating

History Preparation for construction of the tailings dam commenced in May 1983. Excavation for the dam footprint commenced in early November 1983. In mid December 1983, (when approximately 250,000m³ of excavation had been completed) a landslide occurred in the eastern side of the valley. Approximately 3.4 million m³ was involved in the slide mass. This was followed in early January 1984 by a much larger slide (approximately 35 million m³) covering an area approximately 800m wide and 1.2km long) (Figure 2).

Extent of December 1983 Sl

Extent of January 1984 Slid

Figure 2 An investigation of the slides was initiated (specifically to asses whether the footprint excavation was the primary cause, and whether there were grounds for an insurance claim under the contractors insurance). Initial Assessment Initial assessment revealed: –

significant variability in the magnitude of translational movement within the slide mass:

–

ratio of excavation volume to slide mass volume for first slide of approximately 1:15;

–

ratio of excavation volume to slide mass volume for second slide of approximately 1:154;

–

-3the disposition of the slide mass with respect to excavation was particularly asymmetrical;

–

slide had apparently occurred on a very low angle failure plane (Figure 3).

Figure 3 There was considerable doubt as to whether the excavation was the sole (or even a significant) contributor to the slide. Other potential causes canvassed at the time include earthquake and/or an intense rainfall event. Clearly assessment of the cause would not be straight forward. Assembly of a team with a wide range of geo-scientific expertise was the essential first step. Lesson 2:

Complex tasks require high performance teams

The Investigation Geologically, the Star Mountains are part of the deformation zone associated with the active boundary between the Pacific an Indo-Australian tectonic plates (Figure 4). At this boundary the sub-oceanic crust of the Pacific Plate plunges beneath the more stable continental crust of the Indo-Australian plate caulift (Figure 5). There is considerable seismic and volcanic activity associated with this plate sub-duction and the associated rapid uplift has resulted in thick layers of limestone and mudstone, formed from marine deposits, being exposed in the rugged heights of the Star Mountains. A simplified cross-section through the PNG shows the impact of this tectonic activity (Figure 5), including the high central mountain spine and zones of low amplitude folding to the north and south of the spine. This reminder of the mechanics of continental creation proved to be the first vital clue in understanding the cause of the slide.

Figure 4

-4-

Figure 5 Initial helicopter reconnaissance surveys to assess the magnitude and extent of the damage caused by the slide also highlighted an unusual feature of the Ok Ma valley (viz. that the valley was broadly leaf-shaped with a narrow, meandering entrance gorge upstream of the dam site, and another similar exit gorge downstream of the site). This feature was repeated regularly throughout the surrounding area, and this repetition of topographic form proved to be the second important clue in understanding the slide. Detailed study of the aerial photography of the site and its surrounds revealed a number of regional characteristics, in particular several sets of linear features crossing the folded topography surrounding the site (Figure 6). These lineaments (identified under the stereoscope as distinct partings in the tree cover) are indicators of geological structures such as faults, fractures and shears in the rocks underlying the area.

Figure 6 The orientation of the predominant sets of defects was noted. Lesson 3:

Take time to look at the big picture – get above the trees so you can see the form of the forest.

Review of the considerable volume of existing investigation data and records showed:

–

-5no significant earthquake or rainfall events that were likely to have initiated the slides,

–

the stratigraphic sequence at the site (Figure 7) with the upper levels of the sequence removed by erosion exposing the Warre limestone as a relatively thin cap overlying approximately 1000m of dark fine grained mudstone (Pnyang formation) in which the slide occurred.

–

failure could not be readily explained based on measured strength in mudstone.

Figure 7 Field investigations concentrated on detailed geological mapping of the site, and a drilling and instrumentation installation programme. A number of inclinometers and piezometers were installed throughout the slide mass to record any on-going movement of the slide mass, as well as details of fluctuations in groundwater levels which have a major influence on slope stability. Airphoto interpretation and field mapping yielded valuable understanding of the regional geomorphology. The leaf-shaped valleys were found to be the results of transverse erosion through the anticlinal folds formed by the compression and uplift caused by regional tectonic forces. A mechanism was deduced for the valley formation processes as shown in Figure 8: –

Streams originating in overlying formations (since removed)erode down through the upper moderately resistant limestone cap.

–

As uplift and erosion continue, the streams enter the underlying less resistant mudstone.

–

Progressive erosion of the mudstone and removal of the toe of the valley slope leads to undercutting and breaking up of the limestone cap.

–

Evolution of the slope by repetitive slides in the mudstone creates weak zones approximately parallel to the bedding with very much lower strength than the surrounding materials.

-6-

Figure 8 The establishment of the mechanism of formation of the characteristic leaf shaped valley (Figure 9) was the next critical factor in understanding the cause of the slide.

-7-

Figure 9 With a widely scattered team, daily debriefings were held to enable the whole team to stay in touch with progress of the various facets of investigation work, each proceeding at its own characteristic pace. Lesson 4:

Good communication is critical to the effective functioning of high performance teams.

Geological mapping of the immediate area surrounding the slide revealed a closely spaced grid of near vertical ts in the mudstone (Figure 10) and the orientations of these defects were found to align closely with the large-scale lineations observed on the aerial photography.

Figure 10 The orientations of these lineations at both macro and micro scale aligned closely with the theoretically calculated directions of failure planes which would be induced in the mudstone by the tectonic compression forces causing the uplift and folding at the site. Laboratory scale triaxial compression tests on mudstone specimens confirmed this. All of the major defect sets identified at the site were now able to be explained. Figure 11 shows the defects: –

bedding planes originally horizontal, but now distorted by folding,

–

axial defects parallel to the fold axis (major tensile zone),

–

conjugate t sets induced by tectonic compression.

-8-

Figure 11 Figure 12 shows a diagrammetric representation of a typical wedge-shaped block of mudstone which would be produced as a result of the intersection of the various defects sets at the site.

Figure 12 Instrumentation records showed that movement was continuing in the slide mass and that the movement appeared to be occurring on multiple planes stepping up in elevation progressively in the downstream direction. Measured values of shear strength on samples recovered from the vicinity of the shear planes identified in the inclinometers confirmed that actual strengths on these planes (formed initially in the valley forming process) was about one third of the average strength of intact mudstone. These measurements confirmed values of shear strength estimated from back analysis of the stability of the slope to be necessary for slides to be mobilised on planes at the locations indicated by inclinometers. They also confirmed that use of the average values of mudstone strength (as had apparently been done by the original designers) would seriously under-estimate the likelihood of instability. Lesson 6:

Averages are statistical constructs. Understand the numbers that went into their calculation.

Piecing all of the data together allowed a model for the slide mechanism to be developed involving uniform planar movement of a relatively few large wedge-shaped blocks of mudstone. Figure 13 shows a schematic representation of the site and the three dimensional spatial relationships between the slide blocks and the terrain surface. Initial failure of a relatively small block at the upstream end following removal of at the toe of the slope by the footprint excavation in turn removed from a succession of further blocks (with different failure plane elevations) resulting in progressive enlargement of the failure downstream.

-9-

Figure 13 This model actually recreated the form of the landslide remarkably closely. Lesson 7:

Sketch what you think is happening. It allows a better understanding of complex relationship.

LESSONS The lessons highlighted during this investigation while significant in their own way, and useful in informing the author’s approach to project management since, have greater real value than simply providing guidance for the future solving of similar technical problems. While technical papers are an essential tool in the advancement of the science that underpins our professional practice, stories of the execution of major iconic projects are considered to be as important (or more important) in the advancement of the art of management and leadership in our profession. Stories can identify heroes who have executed iconic projects, point to these pioneers, and provide a real-life point of reference for the next generation of aspiring leaders as a guide to how it is done. According to Peters, the Essentials of Leadership are to : •

INSPIRE

(enthusiastic followers);

•

LIBERATE

(team ’ imagination and ion);

•

ACHIEVE

(outstanding results through others).

In these the more profound learnings from the study can be set out as follows: •

INSPIRE

−

Leadership is not a matter of doing excellence------ it is a matter of inspiring excellence in others.

−

Recognise your dependence on the effective functioning of a high performance team.

−

Honour those team whose perspective may appear unconventional or off-beat.

(In this study, listening to an initially improbable notion that plate tectonics was a root cause of the slide helped inspire a cascade of lateral thinking breakthroughs). • LIBERATE −

Create opportunities: encourage the team to apply their talents and grasp these opportunities;

−

“Forget command and control; Forget knowing one’s place; Forget hierarchy”

- 10 (In this study, liberating the team from the constraints of adhering solely to conventional quantitative engineering analysis was a key to the successful modelling of the slide mechanism). • ACHIEVE −

Honour those who try to achieve something rather than those who try to stop something wrong being done;

−

Achieving outstanding results requires ion, persistence and imagination

(In this study persistence in pursuing the enigma of failure on planes apparently too shallow to allow movement assuming average material strengths, allowed the achievement of a quantitatively able slide mechanism). THE IMPORTANCE OF STORYTELLING As Peters attests: “A scintillating story makes an abstract strategy real.” On completion of a WOW project (according to Peters; one that matters; one that makes a difference; one that transforms the organisation; one that it is worth bragging about) write up the story for inclusion in the library of leadership literature. And then, recount the story (often) to help inspire and liberate aspiring leaders to achieve their potential. REFERENCES

1) Hollingswoth D’Appolonia, July 1984 “Ok Tedi Project, Ok Ma Tailings Dam Site, Evaluation of Landslide”. Report to Ok Tedi Mining Ltd. 2) Griffiths, J.N. Hutchinson, D. Brunsden, D.J. Petley and P.G. Fookes; “The Reactivation of a Landslide During the OK Ma Tailing Dam, Papua New Guinea”, J.S. Quaterly Journal of Engineering Geology and Hydrogeology Vol. 37, 2004 pp 173-186. 3)

Tom Peters “Tom Peters Essentials, Leadership”, DK Publishing, 2005. U:\MIKE MARLEY\TURNING HINDSIGHT INTO FORESHIGHT-REV0.DOC

Case studies 2

Forensic Studies for Failure in Construction of An Underground Station of the Kaohsiung MRT System Richard N. Hwang 11F, 3, Dunhwa S. Road, Sec. 1, Taipei, Taiwan 10557 [email protected]

Kenji Ishihara Chou University, Japan [email protected]

Wei F. Lee Taiwan Construction Research Institute, Taipei, Taiwan [email protected]

Abstract: Several buildings collapsed as a result of leakage on diaphragm wall during the excavation for constructing O1 Station of the Kaohsiung MRT System. Resistivity image profiling was carried out to check the quality of diaphragm walls and the effectiveness of ground treatment using C. In addition, undisturbed soil samples were taken by using GP-75S sampler and laboratory tests were conducted for determining the characteristics of soils. This paper presents the findings of these studies. 1

THE PROJECT

The construction for Kaohsiung MRT System (KMRTS) was commenced in October 2001. The system comprises 2 lines, i.e., the Red Line in the NS direction and the Orange Line in the EW direction with a total of 37 stations and 3 depots and is expected to be open for revenue services at the end of year 2007. The station of interest, i.e., O1 Station (Sizihwan Station), is a terminal station at the western end of the Orange Line. It is located on the north bank of the Kaohsiung Harbour as depicted in Fig. 1. Fig. 2 shows the layout of O1 Station which is 287m in length and 16m in width and is connected to a section of tunnel at its eastern end. 2

was successfully completed and sub-structure erected in June 2006. 3

GROUND CONDITIONS

Fig. 4 shows the soil profile obtained at Borehole WB-11. Although the thicknesses of sublayers vary from place to place, the sequence shown in the figure is quite typical. Soil properties adopted in designs are listed in Table 1.

THE EXCAVATION

iver Ai R

This 2-level station with side platforms is buried at a depth of 5.12 below ground surface. Excavation for the station was carried out to a depth of 20m by using the cut-and-cover method of construction. Diaphragm walls of 800mm in thickness were installed to a depth of 39m in Zones A and B and walls of 1000mm in thickness were installed to a depth of 37.5m in Zones C and D. The pit was propped by six levels of struts as depicted in Fig. 3. On 9 August 2004, a sinkhole was formed behind S60M on the south side when excavation already reached a depth of 15m and 4 buildings (3-story) collapsed due to leakage of the diaphragm wall. Excavation was halted and measures were taken to improve the watertightness of diaphragm walls. Excavation in Zone B was resumed in July 2005 but a sinkhole was formed shortly at the same location for the second time. Additional ground improvement work was carried out and excavation resumed at the end of November 2005. The excavation

Kao

Fig. 1

hsiu n

gH

arb o

ur

Location of O1 Station, Kaohsiung MRT System

1

Basic soil properties adopted in design of diaphragm

Unit weight KN/m3 20.1 18.9 20.2 19.5 19.5

SF CL SM CL CL

Table 2

Φ’ degrees 28 27 32 29 28

C’ kPa 0 0 0 0 0

Su kPa

determined in major projects to check whether groundwater is aggressive. Table 2 shows the results of chemical tests on soils and Table 3 shows the results of tests on groundwater in boreholes in the section of route between O1 and O2 Stations. The aggressiveness of chemical attack is classified as “severe” in accordance with Table 4 which ….(ref. )

18 N

Table 1 walls

85 113

Leakage

Results of chemical tests on soils

Hole

Depth, m

PH

Organic

OA-1 OA-3 OA-4 OS-1 OS-2

20 13 15 16 16

7.7 9.5 8.3 8.3 8.3

7.38 3.96 5.29 1.8 1.8

Cl(%) 0.053 0.038 0.11 0.049 0.049

Entrance B

SO4-2 SO3-2 (mg/kg) (mg/kg) 2176 5.1 99.6 11.9 324.7 7.5 341.4 159 323.3 2.6

Entrance A

Hole OA-1 OA-3 OS-1 OS-2

Results of chemical tests on groundwater Depth m 15 15 17 15

Table 4 (ref?) mild moderate severe very severe

PH 7.9 7.2 6.8 7.7

Cl(mg/l) 180 5030 15800 17300

SO4-2 (mg/l) 112 680 2450 2450

SO3-2 (mg/l) <2.5 <2.5 <2.5 <2.5

Ca+2 (mg/l) 59.7 113 984 1030

Zone B

Zone C

Zone D

Fig. 2 Layout of O1 Station, Kaohsiung MRT System Ca+2 (mg/l) 80.3 2480 5530 4780

Classification of aggressiveness of groundwater Cl- (mg/l) 0 ~ 1,000 1,000 ~ 5,000 > 5,000

Cut-and-cover Tunnel

O1 Station Zone A

Table 3

4 buildings (3-story) collapsed

SO4-2 /SO3-2 (mg/l) 0 ~ 150 150 ~ 1,500 1,500 ~ 10,000 > 10,000

4

THE EVENT OF 9 AUGUST 2004

Water spurted from the bottom of the excavation in front of S60M, refer to Fig. 2 for location, on the southern wall at 13:20 on 9 August 2004 as excavation reached a depth of 15m. A sinkhole of 3m maximum depth was formed behind the diaphragm wall and the area affected was around 500m2. Four 4story buildings collapsed in less than an hour and were demolished overnight. Several low-rise shops were severely damaged and were demolished sometime later. 15.78m Ground Level

0 1

The subsoils on west coast of Kaohsiung City are young sediments with high water contents and low plasticity. Such soils can easily be softened once disturbed, or liquefied when subjected to steep hydraulic gradients and, therefore, failures of trenches were quite common during installation of diaphragm walls. In many cases, mini-piles and/or micro-piles were installed to prevent trench from collapsing. Even so, necking still frequently occurred and reduced the sectional areas of diaphragm walls. One row of C piles was indeed used along the perimeter of the area to be excavated for maintaining the stability of trenches prior to the installation of the diaphragm walls. In some cases, the soil pockets were continuous from back to back of the walls and became water paths. Ground treatment had to be applied behind defective diaphragm walls to stop the ingress of groundwater into pits. The site was a salt pan a century ago. It was later used for fish farming for decades. Groundwater table was much lowered as fish farmers relied groundwater as supply because surface water was contaminated. Ingression of sea water increased the high chloride content in groundwater. It has been suspected that the quality of diaphragm walls may deteriorate as a result of chloride attack and, therefore, chemical contents are routinely

2

GL-5.12m

2

3

4 5

6

GL-19.85m

Diaphragm Wall Zones A & B t = 800mm L = 39m Zones C & D t = 1000mm L = 37.5m

Fig. 3 Excavation scheme and retaining system

JSG piles 1200mm diameter at 1m spacing GL-12m to GL-29m

Bored piles 800mm diameter at 1m spacing GL-1m to GL-30m

1,000m2.Subsoils at this site, being closer to the sea, are even worse than those at O2 Station so the incident was not a surprise. 6

S58M 1

2

3

4

FORENSIC STUDIES

S60M

5

JSG Piles

JSG piles 1200mm diameter at 1m spacing GL-12m to GL-29m

(#1, 3 and 5 were in place #4 was being installed)

Plan

Subsequent to the event of 9 August 2004, the contractor engaged Kaohsiung Professional Civil Engineers Association, Kaohsiung Professional Geotechnical Engineers Association and Kaohsiung Architects Association on 16 August to form a t committee to investigate the causes of failures. After the event of 7 July 2005, Taiwan Construction Research Institute (TCRI) was engaged by the contractor in December 2005 to further investigate the causes of failures.

6.1 Investigation by the Professional Associations CL

JSG Piles

2 <3 SM

GL –12m

GL –15m 10 <28

CL/ML 20 <27

CL

GL –15m

Defective between GL-15.95m to GL-16.55 GL –29m

Diaphragm Wall t = 800mm to GL-39m

JSG Piles

Section Fig. 4 The event of 7 July 2005

One row of 11 bored piles was installed behind s S58M to S60M. Pumping tests were performed in November 2004 to see if there were other defective s. A total of 3,285 cubic meters of water was drawn from 6 wells and water levels inside the excavation closely monitored at 60 wells. The groundwater table inside the excavation dropped; on an average of 2.2m as a result of pumping. The recovery of water levels was monitored for more than 10 days, however, the desired purpose was not achieved as the locations of leakages could not be identified (Ho, et al., 2007). 5

THE EVENT OF 7 JULY 2005

To be on the safe side, one row of JSG piles was added along the entire perimeter of the station. Furthermore, the ts between JSG piles were treated by using C piles. Pumping tests were again performed subsequently to confirm the effectiveness of these measures. The results were not satisfactory as the rate of recovery of groundwater inside the excavation was only slightly smaller than that obtained previously (Ho, et al., 2007). It was decided to add more JSG piles at the back of S58M. Three new piles were installed without problem. As No. 4 pile, refer to Fig. 9, was installed on 7 July 2005, groundwater brought a large quantity of soil into the pit. A nearby hospital was endangered and the patients in the hospital were urgently evacuated as a precautionary measure. It however survived with only minor damage. The sinkhole was about 1m in depth and settlement spread over an area of about

The committee studied the field records of diaphragm wall installation and checked the structural design of diaphragm walls. Upon the recommendation of the committee, additional soil exploration was carried out to determine ground conditions and tests were carried out to determine the quality of diaphragm walls. The design was found to be adequate and the failure was attributed to defects in wall s. Coring was carried out on s S59F, S60M and S61F in September 2005 to check the integrity of the walls. S60M appeared to be in good quality within the depth of coring of 16m. Segregation of concrete was detected in S59F at depths of 15.95m to 16.55m. Coring was abandoned at S61M as rebars were encountered at a depth of 6m. Laboratory tests were carried out on the cores obtained from s S59F and S60M and an average unconfined compressive strength of 58.7 MPa (with a standard deviation of 83.8 MPa) was obtained on the 11 specimens from S59F and an average strength of 713.1 MPa was obtained on the 9 specimens from S60M, while the design strength of concrete is ? MPa. Chemical tests on the cores indicate chloride contents varying from 0.016 to 0.022 kg/m3, less than the tolerance of 0.30 specified in CNS 3090 for ready-mixed concrete. The thicknesses of s S60M and S61F and the locations of rebars were confirmed to be adequate by using radar. However, S59F was found to be defective. 6.2 Investigation by Taiwan Construction Research Institute In addition to routine tests, such as T and SPT, TCRI carried out resistivity image profiling (RIP) of diaphragm walls and laboratory tests on soil samples taken by using GP-75S sampler which is a tool specifically developed for undisturbed sampling of difficult soils. Cross-hole resistivity image profiling was carried out at S72MF. Basically, electrical current is applied to the ground through two current electrodes and a electrical field is generated. The resulting voltage difference between two potential electrodes is measured to determine the resistance of the ground. By moving the current electrodes and the potential electrodes, ground resistances can be measured for different situations. With the modern computer technology, the data obtained can be analyzed to obtain the resistivities of underground materials at various locations and the results can be can be presented in a graphic form. Table 4 shows the typical values of resistivities of different types of soils and rocks. As can be noted, the resistivities for concrete vary from 10,000 ~ 40,000 ohm-m, while those for

3

Crosshole RIP

saturated sands, silts and clay vary from 15 to 1,000 ohm-m. The resistivity of sea water is as low as 1 to 5 ohm-m. It is therefore possible to identify defects, if any, in diaphragm walls accordingly. Pole-pole array was adopted and 3 holes were sunk on each side of the as shown in Fig. 5.Sensors were installed at 1m intervals to a depth of 40m below ground surface. (more description on operation)

Y=3.5

Crosshole RIP

Y=1.5

Y=3.0 Y=2.0 Y=2.5

Y=1.0

X

X=3.5 X=3.0 X=2.5 X=2.0 S72

X=1.5 X=1.0

Y

0

開

Station 挖 Excavation區

Table 4 Resistivities of different material JSG

Z

Material

Resistance, ρ ( Ω -m)

Fresh water Sea water Coarse sand, gravel (dry) Coarse sand, gravel (saturated) Sand (dry) Sand (saturated) Silt (dry) Silt (saturated) Clay (saturated) Sandstone Granite, gneiss Concrete, gabbro Quartzite

10,000 1~5 20,000 ~ 80,000 1,000 ~ 5,000 5,000 ~ 20,000 200 ~ 1,000 400 ~ 2,000 30 ~ 200 15 ~ 30 100 ~ 8,000 7,000 ~ 15,000 10,000 ~ 40,000 5,000 ~ 10,000

Fig. 6 3D presentation of Resistivity Image Profileing

Crosshole RIP Diaphragm Wall

Crosshole RIP JSG

12m

17m

40m

Fig. 7 Side view of Resistivity Image Profiling One row of JSG piles were previously installed at the back of the as shown in Figs. 6 and 7. To further improve the watertightness of the wall system, C piles (not shown) were installed at the ts to fill up the gaps, if any, between JSG piles. RIF was carried out between depth of 11m and 40m before the installation of C and again after the installation of C to see the differences made. Fig. 8 shows the results obtained for a longitudinal section (XZ plan, at Y = ) of the wall and soil pockets are identifiable.

GL

Low resistivity indicating soil pocket GL

JSG Gr out GL

BH-1

BH-2 1.8m

1.4m

1.2m

1.35m

BH-3 2.2m

S72MF

Low resistivity indicating soil pocket with sea water GL

Resistivity (ohm-m)

S73M 0

Y

1.7m

1.9m

BH-4

1.95m 1.8m

1.57m B1

BH-5

BH-6

X

Fig. 5 Resistivity image profiling at 72

4

0

3

180

10,000

Fig. 8 Interpretation of Results of Resistivity Image Profiling

After C Treatment

Before C Treatment

New C Tr eat ment

Pr evi ous JGP Tr eat ment

X=1.0

X=1.5

X=2.0

X=2.5

X=3.0

X=3.5

Fig. 10 Results of Resistivity Image Profiling in YZ sections Resistivity (ohm-m) 0

0

3

180

10,000

Fig. 12 3D presentation of results of Resistivity Image Profiling

Fig. 13 compares the results of RIP before and after C treatment.

Y=1.0

Y=1.5

Y=2.0

Y=2.5

Y=3.0

Y=3.5

Fig. 11 Results of Resistivity Image Profiling in XZ sections ACKNOWLEDGMENTS The authors are grateful to Da-Cin/Shimizu t Venture for the authorization to publish the data presented herein.

REFERENCES

Ho, S. K., Chou, C. C., Chen, D., Chung, L-J, Liao, Z-B, Chen, Y-Y (2007), The pumping tests at KMRT O1 Station, Proc., 2007 Cross-Strait Symposium on Geotechnical Engineering, April 15~17 (in Chinese)

5

CASE HISTORY 4

IMPORTANCE OF UNDERSTANDING LANDFORMS WHEN DEVELOPING ON LANDSLIPPED TERRAIN R.P. Thimpson Coffey Geotechnics Ltd

INTRODUCTION The landscape in northern England, U.K, includes many areas where there are relict landslides which are a reflection of former glacial or periglacial conditions. Valley landslides were typically oversteepened by meltwaters and the rapid dropdown of glacial lakes following the breaching of ice or moraine dams. As the climate became warmer, stability of the hill slopes returned but was often only marginal. Increased demand for building land has led to areas of former landslipping being considered for development. However, the first stage is to recognise and understand that the landforms are actually relict landslides and that the margins of safety against renewed movement are modest. Before even embarking on an intrusive ground investigation, it is important to review both the geology and the landforms so that the possible, if not probable factors dictating the stability can be established. These include the depth to the plane of sliding, the shear strength operating on this plane, and the groundwater conditions. The sensitivity of the slope to renewed changes needs to be understood so that the appropriate range of hypotheses that could again trigger failure is identified. Two cases are described below in which development of adjacent areas of hillside was planned. Each had somewhat different landforms and therefore required separate failure hypotheses to ensure properly targeted ground investigations were conducted and suitable geotechnical works designed and constructed. Site A The first site was on east-facing hillside and was planned for multi-storey industrial development, which was to form a further part of an already developed location (ref1). No alternative sites were available. Published information (ref2) on the geology of the area indicated that it comprised inter-layered shaley mudstones and sandstones belonging to the Carboniferous period. A particularly distinctive marine fossil band, Gastrioceras cumbriense, was noted as being present in the relevant stratiagraphy. The site itself lay about 8km south of the ice sheet at its maximum glaciation. It is thought that a glacial lake had been formed which subsequently drained rapidly and caused landslipping of the valley side. The movement would have been on the shaley mudstone layers. As the water levels fell the reprofiled valley side adopted a stable condition again, albeit marginal. Concerning the topography of the area, the lowest part of the valley side had slopes of about 10 degrees which gradually rose to about 15 degrees. However the site for the development itself was backed by a mound with slopes of about 40 degrees which in turn was followed by a stretch of level ground before the main hillside rose at an angle of about 60 degrees. The stretch of level ground and the mound were tentatively classified as comparable to graben and horst structures. The graben was backed by faults with the main slope and with the back face of the horst while the

ground downslope of the front face if the horst was also landslipped terrain, see fig. 1. The development was to be sited in the last area, downslope of the horst. The proposed construction included a semi-basement which entailed excavation works that had a serious potentonal to create renewed landslip movement. It was therefore essential to gain the fullest understanding of the ground conditions and test the hypothesis that the location was a series of relict landslides. The initial step was to propose a ground model in which the depth to the plane of former sliding was comparable to the width of the graben i.e. the level stretch of ground (ref 3). A second step was to then propose that the mound or horst had undergone a planar slide on gently dipping plane consistent with the general dip of the strata. In addition it was recognised that it would be benefit if the marine fossil band could be detected. Once the hypothetical ground model for the area upslope of the site to be developed was established, the next stage would be to focus on the site itself so that it could be safely excavated and that suitable foundations could be designed. Boreholes were drilled firstly in the area of the supposed graben. The core revealed steeply inclined mudstone layers intermixed with more broken rock, consistent with slope movement. At a depth of nearly 20m, similar to the width of the graben, intact shaley mudstone rock was encountered which thereby demonstrated that the base of the graben had been reached and allowed location of the plane of former sliding to the established. Sandstone was encountered after a further 20m of drilling. The second part to be investigated was the horst block. This was found to comprise essentially intact layers of gently dipping shaley mudstone. Marginally above the plane of sliding, as determined from the investigation in the area of the graben block, a marine fossil band was recovered which was identified as being the Gastruiceras cynvruese band. Having found the fossil band, its plan position and elevation was measured at this location. By relying on the published value for the geological dip of the strata a search was made for possible evidence of the fossil band in the main intact hillside. Fortunately there was a deeply incised narrow gorge with fresh exposures of rock and the fossil band was able to be located. Its exact plan position and elevation in this second location was measured so that an accurate appreciation of the dip of the undisturbed strata could be determined. The third part that was targeted for the investigation was the site selected for the development and the area immediately downslope. With the knowledge that the fossil band was likely to be present and that it lay marginally above the plane of sliding , the drilling was conducted with great care. On the site itself, landslide debris was initially encountered and close to the base of this fossil band was again met and was directly underlain by mudstone with one thin layer, exhibiting a polished surfacing consistent with material that had been subject to former sliding. This was carefully sampled for subsequent laboratory testing. It was determined that the residual shear was between 11 degrees and 13 degrees, the slight variation being a reflection of testing under a range of effective stress conditions (ref 4). Immediately downslope the interface between the landslipped debris and the intact mudstone was at a slightly steeper grade. This raised particular cause for concern regarding its stability, noting that the toe of this slope had been cut away with only a low height retaining wall being in place at this location. By this stage the surface topography was known and the original hypothetical ground model was proven with the location and strength of material relating to the plane of

sliding established. The final main factor was information the ground water conditions. There were two possible sources. Firstly the underlying sandstone rock, which would be relatively permeable, could collect rainfall and run-off from the main hillside and feed this from below into the overlying mudstone, in effect giving rise to an unward flow and critical porewater pressures at the plane of sliding. Alternatively, the ground surface of the graben block was seen to become flooded after heavy rain. The water could then, possibly , drain down through the broken rock fragments forming the block until the water met the relatively low permeability intact mudstone at a depth of the plane of sliding. The water would then flow along this plane in preference to downwards through the intact mudstone. Targeted instrumentation and careful monitoring established that this later hypothesis was most likely to be correct. Furthermore it revealed that the downslope of the mound, on the site for the development itself, the phreatic surface could lie very close to the current ground surface. Stability analyses of the existing slope gave varying factors of safety with the area downslope of the proposed development site having only marginal stability, of the order of 1.1. The next stage was to consider the implications of excavating the site for the development. Again such works were found to give rise to unacceptably low margins of safety in respect of the stability of the mound which could potentially move on its former plane of sliding. Finally the weight of the proposed development would render unstable the downslope section of the slope. The design to address the various geotechnical problems was divided into three parts. On the uphill side of the proposed excavation it was decided to install a contigious bored pile wall with temporary ground anchors. For the permanent condition the piles would be propped by the base slab of the development. This part was designed to ensure that movement of the uphill side did not occur under any of the foreseeable temporary or permanent conditions. The next part concerned the foundations of the building itself. It was important that not only was adequate bearing capacity achieved ,but that the building weight did not impose load on the downslope area. Accordingly bored pile foundations were selected with the piles being founded below the plane of sliding. To increase stability of the downslope area itself, trench drains were designed to lower the ground water levels. By this means the factor of safety against instability was raised to at least 1.3 even under the most severe conditions. Conclusions for Site A i)

ii) iii)

iv)

v)

The initial ground model and associated failure hypothesis were rigorously researched and were based on a comparatively complex series of landslipped blocks. The location of the plane of sliding was provisionally estimated together with the groundwater conditions and formed the basis of a tentative model. Ground investigations and detailed forensic geotechnical examination of the data obtained confirmed the validity of the original model and failure hypothesis. With this information to hand, the appropriate geotechnical design involving an anchored bored pile wall, bored pile foundations and drainage trenches could be determined and implemented. The case history A has shown that a comparatively complex landslip area can be effectively and safety developed provided the mechanism of the landslip is fully understood.

Site B The second site included a section of the access road to site A and was also on a hillside that included landforms consistent with relics landsliding. Similar geological conditions to site A existed with the solid strata comprising mudstone ad sandstone. The general topography of the area included a gentle convex toe section of the slope which then rose at a gradient of about 10 degrees, increasing to about 15 degrees before reaching a level section or plateau area at about 100m distant from the tone, see fig. 2. Beyond this the ground rose relatively steeply. Some excavation works in the toe area of the slope had been carried out already to allow constructon of the access road to be started. The contractor seemed to be peridiocally finding that after excavating a section on one day to a specified profile he was returning the next day only to find that he had to repeat the works to achieve the required cut profile. The problem was noted as being more marked in wet weather. Examination of the overall slope area revealed that fissures in the ground were appearing about 100m upslope of the toe at the junction between the top of the 15 degree slope and the plateau area. There was little doubt that the excavation works for the access road had triggered movement of a relict landslide. In addition it was the lower stage being disturbed by the construction works. However the detailed mechanism controlling the behaviour of the slope was not known. Two possible hypotheses were proposed. The first model involved a relatively shallow failure surface at about 3m depth with surface run-off ponding on the plateau area and then infiltrating the ground to give rise to periodic increases in the pore pressure on the failure surface. The second model included a failure surface at 6m to 7m depth with the pore pressures being controlled by variations in groundwater flow rates and therefore rainfall. Boreholes were drilled on the landslipped area in order to determine the actual ground conditions with depth and piezometers were installed so that the groundwater conditions could be recorded. The piezometers were also designed to act as basis slip indicators. In addition movement surface monitoring pins were placed spanning the zone at the edge of the landslipped surface so that magnitude and direction of movement between the landslip and adjacent stable ground could be established. Remedial solutions to stablise the movement were planned to be centred around the installation of drains to reduce the porewater pressures on the sliding plane. However until the ciritical plane and associated water pressures were known, it was not possible to proceed with the design of a solution let alone its implementation. One aspect that became apparent from the borehole investigation was that a geological fault diagonally traversed the centre of the landslip area so that mudstone rock was displaced downwards over the lower section of the landslip. At the fault interface, therefore, more permeable sandstone rock was juxtaposed against less permeable mudstone rock. The boreholes also revealed that non-solid deposits, being drift or landslipped material, extended to about 6m to 7m over most of the landslipped area increasing to about 10m depth near the head of the slip. However there was no horizon that could be conclusively determined as being the slip plane, other than possibly the interface between non-solid and the solid deposits.

A rockfill berm was placed at the toe of the slope to provide a modest improvement in the short-term stability of the slope. It was not until about nine months afterwards that there was a period of particularly heavy rainfall. The piezometers revealed that there was a marked increase in pore pressure at the interface of the non-solid and solid deposits. Futhermore the slip indicators in the piezometer tubes confirmed that movement was taking place on this horizon. The original ground model and its associated failure hypothesis were duly modified to embrace the information that had been established regarding the location of the movement plan, the presence of the faulting and that this fault dictated the critical porewater pressures. From this information, a 7m deep drainage trench was designed and constructed running diagonally across the slope and positioned immediately downslope of the line of the fault. This ensured that under very wet conditions the groundwater was drained away from the fault zone before it had the opportunity of reaching and disturbing the stability of the toe of the slope. With the toe stablised, the upper pat of the slope was effectively buttressed so that it likewise was rendered stable. Conclusions for Site B i) ii) iii) iv) v) vi)

The intial ground model and associated failure hypotheses were reasonably based on the site being a straightforward two-stage relict landslip. The details relating the plane of sliding and critical porewater pressures were not known but two tentative models were proposed. Ground investigations revealed that the original ground model required modification in order to properly understand the behaviour of the landslip. With this information to hand, appropriate drainage works to address the critical porewater pressures could be determined. The tentative drainage solutions based on either of the original modesls would have been seriously flawed and would not have provided stability to the landslip. The case history B has shown that an apparently relatively straightforward landslip may, in fact, be complex and that rigorous forensic geotechnical engineering is essential to gain a correct understanding.

Overall Conclusions 1. The examples for the two sites, which concerned relict landslides, demonstrate that it is essential from the outset to develop well thought out and researched failure hypotheses together with ground models and that this should be done prior to intrusive ground investigations being carried out. 2. On the basis of the hypotheses, the ground investigations can be target to give the optimum results. It is essential to review and modify, if necessary, the initially chosen hypothesis in the light of the findings from the investigations. Provided this is achieved, the geotechnical aspects of a development can be designed and implemented with known levels of safety. 3. One of the sites, Site A, was complex but the rigorous forensic geotechnical examination of the findings from the borings and field exposures allowed the original ground model and associated failure hypothesis to be confirmed. 4. The other site, site B, appeared to be straightforward two-stage landslide. However the two originally proposed possible failure hypotheses were found not to match the actual ground conditions as revealed by intrusive investigations. It demonstrated the requirement to carefully examine the information obtained and then to modify the original hypothesis accordingly.

5. From the examples it is evident that if proper failure hypotheses and targeted ground investigations are not conducted, the subsequent development could lead to renewed slope movement. This is exemplified, in particular, by the initial works that were carried out at site B. References 1. Leach B.A. and Thomson R.P. (1987), The influence of underdrainage of the stability of relic landslides. 9th ECSMFE Dublin Vol.1pp 447-450, Balkema 2. The Geology of the Country around Huddersfield and Halifax (1930), Explanation of Eological Sheet 77, Memoirs of the Geologifcal Survey of England & Wales. DSIR His Majesty’s Stationary Ofrice London. 3. Holmes A. (1965), Principles of Physical Geology. Second Edition pp 1046-1051, Nelson. 4. Thompson R.P (1991) Stabilisation of a landslide on Etruia Marl.International Conference on Slope Stability, Isle of Wight UK. Pp 403-408. The Institution of Civil Engineers. Thomas Telford.

Case Studies 5

Monitoring in Forensic Geotechnical Engineering - Dam Construction – Yoshi Iwasaki, Ph.D., Dr of Eng, P.E. Managing Director Geo-Research Institute, Osaka 1 Observational Procedure The text book of “Soil Mechanics in Engineering Practice, Second Edition” by K. Terzaghi and R.Peck(1967) added Chapter 12 Performance Observations to the first edition. It was the first to point the enrolment of the monitoring soil and structural performance during construction work as to provide evidence in lawsuits. It states that a fair decision is to be expected only if the causes and the nature of the mishap are known. If the contractor or owner can prove that he has anticipated the undesirable condition, has observed its progress during construction, and has done everything possible to avoid it, he is in a much more favorable position than if the condition has taken him by surprise. The element of surprise not only injures his professional reputation, but it may also injure his financial standing. When the third edition by the authors of K.Terzaghi, R.Peck, and G.Mesri was published, the Chapter 12 in the second edition was deleted and refers observational procedure only in a single paragraph of “10.5 Observations during Construction,” where no legal aspects was mentioned for usage of monitoring. 2 Enrolment of observation in forensic geotechnical In the third edition of “Soil Mechanics in Engineering Practice”, it is stated that “in earth-dam and foundation engineering the permanent structures are designed before the construction operations start, and the consequences of unanticipated sources of trouble do not appear until the structure is in an advanced state of construction or is in service.” This statement sounds observational method is not useful for earth dam and foundation engineering. The author will review several case histories that are constructed with or without monitoring including forensic problems and try to consider any meaning of monitoring in dam construction. In very early stage of dam construction, St.Francis Dam, Los Angeles County, California, U.S.A. was constructed as a concrete gravity dam in March 1926. During the

1/7

initial filling, cracks in the dam began to appear and was tried to seal the cracks. During the final stage of the second filling in March 1928, new leakage was found and the dam was failed. The St.Francis Dam is considered the greatest American civil engineering failure of the twentieth century. Approximately 450 mostly sleeping people died by flood caused by the failure of the dam. When the dam was constructed, dam engineering was just to start to develop based upon these failures and lessons learnt. Another example is Vajont dam, Italy that was a concrete arch dam constructed in 1962 and failed in 1963. 2.1 Vajont Dam Vajont Dam is a dam completed in 1961 under Monte Toc, 100km north of Venice, Italy. It was one of the highest concrete dams in the world measuring 262m in height with 27m in base thickness and 3.4m in top thickness. Shifting of rock was noticed during the first filling in February 1960 before final completion of the dam of September 1960. By March 1960, the level of the reservoir reached 130 above the level of the river, when the first small slide was observed. Table-1 The Sequence of Filling/Draw Down leading to Failure date

The First

Filling

Stage

1960/2 1960/3 1960/10 1960/11/4

Draw Down

The Second Stage

1961/10 1962/2 1962/11 1962/11 Draw Down 1663/2

Filling

1663/3 1993/5 1993/6 Filling 1993/7 The Third 1993/8 Stage 1993/9 1993/10 Draw Down 1993/10/9 1993/10/9

water

movement

level

rate

130m 170m 180m 185m 135m 135m 185m 235m 185m 185m 231m 237m 240m 240m 245m 245m 235m 22:38

mass movement/land slide

3.5cm/day potential huge slide area recognized 0.7millon slid 8cm/day 0.1cm/day

1.2cm/day 1cm/day 0cm/day

0.3cm/day 0.4cm/day 0.5cm/day 0.8cm/day 3.5cm/day 20cm/day heavy rain

2/7

260million m3 slid

In October 1960, the level reached 170m, a rapid increase in the rate of displacement to approximately 3.5cm/day was observed. At this time, a huge t of 2km in length was found to open up in an area of slope 1700m long and 1000m wide, suggesting very large slide was mobilized. On November 4, 1960, there took place a land slide of 700,000 cubic meters at the level of 180m. The water behind the dam was lowered to 135m and the movement was reduced to 1mm/day. The safety of the slope of the dam was discussed and calculations ordered by management showed that catastrophic failure was unlikely and the reservoir allowed to refill water under controlled monitoring. Thus the reservoir was filled and slowly emptied three times. On October 9, 1963 at 10:38 pm, the combination of third drawing down of the reservoir and heavy rains triggered a big landslide of about 260million cubic meters of forest, earth, and rock, which fell into the reservoir. The resulting splashed water caused some 50million cube meters of water to push up the opposite bank and destroyed the Village Casso, 260m above the lake level before overtopping the concrete dam of 250m in height. The water, estimated about 30million m3 then fell more than 500m onto the downstream villages of Longarone, Pirago, Villanova, Rivalta, and Fae destroying them and killed some 2500 persons. The court condemned eight persons in guilty not to take any action of evacuation of the inhabitants. In the case of Vajon dam, monitoring was being carried out, however, it was very difficult to predict the failure of the slope. 2.2 Teton Dam Teton dam constructed and failed in1976, Idaho, U.S.A. is another case that has no monitoring except measurement of deformation. Teton dam, Idaho, the major feature of the multipurpose Teton Basin Project of the United States Bureau of Reclamation, failed on June 5, 1976 while the reservoir level was 1m below the spillway sill. The dam was a center core zoned earth fill structure, with a height of 93m above the river bed and 123.4m above the lowest foundation. The instrumentation of Teton dam was not prepared because performance records were considered to be available for dams “constructed of similar material and on similar foundations” and only settlement monuments were provided. 2.2 Takase Dam

3/7

Takase dam is a central core earth-rock dam, 176m in height, 362m in crest length built in 1978 by Tokyo Electric Power Company Inc. as a part of Takase River Hydraulic Power Development Project developed in Takase Basin, central Japan.

Fig.1 Pore pressure measurements at Takase dam during initial filling (1)

4/7

Fig.2 Pore pressure measurements at Takase dam during initial filling (2)

Fig.3 Change of equi-pore water pressure line during the first filling (Dec.,’78, Feb.,79, Apr.,79, and July, 79) Instrumentation of Takase dam was arranged to provide pore pressure, earth pressure, displacement in dam section, and surface targets for measurement of surface movements. In placing the pore pressure gauges for control of the dam construction work, it was decided to concentrate mainly on the core and transition zones, where comparatively

5/7

high pore pressure was expected to develop, and also to a central extent on the inner shell zone, which has a relatively high fine grain content. Pore pressure gauges for dam security control during and filling of the reservoir were located on the core area and within the foundation bedrock, and also in positions that enable tracking of the downstream phreatic line. Fig.1 and 2 show changes of pore water pressure during the initial filling water in the reservoir. The filling started in December, 1978 and reached nearly completion in July, 1979. Each figure shows the measured pressure at three different heights within the embankment of EL=+1107m, EL=+1165m, and EL=1135m. Fig.3 shows change of equi-contour line during the filling (Dec,’78(left top)-Feb., 79(left bottom, Apr.,79(right top), and(July,79(right bottom )). It is seen that the distribution of the pore pressure in the field is sound condition from which no leakage is expected. 2.3 Dams that were saved from failure based upon monitoring ASCE Task committee on instrumentation and monitoring dam performance summarized various case histories that showed saved dams from catastrophic failures based upon instrumentation and monitoring. The following three examples are from the case studies shown in the guidelines for Instrumentation and Measurements for Monitoring Dam Performance. Findlay(1988) reports a dam for a hydroelectric project in 1986 included the installation of instrumentation and monitoring system. Pneumatic piezometers were installed to monitor hydrostatic pressure during and following the initial filling. It was found the reading of two sensors increased equivalent with rising water elevation. Piping might have occurred if the head condition were continued. The reservoir was immediately dewatered. The pore water gages responded with the same amount of the dewatered level. The two piezpmeters were installed near a contraction t. Modification was made for inadequate sealing and a safe condition was achieved. Only after one year of a dam of USBR(Fontenelle Dam) in 1965, piping of the embankment and surface erosion caused a partial failure. Damage to the 36.6m high, 1626m long rolled-earth embankment was repaired and the dam put back in service. In 1982, instrumentation and monitoring indicated the presence of new seepage. A cut-off wall of about 60cm thick and 48m deep along the whole embankment of about 1.6km was installed to stop the leakage. (ENR,Nov.26, 1987) Navajo Dam of U.S. Bureau Reclamation was detected seepage problem during the initial filling in 1963. Seepage reading of the dam had been increased until the installation of a 100cm thick, 120m deep, 135m long cutoff wall.(Civil Engineering,

6/7

March 1987) It is clear that the observational procedure has been applied successfully in dam constructions based upon monitoring of the dam behavior during the first filing as the final stage of construction or even later phase under service stage. 3 Conclusions If you search for “Teton Dam,” through internet Home Page, you will find hot discussions still going on the true reason that made the dam failed. If the performance was monitored, the discussion might be more fruitful and practical based upon real data. Failure of Teton Dam clearly shows the measurement of power pressure within dam section is necessary and if it had been measured, the failure of the dam itself had been prevented. Vajon dam shows inadequate theory and practice of the slope failure in geotechnical engineering. The statement in the textbook of “Soil Mechanics in Engineering Practice” the consequences of unanticipated sources of trouble do not appear until the structure is in an advanced state of construction or is in service” should be replaced by “the construction of a dam continues until the initial filling water and should be confirmed as assumption by design or ready to be modified based upon the monitoring results during the filling water in the dam. References: W.L.Chadwick(1977), Case Study of Teton Dam and Its Failure, Proceedings of the Ninth International Conference on Soil Mechanics and Foundation Engineering, Case Study Volume, Tokyo, Japanese Society of Soil Mechanics and Foundation Engineering, pp431-455 R.Takai, T.Iwakata, and Y.Miyata (1977), Results of Soil Tests and Measurements during and after Construction of the Takase Dam, Proceedings of the Ninth International Conference on Soil Mechanics and Foundation Engineering, Case Study Volume, Tokyo, Japanese Society of Soil Mechanics and Foundation Engineering, pp497-554 ASCE Task Committee(2000)”Guidelines for Instrumentation and Measurements for Monitoring Dam Performance” Findlay, R.Craig (1998), “Hydrostatic Pressure at a Soil-Structure Interface,” Proceedings of the Second International Conference on Case Histories in Geotchnical Engineering, University of Missouri-Rolla, St.Louis

7/7

Case Study 7

FORENSIC ANALYSIS ON RE WALL DISTRESS: A CASE STUDY Sivakumar Babu, G. L.1, Murthy, B. R. S.2 and Singh, V. P.3 1. Associate Professor, Dept. of Civil Engg., IISc, Bangalore, India, 560 012, [email protected] 2. Retd. Professor, IISc, Bangalore, India, 560 012, [email protected] 3. Research Associate, Dept. of Civil Engg., IISc, Bangalore, India, 560 012, [email protected]

ABSTRACT The article presents a forensic analysis of the observed distress of an RE wall. The investigation involved field observations, extensive laboratory testing of backfill soil properties and reinforcement material (i.e. geogrids), analysis of failure cause and justification for the soil nailing technique as a rehabilitation measure. The article also highlights soil nailing technique as a viable alternative for rehabilitation of failed reinforced earth (RE) structures.

1.

INTRODUCTION

Forensic geotechnical engineering (FGE) may be defined as the discipline that prepares geotechnical engineers to investigate failures of structures such as embankments, pavements, foundations and earth retaining systems, in purview of professional ethics and legal bindings. FGE being a relatively new terminology in geotechnical engineering profession, it is necessary to create general awareness and understanding about the subject among geotechnical engineers, clients, and contractors. In this article, a case study of RE walls constructed for a flyover is presented wherein a detailed investigation was sought by the client with regard to the examination of the distress noted during the construction. The methodology adopted to carry out detailed investigation involved following sequence of steps. 1. Survey and characterization of the distress i.e. site visit and observations. 2. Diagnostic tests i.e. collection of material (backfill and geogrid) samples for laboratory testing to evaluate the behavior of soil and the strength characteristic of geogrids used. 3. Development of most probable failure hypothesis and cross check with original design. 4. Back analysis of RE walls stability using conventional methods, and rigorous numerical analysis using computational code. 5. Suggestion and justification for the rehabilitation measure in the form of driven soil nails.

2.

SURVEY AND CHARACTERIZATION OF DISTRESS

Site visits were made to (a) identify and study locations of distress, (b) collect the samples of backfill and geogrids for laboratory testing, and (c) procure a preliminary report on the possible cause of distress from the client/contractors. For the purpose of identification, the two ends of

flyover were named as the Siliguri end and the Dalkola end. The construction work of the RE wall started in June 2006 and the Siliguri end of the flyover was completed in April 2007 and the Dalkola end by August 2007. During June –October 2007, the area experienced severe rains which were of unprecedented magnitude causing distress in the RE wall that was being constructed. On the Siliguri end, distresses in RE wall were noted in the following forms: i) damage to reinforced earthwork ii) damage near abutment leading to misalignment of s, and iii) sand flow was noticed and one of the s got dislocated in the region near the design height of 8 to 8.8 m. On the Dalkola end, the client informed that the sand removal led to overstressing of the connection, resulting in the breakage of grid and s. A few photographs of the distress locations in the RE walls taken during the site visits are shown in Figures 1(a)-(c).

Figure 1(a): Damage near flyover abutment

Figure 1(b): Bulging of facing

Figure 1(c): Damage near top facing and the approach slab

3.

DIAGNOSTIC TESTS ON MATERIAL SAMPLES

3.1

Test Results of Backfill Soil Samples

Backfill soil samples collected during site visits were tested in the laboratory and used to obtain shear strength parameters in both the undisturbed and saturated conditions. Table 1 shows the summary of the shear strength parameters obtained from laboratory tests on backfill soil samples.

Table 1: Shear strength parameters of backfill Soil state In situ condition In situ condition Saturated condition

Bulk density 3 (kN/m )

Peak friction angle (degrees)

Residual friction angle (degrees)

Cohesion (kPa)

17.50

40.0

35.1

0

16.08

39.1

33.2

0

17.50

30.54

29.8

0

It is noted that the shear strength parameters for the in situ condition are higher than the design value of 32o. Tests on saturated sand indicate that there is a marginal decrease in the friction angle. The permeability coefficient of the backfill corresponding to the bulk density of 17.50 kN/m3 is in the range of 0.089 mm/sec. It is likely that the rate of infiltration during the intense rainfall was higher than this value.

3.2

Strength and Permittivity Characteristics of Geogrids Samples

The permittivity of the geogrid (geotextile) used is 1.85 sec-1 and the thickness of the geogrid is 1.68 mm which corresponds to permeability coefficient of 3.1 mm/sec. For proper filter behavior, as per NCMA (1996) guidelines, the geogrid permeability coefficient needs to be not less than 10 times higher than the permeability coefficient of the backfill. In this case, the geogrid is highly permeable and the factor of safety with respect to filter criteria is in the range of 3.48 (3.1/10 x 0.089) and hence, the geogrid used was satisfactory and a separate drainage medium behind the facing s was not essential. Further, the design strength and other properties of geogrid samples were assessed by another agency and the result supplied by them showed that the used geogrids satisfied their technical specifications.

4.

DEVELOPMENT OF MOST PROBABLE FAILURE HYPOTHESIS

Rainfall leads to saturation of the backfill and/or pore water pressure development which needs to be addressed properly. The effect of rainfall on reinforced wall can be understood from design charts for reinforced soil slopes for dry and saturated conditions developed by Jewell (1990). Figures 2(a) and (b) show the minimum required force coefficient (kreq) for reinforced soil structures with slope angles varying from 0 to 90 degrees. The state of pore water pressure is expressed in of pore pressure coefficient ru = u/γz, where u is the pore pressure, γ is backfill bulk density and z is the depth at the point of considerations of pore water pressure (u) and vertical stress (γz). The pore pressure coefficient varies from 0 to 0.5 and dry/fully drained backfill has ru = 0, whereas presence of saturated water in the backfill is indicated by ru = 0.5. It is expected that minimum required force expressed in of (kreq) in the reinforcement for equilibrium is lesser for fully drained backfill. From Figures 2(a) and (b), it can be noted that for fully drained backfill, for the slope angle 900 representing wall, for design friction angle of 320, the kreq value is 0.31. This value increases to values in the range of 0.60 when pore pressure in the backfill exists. The pore pressure generated in the backfill in a momentary condition as a result of heavy downpour and is not considered in design. Thus, when rain of unpredicted magnitude occurs and if the infiltration rate is higher than rate of discharge, seepage pressures build up and momentarily induce additional pressure on the retaining system. The local bulging observed in a few locations in the RE walls of the flyover is attributed to the above phenomenon. Further, for a completely covered surface in the plan area of the wall, percolation of water is not allowed. For this condition, using design friction angle of 32o, the minimum required force coefficient (kreq) for reinforced soil structure of 90o (vertical wall) is 0.31 and the corresponding force required to be supplied by the reinforcement at a depth of 10.40m is (kreq)γH is 0.31x17.5x10.4 = 56.42 kN/m. It is noted that the geogrids provided at this depth have higher strength than required value of 56.42 kN/m. However, under the influence of pore pressures, the minimum required force coefficient (kreq) is in the range of 0.64 and hence, the force momentarily can increase to a maximum value of 116 kN/m, applying considerable force on the reinforcement and s which could have contributed to the distress noted. It is also noted that for the fourth facing from the top, two types of geogrids (SS120 and SS150) which have different stress-deformation characteristics were used. The tensile force

required for stability is marginally in the same range whereas elongations required to obtain the same force in each geogrid layer are different. This could have contributed to marginal drift in the facing s under the influence of earth pressures.

Figure 2: Jewell (1990) design charts for reinforced slopes (a) ru = 0; (b) ru = 0.5.

5.

BACK ANALYSES OF RE WALLS STABILITY

The back analyses to examine the stability of the geogrid reinforced retaining walls was carried out using conventional method (FHWA 2001) and by performing rigorous finite element analysis using a computational code (Plaxis 2006). Various sections of the RE walls with different design height of wall (e.g., 10.40 m, 9.60 m and 8.80 m) were analysed separately. In the present paper, results of the analyses corresponding to the maximum design height (i.e. 10.40 m) are presented and discussed. Table 2 shows the material properties and other parameters adopted for the study. Table 2: Material properties considered for analysis Parameter Design height of the wall, H (m) Unit weight of wallfill, γ (kN/m3) Angle of internal friction of wallfill, φ (degree) Surcharge load, ws (kPa) Length of geogrids member, L (m) Vertical spacing of geogrids , SV (m) Geogrid-soil interaction coefficient, μ Design strength of geogrid Note: Figures in bracket are for the bottom most layer of geogrid.

Value 10.40 18 32 22 7.80 (7.20) 0.80 (0.365) 0.85 As per the data supplied by the client

It is to be noted that the properties of foundation and backfill are assumed to be same as that of wallfill. Tables 4 and 5 shows the typical results. Figures 1 and 2 shows the finite element models of the RE wall of 10.40 m design height with and without use of soil nailing as remedial measure respectively. 5.1

Results of Conventional Analysis

RE wall is found safe against overturning failure with minimum eccentricity of 0.85 mm which less than the permissible vale of 1.30 mm (i.e. L/6). Also, wall is safe against sliding failure having a minimum factor of safety = 2.50 which is more than the minimum recommended value (i.e. 1.50). Further, no evidence of bearing capacity failure were reported or found during site visit. Thus, it can be concluded that the RE wall is externally stable. Table 3 shows the result of the internal stability analyses. According to FHWA (2001), minimum recommended values for factor of safety against pullout FSP and tensile failure FST of the reinforcement are 1.50 and 1.30, respectively. It can be seen (indicated in bold) from Table 3 that the FSP values for the top four geogrid layers and FST values for the bottom four geogrid layers are less than the minimum recommended values.

Table 3: Results of conventional analyses for section with design height 10.40 m

Geogrid layer number ( from top )

Depth z (m)

1 2 3 4 5 6 7 8 9 10 11 12 13

0.44 1.24 2.04 2.84 3.64 4.44 5.24 6.04 6.84 7.64 8.44 9.24 10.04

5.2

Design tensile strength of geogrid, Pd (kN/m) 29.70 29.70 39.60 39.60 59.50 59.50 59.50 74.40 74.40 74.40 74.40 74.40 74.40

Factor safety against pullout, FSP

Factor of safety against tensile failure, FST

0.81 0.99 1.19 1.41 1.64 1.91 2.19 2.51 2.85 3.24 3.68 4.16 4.35

4.36 2.85 2.79 2.17 2.62 2.17 1.82 1.93 1.66 1.43 1.24 1.09 0.90

Results of Finite Element Analysis

As mentioned earlier, the section of RE wall with design 10.40 m was simulated using finite element computational code Plaxis (2006). Mohr-Coulomb model was used to represent soil behaviour and geogrid layers were simulated using ‘Geogrid’ structural elements (yellow color) available in the code. Details on simulation procedure can be found in the manual of the computational code. Figure 3 shows a typical example of numerically simulated RE wall.

Figure 3: Typical example of numerically simulated RE wall.

The results of the finite element analysis are shown in Table 4. From Table 4, it can be seen that the bottom geogrid layers are not safe against tensile failure. Further, the maximum displacement of the wall is observed as 26.53 mm.

Table 4: Results of finite element analysis for section with design height 10.40 m

Geogrid layer (from top)

Depth z (m)

Design strength of geogrid, PD (kN/m)

1 2 3 4 5 6 7 8 9 10 11 12 13

0.44 1.24 2.04 2.84 3.64 4.44 5.24 6.04 6.84 7.64 8.44 9.24 10.04

29.70 29.70 39.60 39.60 59.50 59.50 59.50 74.40 74.40 74.40 74.40 74.40 74.40

Axial force developed in each member (kN/m) 16.34 16.56 12.21 16.52 19.20 18.99 22.36 32.05 37.06 46.30 60.73 78.99 133.59

Horizontal displacement (mm)

FST

26.53 26.44 26.42 26.39 26.36 26.32 26.27 26.22 26.16 26.10 26.04 25.98 25.94

1.82 1.79 3.24 2.40 3.10 3.13 2.66 2.32 2.01 1.61 1.23 0.94 0.56

5.3

General Observation from Back Analyses

From Tables 3 and 4, it is evident that in the bottom layers of geogrids, the axial force developed is more than design tensile strength of geogrids; at depths beyond 6m as the factor of safety is close to 2 or less. The height of the wall above the ground level is in the range of 8m. Hence, in this range it is preferable to increase the factor of safety with respect to tension, and, it is recommended that soil nailing could be employed.

6.

SOIL NAILING AS REHABILATION MEASURE

FHWA (2003) may be referred for detailed information on the analysis, design and construction of soil nail walls. Figure 4 shows the numerically simulated model of RE wall with soil nails. Table 5 shows the results of finite element analyses of the RE wall, additionally reinforced with 25 mm diameter horizontally driven nails of length 7.80 m spaced at 0.80 m vertically between geogrid layers and 1.00 m horizontally along the length of the wall. From Table 5, it is evident that improvement in the values of factor of safety against tensile failure FST is observed. For the bottom most geogrid layer FST value has increased from 0.56 (for geogrids alone) to about 1.11 (for geogrids along with soil nails) which almost two times improvement. It is also observed that maximum axial force developed in the bottom most soil nail is about 58.08 kN which yields a factor of safety value against nail tensile failure FST equal to 3.53. Hence, use of soil nailing significantly improves the stability of RE wall and is therefore, suggested as the possible remedial measure for short term and long term stability of RE wall. Soil nails may be provided where the factor of safety with respect to tensile forces is less than 2 and it can be noted from Table 4 that soil nailing is required at the top as well as bottom. It may also be noted that the wall height above the ground level is 8m. Hence, the results presented in Table 5 needs to be considered for 8 m depth only.

Figure 4: Numerically simulated RE wall with soil nails (black elements indicate soil nails).

Table 5: Results of finite element analysis of RE Walls with soil nails* Axial force Horizontal developed in displacement FST each member (mm) (kN/m) 1 0.44 29.70 8.57 13.07 3.46 2 1.24 29.70 9.12 12.86 3.25 3 2.04 39.60 6.38 12.75 6.20 4 2.84 39.60 7.07 12.69 5.60 5 3.64 59.50 7.50 12.65 7.93 6 4.44 59.50 6.00 12.63 9.91 7 5.24 59.50 6.72 12.61 8.85 8 6.04 74.40 10.42 12.58 7.14 9 6.84 74.40 13.87 12.55 5.36 10 7.64 74.40 18.96 12.53 3.92 11 8.44 74.40 25.15 12.50 2.95 12 9.24 74.40 35.04 12.48 2.12 13 10.04 74.40 67.12 12.47 1.11 * - 25 mm diameter driven nails of length 7.80 m spaced at 0.80 m vertically between geogrid layers and 1.00 m horizontally. Geogrid layer (from top)

6.1

Depth z (m)

Design strength of geogrid, PD (kN/m)

Advantages of Using Soil Nails as Rehabilitation Measure

As indicated from the results of the numerical analyses of RE wall with soil nails (see Table 5), the contribution of soil nails in improving the stability of the RE is found to significant. Introduction of soil nails helped in the aspects: 1. Redistribution of axial forces among soil nails and geogrid layers and thus improving the tensile capacity of the original reinforcement (i.e. geogrids). 2. Reduction in net force at facing connection thereby improving connection stability. 3. Reduction in the lateral displacement of the RE wall to almost 50%. 4. Improvement in the overall stability of the RE wall against any unforeseen event such as unprecedented heavy rainfall.

7.

FINAL RECOMMENDATIONS AND PRECAUTIONARY MEASURES

The analysis shows that the unprecedented rains during construction (temporary phenomenon) resulted in very high lateral pressures contributing to minor movement of s in the area. The geotextile provided for filtration was adequate. The backfill properties measured from undisturbed samples were satisfactory. However, it is essential that proper drainage measures should be implemented so that the backfill is not saturated during the remaining part of construction.

Soil nailing to improve the stability of RE wall is recommended as the rehabilitation measure. Driven soil nails of 20 mm diameter and length 5.6 m are suggested near the Siliguri end abutment. In general, nail lengths of 0.7H may be provided at locations of distress. The height shall be reckoned as the height above the ground level. Two nails per facing may be provided when the height exceeds 6 m. It is to be noted that the above recommendation about the provision of soil nails is based on the similar analysis performed for other sections of RE wall with smaller heights, keeping in view the cost of rehabilitation process, and ease of application. 7.1

Suggested Precautionary Measures

1. It is essential that the nailing process should not leave the facing ugly. 2. As noted during the field observations, gap was not maintained between the top of the facing and the approach slab, creating additional loads on the facing s. This contributed to movement of the facing . It is suggested that gap may be created such that the stress transfer is avoided. 3. It is reiterated that normal conditions considered in design shall be ensured during execution. 4. Unexpected loads in the form of seepage/pore pressures, loading on the facing from approach slab shall be avoided. 5. It is advisable that during impending monsoon, the embankment area be covered properly and handle local problems if any from outside only.

8.

CONCLUDING REMARKS

The article presents a case study of failed RE wall, wherein, a through investigation about the cause of distress observed during the construction of RE walls was sought by the client. Although, this study do not addresses the legal issues related to the investigation, it attempts to describe the general procedure that may be followed in the forensic geotechnical engineering.

9.

REFERECNES

FHWA. (2001). Mechanically Stabilized Earth Walls and Reinforced Soil Slopes: Design and Construction Guidelines. Report FHWA-NH1-00-043, U. S. Department of Transportation, Federal Highway istration, Washington D. C. FHWA. (2003). Geotechnical Engineering Circular No. 7 - Soil Nail Walls. Report FHWA0-IF-03017, U. S. Department of Transportation, Federal Highway istration, Washington D. C. Jewell, R. A. (1990). Revised design charts for steep reinforced slopes. Proceedings of the Symposium on Reinforced Embankments: Theory and Practice, Thomas Telford, 1-30. NCMA. (1996). Design Manual for Segmental Retaining Walls. National Concrete Masonry Association, 2nd edn, NCMA, Herndon, VA. Plaxis (2006). Reference Manual. Delft University of Technology & PLAXIS B.V., The Netherlands.

Case Study 8 DISTRESS TO REINFORCED EARTH WALL EMBANKMENT: A CASE STUDY Prof. T.G. SITHARAM Professor and Chairman, PMG, Prof. K.S. SUBBA RAO Professor (retd) Department of Civil Engineering Indian Institute ofScience Bangalore – 560012 SUMMARY In this case study, the distress to the Road over bridge (ROB) with a 6 lane highway (about 30m wide) constructed using reinforced earth wall technique with a maximum height of about 11m is discussed. A view of constructed RE wall is shown in Fig1. The longitudinal cracks as seen on the ROB-II are shown in Figure 2. The typical cross-sectional details of ROB are given in Fig 3. After the completion of the ROB-II, it was opened for traffic on 15.07.2007. During inspection on 20th September 2007, some longitudinal cracks were noted on the pavement parallel to the reinforced earth wall face (As in Fig 2) on both Right Carriage Way (RCW) and Left Carriage way (LCW) of the road between chainages 195.0m and 390.0m (see Fig 4). On both sides of the road cracks have appeared just at the end of the GI mesh reinforcements of the RE wall which is at a distance of 5.0 to 8.5 m from the Reinforced earth wall s. The maximum width of crack was mostly 1 to 3mm and at one place it was 9mm. Immediately, the longitudinal cracks were filled up and rolled. At that time, some cavities have also been observed as shown in fig 4. The asphalt filling up which had been done on the observed cracks showed partial success in that the crack on the left carriage way from chainage 195 to 295 did not reappear. But the one on the right carriage way from 295 to 390 did reappear. In addition, on the 19th of October, 2007 new hair line cracks were observed on the left carriage way and the right carriage way as indicated in the Figure 2. These hair line cracks were seen to be confined to the top bituminous concrete layer, however extending upto Ch 405 on LCW and Ch +205 on RCW. The observation on the verticality of the end s were also recorded from 26.09.2007 to 15.10.2007. The recorded movements during the above mentioned period remained fairly stable and showed no evidence of perceptible distress propagation. The movements mostly are in the range of 5 to 30 mm which are generally expected in RE walls for mobilization of shear forces on the soil-reinforcement interfaces. In order to get some additional information on the compacted state of the backfill, geotechnical investigations at three more locations in the ROB section of the carriage ways were suggested and carried out.

Fig 1. View of RE wall which is constructed for ROB

Fig 2. Longitudinal cracks on the carriage way

Fig 3. Typical cross section of the ROB

Fig 4. Cavity observed on RCW

Fig 5. Longitudinal cracks as seen on both RCW and LCW of ROB –II Most of the boreholes in the reinforced earth wall embankment section show a good penetration resistance with depth but only in one o the borehole (which is close to the longitudinal crack on the right carriage way) shows some evidence of loose pockets and with very low penetration resistance (at a depth of 4.5m from the road surface) and the presence of boulders (between depths of 5 to 8m from the road surface). The results of ground penetrating radar survey indicate the presence of loose pockets and cavities in the top layers of compacted fill. Design of RE wall was correct and the Factor of Safety (FS) for the global stability is high, which rules out the global stability concern (as expressed by the designer and the

owner) for the ROB-II section with distress. GPR survey has also not indicated any formation of slip circle type of failure ing through the embankment section. The design calculations for the retained earth wall consisted of internal stability checks (reinforcement rupture and reinforcement pullout) external stability checks (for overturning and sliding) and they are found to be in order. It is unlikely that the longitudinal cracks that appeared might have been caused due to these internal and external stability issues. Some settlement might have been caused due to poor foundation (as observed in BH-10 with values of N=10 corresponding to a depth of 8.5m from the road level). However, this may not be very significant as we did not observe any level difference across the crack. Selection of borrow materials and Compaction control was found to be the cause of distress and cracks. Most of the cracks are confined up to the bottom of the granular sub base layer. Granular sub base layer seems to be most affected layer. From the detailed investigations (forensic), the cause of longitudinal cracks are attributed to: a) From the median area or the voids that might have been present in the top layers of pavement (BC, DBM and BM), rain water got seeped into the granular sub base layer and further onto subgrade which might have caused differential settlement of soil layers at the junction of reinforcement section and un-reinforced section of the embankment and resulted in the form of longitudinal cracks on the pavement. This is evident from the differential settlement in various soil layers as observed in GPR survey. b) These longitudinal cracks are mainly confined in the top 1.5m from the road surface. However, there are indications that these cracks might have extended up to 4 to 5m into the compacted soil (as indicated in one of the boreholes) during the heavy rainy days and subsequent seepage at some locations, where it had small boulders causing a high permeability zone. At some locations, densities of the backfill material obtained in the embankment section did not meet the desired values. c) Selection of granular sub base material and improper drainage through the granular sub base layer. Cement grouting is recommended for the compacted backfill and has been done all along the longitudinal cracked zones at regular intervals of 3m c/c. c. The median portion was recommended to be made impermeable by providing LDPE layer along with PCC of about 50 mm thick all along the length of the median to prevent water entering the pavement. Premixed seal coat comprising of a fine aggregate premixed with bituminous binder was also recommended for sealing the voids in a bituminous surface laid to the specified levels, grade and camber for the full width of the carriage way.

Case Study 9

EXPERIMENTAL AND NUMERICAL TECHNIQUES FOR ANALYSIS OF SOIL NAILED STRUCTURES 1

K. Rajagopal1 and T. Rambabu2 Professor, Dept. of Civil Engineering, IIT Madras, Chennai 600036, e-mail: [email protected] 2 T. Rambabu, M/s Utraco-Ryobi Ground Engineering Pvt. Ltd., Chennai-87, e-mail: [email protected]

ABSTRACT This paper discusses the comparison in the response of soil nailed retaining walls and slopes obtained by different methods i.e. experimental, limit-equilibrium and finite element analysis. The paper describes some laboratory tests in which 1.75 m high soil nailed retaining walls were subjected to failure by applying uniform surcharge pressure. Their response was studied by three different methods and comparisons were made. Later, the paper describes the stabilization of deep excavations at a construction site in Chennai using pre-stressed nailing technique. The construction site is surrounded by fully-developed residential areas. Any slope failure at the site will result in loss of foundation to the adjacent structures which may lead to many legal problems. After carefully considering various methods for stabilizing the excavated faces, it was decided to use prestressed grouted nails for stabilizing the slope faces. This paper discusses the results from different parametric studies and the results obtained. The implications of different possible alternatives are discussed in the paper.

1.

INTRODUCTION

The use of soil nailing is a new and innovative technique that can be used even in very difficult site conditions. It involves in driving in steel rods into the soil beyond the rupture plane to bind the active soil mass to the interior mass. This technique falls under the category of earth reinforcement and consists of ive bars, named nails, which are generally used to retain excavations, and to stabilize unstable slopes. This technique has been widely used in different types of projects, e.g. soil slopes (Nicholson and Boley 1985, Bruce and Jewell 1987), retaining structures (Stocker et al. 1979, Shen et al. 1981, Guillox and Schlosser 1982), tunneling and other civil and industrial projects. Babu and Singh (2009) have described some latest techniques for the design of nailed soil walls and their verification by finite element based numerical methods. Gosavi et al. (2009) have described some analysis method for nailed soil cuts and their verification by laboratory model tests. The design and construction aspects of soil nailing for stabilizing of slopes and excavations was presented by Elias and Juran (1991). In this paper, the response of the nailed soil walls is studied through different methods, viz. limit equilibrium method based on bi-linear wedge analysis, laboratory model tests and finite element analyses. The results are analysed for comparing the accuracy of different methods and the implications for the field performance. Later, one case study of a real field case is described in the paper.

3

2.

BI-LINEAR WEDGE ANALYSIS METHOD

This limit equilibrium analysis method assumes the failure surface to consist of two surfaces intersecting within the soil mass. The reinforcement elements are assumed to develop only tensile forces and the effect of the bending stiffness is ignored in the analysis. The length of the reinforcement element beyond the rupture wedge is defined as the embedment length and the pullout resistance is calculated based on this length of reinforcement only. These tensile forces are divided into the components parallel and perpendicular to the assumed rupture plane. The normal force and tangential component in each reinforcement element crossing the potential surface are added to the resisting forces mobilised in the soil when determining the factor of safety of the entire mass. The factor of safety of the soil mass is computed as the ratio between the available and the required forces. The above calculations are performed repeatedly for different failure surfaces using a computer programme developed in C++ language. The computer program is interactive and reads the input data through the console and the results are saved in a file. The output from the program consists of the factors of safety at each and every rupture plane considered in the analysis. The program also gives the minimum factor of safety for the given configuration. 3. EXPERIMENTAL PROGRAMME All the retaining wall tests were performed in a specially constructed retaining wall test facility of 2000 mm long, 740 mm wide and 2000 mm height. This facility consists of a mass concrete base slab of 100 mm thickness and two longitudinal side walls that are anchored into the base slab. The side walls of the facility were constructed with hollow concrete blocks of size 400 mm × 200 mm × 200 mm. These blocks were anchored into the foundation soil through tie rods of 12.5 mm extending through the concrete floor. During the tests, steel channels (ISMC 200) were fixed across the two side walls to create a self-reacting rigid frame consisting of the base slab, side walls and the steel channels as illustrated in Figure 1. The test set up is constructed 1.0 m below ground level and 1.0 m above the ground level for ease of handling large quantities of soil during the tests and also from the stability consideration of the two side walls. A free space of 500 mm was left in front of these walls for construction of the test wall and for access to provide necessary instrumentation. The backfill of the model retaining walls was subjected to uniform surcharge pressure by inflating an air bag placed between the backfill surface and the cross-channels connected between the two side walls. The top of the air bag was covered with a non-woven geotextile and a 25 mm thick plywood plank was placed between the geotextile and the steel channels to distribute the load uniformly on the soil. This air bag was specially made of double faced neoprene coated nylon reinforced fabric. An air compressor capable of supplying air at a maximum pressure of 1000 kPa was run continuously during the test to supply pressurised air. The air is routed through a non-return valve. The regulator was adjusted to control the pressure of the air in the air bag to apply the pressure in small increments.

44

FGE 2009, Bangalore

The lateral displacements of the wall facing were measured through LVDTs that were connected to display unit to directly read the displacements. The vertical and lateral soil pressures were measured using strain gauge type pressure cells of 300 kPa capacity that have an accuracy of 0.1 kPa. The loads developed in the soil nails at the wall facing were measured through pre-calibrated load rings fabricated within the geotechnical engineering laboratory of IIT Madras. These load rings were made of 45 mm wide steel pipes (85 mm internal diameter and 2.5 mm wall thickness) fixed with resistance type strain gauges on diametrically opposite sides. Strain gauges were also fixed at four locations on the soil nails to measure the strains along the length of the nails. These strain gauges were protected from damage during the installation by placing them in side a thin plastic membrane wound around the soil nail. 3.1 Construction of the Model Retaining Walls The soil used in the present investigation was a dry river sand with coefficient of uniformity (Cu) of 1.96, coefficient of curvature (Cc) of 1.05 and effective size of particle (D10) of 0.25 mm. The soil can be classified as poorly graded sand with the letter symbol SP according to the Indian soil classification system. The minimum and maximum dry unit weights of the soil are 14.97 and 18.5 kN/m3. The direct shear tests have given the peak friction angle of the soil as 41°. reaction frame wooden plank and nonwoven geotextiles 0.2 m

inflatable air pressure bag

facing LVDT

2.0 m

load ring

sand

G.L pressure cells strain gauge

2.0 m Figure 1. Set up for model retaining wall tests The model walls were then built in the retaining wall test facility with the same configuration to evaluate the accuracy of the limit equilibrium method of analysis of this type of earth structures. All

45

FGE 2009, Bangalore

the tests were performed with a model wall height of 1800 mm. This has left a clear gap of 200 mm to accommodate the air bag between the surface of the backfill and the cross channels to apply the uniform surcharge pressure. The front facing was made of individual facing blocks (8 rows and 3 columns). The soil nails used in these test walls consisted of 12.5 mm diameter plain mild steel rods. The length of the soil nails was varied from 600 mm to 1600 mm. The number of soil nails also was varied from 9 to 24. All the tests were conducted at a relative density of the backfill soil of 70%, which corresponds to a peak friction angle of 41°. The backfill sand was placed in the tank in layers of 225 mm thick. For, each lift, the amount of soil needed to produce the desired dry density (at relative density of 70%) was weighed out and placed loosely in the tank. Each layer was then compacted manually using a drop weight of 65 N falling on a wooden plank of plan size 250 mm × 600 mm from a height of 350 mm. This procedure of soil compaction was arrived at by trials to achieve a uniform soil density throughout the test tank. The compacted density of the sand was monitored by collecting samples in steel cups of 90 cm3 volume placed at different depths in the sand bed. The difference in the density of the soil at different locations was found to be less than 1%. The front facing units were temporarily ed laterally using cement blocks. After the backfill soil was placed to the full height of the model wall, the bricks in each layer were successively removed after the nails in that layer were installed. The nails were installed by driving. A template was used to guide the inclined nails at 10° below the horizontal. After the nails were driven in, the facing s were connected to the nails through pre-calibrated load rings as shown in Figure 1. This method of construction of the test walls corresponds to the top-down construction technique adopted for slope stabilisation works in the field. 3.2

Test Procedure

The constructed walls were subjected to failure by applying uniform surcharge pressure on the backfill. The surcharge pressure was applied in small increments until the failure of the wall by controlling the flow of air through the non-return valve. The applied surcharge pressure was monitored through the vertical pressures read from the pressure cell fixed at the back end of the wall. Each increment of surcharge was kept constant until the deformations under that particular load increment have reached a steady state. As the surcharge pressure was increased, the lateral deformations and the load in the load rings were observed to increase. As the failure pressure was approached, the load in the load ring connected to the upper row of the soil nails became constant and had suddenly decreased at some surcharge indicating the pullout failure of the nails at that elevation. The lateral earth pressures measured behind the facing s also had shown a similar trend. At that stage, the walls were deemed to have failed due to the pullout of the soil nails. The top row of nails became so loose that they could be pulled out by hand. The test was stopped at that stage and the entire test set up was dismantled. Some of the initial tests were repeated again with the same nail configuration to ascertain the repeatability of the test results. Very small differences were obtained in the results obtained from such tests. The tests were repeated with different nail configurations to examine the influence of the length, inclination and spacing of nails.

46

FGE 2009, Bangalore

4. FINITE ELEMENT ANALYSIS OF TEST WALLS All the finite element analyses in this research work were carried out using the general purpose finite element program GEOFEM developed at IIT Madras, Rajagopal (1995). The behaviour of retaining walls ed by soil nails was analysed by two dimensional finite elemental analysis using plane strain idealisation by smearing the effect of soil nails over 1 m width perpendicular to the plane of the analysis. The influence of some of the important parameters which influence the behaviour of these walls was studied through parametric studies. In order to compare the finite element results with those from experimental work, the exact dimensions used in the laboratory experiments were employed in these analyses. The finite element mesh used for the analysis of nailed walls on sand consisted of eight-node quadrilateral elements to represent the soil and the front facing units, 3-node bar elements to represent the soil nails and six-node t elements of zero thickness to represent the interfaces between the front facing and soil and between the nails and soil. The elastic-plastic behaviour and the failure of the soil were simulated by using Mohr-Coulomb yield criterion. The data from the pullout tests was used to estimate the interface friction angle between the soil nails and the soil. The interface elements consist of two stiffness values, one in the shear direction and the other in the normal direction. The shear stiffness was given a large value to maintain the compatibility between the nail and soil before the failure and was set to a small after the failure to allow for relative deformation. The stiffness in the normal direction was set to a large value to maintain the in the normal direction at all stages of the analysis. The interface element is not allowed to develop tensile stresses in the normal direction. The construction of the wall was simulated by placing the soil in the mesh and applying the self weight as gravity forces with a Ko of 0.50. During this stage, the facing s were laterally ed by external props. The resulting deformations and strains were set to zero at the end of this stage. During the next stage of analysis, the front props were removed and uniform pressure was applied on the backfill surface in small increments to simulate the pressure loading in the laboratory tests. The analysis was iterated to reduce the out-of-balance forces at each load step to less than 0.5%. The failure of the wall was defined at the stage at which the out-of-balance force increases to a high value and divergence takes place in the solution. The finite element analyses were performed with nail lengths of 0.6, 0.8 and 1.2 m, and with nail inclinations of 0°, 10° and 20°. 5. COMPARISON OF DIFFERENT RESULTS Typical lateral deformations along the height of the wall at different surcharge pressures are shown in Figure 2 for 1.2 m long horizontal nail. The forces measured in the load rings were found to increase with increase in surcharge pressures, Figure 3. At the ultimate surcharge pressures the top row of facing s deformed excessively accompanied by the sudden drop of force in the load ring. At that stage, the top row of nails became so loose that they could be pulled out by hand easily. This behaviour shows that the top row of soil nails had failed in pullout. The lateral pressures measured behind the facing units also showed a linear increase with surcharge pressure until failure and showed a sudden decrease at failure. In general, the walls with inclined nails failed at higher surcharge pressures because of the higher pullout capacity of these nails. In all the walls, the lateral wall displacements have decreased towards the

47

FGE 2009, Bangalore

toe of the walls. This is because the length of embedded portion (where the nail resistance is developed) beyond the rupture plane increases towards the toe of the wall. As the deformations are lower towards the toe of the wall, the relative deformations between the soil and nail are also lower and hence the mobilised force in the nail is lower in the bottom row of nails as shown by the results in Figure 3. 2000

elevation (mm)

1500

applied pressure (kPa) 10

1000

20 30 40

500

50 60

0 5

15 10 lateral displacement (mm)

0

20

Figure 2 Lateral displacements at different surcharge pressures (1.2 m long horizontal nail) 2000

elevation (mm)

force in load rings (N)

1687.5 1237.5

1500

1012.5 562.5

1000

112.5

500

0 0

20

60 40 surcharge pressure (kPa)

80

Figure 3. Load ring forces in soil nailed wall (1.2 m horizontal nail) The measured lateral earth pressures behind the facings s are shown in Figure 4. It may be observed that the lateral earth pressures have also increased with the applied surcharge pressure until

48

FGE 2009, Bangalore

the collapse state of the model walls. When the failure pressure was approached, the facing s had excessive lateral outward deformations leading to a decrease in the lateral earth pressures in the top most pressure cell. The reduction of earth pressures was taken as one indication of the failure of the walls. The comparison of the internal rupture plane measured in the laboratory tests and those predicted by the finite element analysis and the bilinear wedge mechanism is shown in Figure 4. The points with the peak strain in the nails were ed together to obtain the possible rupture plane in both FEA and experiments while the surface with the least factor of safety is taken as the rupture plane in the case of limit equilibrium analysis. It could be seen that the rupture plane predicted by the bilinear wedge method is widely different from the measured surface. This may be another reason for the low ultimate pressures estimated by the limit equilibrium method. The responses predicted by the finite element analyses with 0.6 m long nails placed at three different inclinations are shown in Figure 5. These responses closely matched those measured in the laboratory tests. It could be observed that the deformations decrease (at the same surcharge pressure) with increase in the nail inclination as also noted from the experimental results. This result clearly shows that the finite element model is able to replicate the improvement in the performance of nailed soil retaining walls with the inclination of the nails. The comparison between the measured ultimate surcharge pressures on the soil nailed model walls and the corresponding pressures predicted by the limit equilibrium and the finite element methods is presented in Table 1. It can be observed that the pressures estimated from the limit equilibrium method are lower than the measured values in all the cases while the values predicted by the finite element analyses are reasonably close to the measured values. One of the reasons for the discrepancy between the predicted and the estimated results could be the effect of sidewall friction. 2.0

elevation (m)

1.5

1.0

fem

0.5

exp design method 0.0 0.50 0.0 1.0 points of maximum strain locations over the length of the nail (L=1.2 m)

Figure 4. Internal rupture surfaces from different methods

49

FGE 2009, Bangalore

Table 1. Comparison of ultimate surcharge pressures on retaining wall Configuration of soil nails

Limit Equilibrium analysis (kPa)

Laboratory FE analysis (kPa) model tests (kPa)

1.6 m (horizontal)

100

130

125

1.2 m (horizontal)

50

60

55

1.2 m (10° inclination) 65

80

85

0.8 m (horizontal)

35

45

40


60

55

0.6 m (horizontal)

25

35

30


45

45

80

pressure (kPa)

60

40

20o inclination

20

10o inclination horizontal 0 0

10

30 20 displacement (mm)

40

Figure 5. FE predicted retaining wall behaviour with different nail inclinations 6.

CASE STUDY OF A PRE-STRESSED NAILED EXCAVATION

At one construction site in North Chennai, it is required to make deep excavations for construction of two basement floors of a shopping complex. The proposed construction site is surrounded by one to two storey residential and commercial buildings. Due to heavy rain fall in the area, the surface rain water collects into the construction area. Hence, dewatering system had to be installed to keep the construction area dry.

50

FGE 2009, Bangalore

In general, the soil profile at the site consists of four distinct layers. The top one meter or so from the ground level is loose silty reddish brown fine to medium sand. This upper soil layer is followed by a three meter thick medium dense to dense brown silty sand of fine to medium grade with average SPT values of 20. This second layer is underlain by a firm to medium dense brown silty sand of 3 m thickness with average SPT values of around 10. The soil layer below consists of silty sand of approximate thickness 6 m thick extending up to 13 m. This layer is a medium dense to dense and is brownish in colour with an average SPT of 30. This layer is followed by dense to very dense, hard silty sand with average SPT values in excess of 50. The ground water table at the site is around 3 to 4 m from the ground level as per the local information. However due to effective dewatering at the site, the water table is maintained as a depth lower than 8 to 9m below the ground level i.e. at one to two meters below the excavated ground surface. Assumed profiles and proposed slope is indicated in Figure 6.

Figure 6. Schematic sketch of the proposed excavation and ground

The factor of safety of the slope without any stabilization comes to 0.63 which is unacceptable. The factor of safety of the slope was increased to 1.20 by providing pre-stressed soil nails as illustrated above. This factor of safety was adequate as the excavated slopes are only of temporary nature. The details of the soil nails provided at the site are shown in Table 2. The factors of safety obtained from finite element analyses are slightly larger than those obtained from these analyses. Table 2. Details of the pre-stressed soil nails at different depths Row name Location of Force in Free length Required soil nails kN/m (m) fixed length (m) A -1.0m 10 3.0 1.5 B -2.5m 20 2.0 3.0 C -4.0m 25 1.5 3.75 D -5.5m 30 1.0 4.5 E -7.0m 40 0.0 4.2

51

Total required length (m) 4.5 5.0 5.25 5.5 4.2

Provided length (m) 6.0 6.0 6.0 6.0 6.0

FGE 2009, Bangalore

The slope stability analysis was performed using a limit equilibrium based slope stability software “DC slope” from DC Foundations, . A uniform surcharge of 20 kPa was assumed at the ground level to take of the loads from moving machinery etc. The results from these analyses were independently verified from finite element analysis using the program PLAXIS-2d. Typical result from one analysis of limit equilibrium analysis is shown in Figure 7.

Figure 7. Typical result from program DC slope The surface was treated by spraying 50 mm thick concrete to prevent surface erosion. Some typical pictures from the site are shown below to illustrate the nail pre-tensioning and the beam arrangement.

Figure 8. General view of the construction site

52

FGE 2009, Bangalore

Figure 9. Pre-tensioning of the soil nail after the grouting 7.

CONCLUSIONS

Any design of soil structure is complicated due to several unknown parameters. Hence empirical methods are employed to come out with safe designs. Invariably, these are based on earlier experience or on some limit equilibrium analysis. The current study has shown that the stability of the nailed wall or excavation is very much influenced by the length and inclination of the soil nails and pre-tension effects.. Some of the major conclusions that can be drawn from this study are as follows: • The design based on the bilinear wedge failure mechanism is always on the conservative side. • The length and inclination of the soil nail plays significant role in the overall behaviour of nailed soil wall. • The nail should project sufficiently beyond the rupture plane to develop significant pullout capacity and also to intercept the potential rupture planes. • The internal rupture plane in the soil nailed walls can be idealised as a bilinear surface and can be obtained by drawing a vertical line at 0.3H behind the facing and an inclined line through the toe at (45+φ/2)°. • The influence of the nail inclination is very well represented by the finite element models. ACKNOWLEDGEMENTS The financial for this work from the Ministry of Human Resources and Development, Government of India, New Delhi is gratefully acknowledged. Some part of the work reported in this paper was performed by Mr. G.V. Ramesh as part of his thesis work for Master of Science degree by research at IIT Madras. RFERENCES Babu, G.L.S. and Singh, Vikas Pratap (2009) Appraisal of Soil Nailing Design, Indian Geotechnical Journal, Vol. 39, No. 1, 81-95. Bruce, D.A. and R.A. Jewell (1987) Soil nailing: application and practice – part 2. Ground Engineering, The journal of British Geotechnical society, Vol. 20, No.1, 21- 28.

53

FGE 2009, Bangalore

Elias, V. and Juran, I. (1991) Soil Nailing for Stabilisation of Highway Slopes and Excavations, FHWA-RD-89-198, Federal Highway istration, McLean, Virginia, USA. Gosavi, M., Swami Saran and Mittal, S. (2009) Model Tests for validation of analysis of nailed cuts, Indian Geotechnical Journal, Vol. 39, No. 1, 96-115. Guilloox, A. and G.Notte (1982) Experiences on a Retaining structures by Nailing. Proceedings of 8th European Conference on soil Mechanics and Foundation Engineering, Helsinki. Nicholson, P.J. and D.L. Boley (1985) Soil nailing s excavation. Journal of Geotechnincal Engineering, ASCE, 111, No. 4. 45-47. Rajagopal, K. (1998) s Manual for the Finite Element Program GEOFEM, Geotechnical Engineering Division, Department of Civil Engineering, IIT Madras. Shen, C.K., L.R. Herrman, K.M. Romstad, S.Bang, Y.S. Kim and J.S. De Natale (1981) An In Situ Earth reinforcement Lateral system. Department of Civil Engineering, Univ. of California, Davis, Report No. 81-03, U.S. National Science Foundation Report No. NSF/CEE81059, March, 130 pp. Stocker, M.F. and G.Riedinger (1990) The Bearing Behaviour of Nailed Retaining Structures, Design and performance of Earth retaining Structures. ASCE Geotechnical Special publications, No.25, New York, 612-628.

54

FGE 2009, Bangalore

Case Study 10

THE FAILURE ANALYSIS AND REHABILITATION OF DAMAGED RE WALL Murthy, B. R. S. Retd. Professor, IISc, Bangalore, India, 560 012, [email protected]

1.

BACK GROUND

A Geo-grid Reinforced earth retaining wall for a fly-over along the National High way No 4 was constructed a few years back. The wall height is varying from 2 to 13 meters. At a location of about 7 meter height of the wall there was snapping of connection between facing element and the geo-grid reinforcement. The authorities requested the author to visit the site and suggest a suitable permanent solution for rehabilitation of the failed zone. Also it was requested to find out the reason for failure. The site was inspected along with all concerned. There was not much information available with the project authorities regarding the quality control or the designs not even a sketch of the actual implementation.

2.

SITE CONDITION

The failure had taken place over a length of about 10 meters. The failure was essentially in the form of collapse of the s due to snapping of the connecting plastic cleats/clamps, perhaps initially of the upper layers. The number of geo-grid reinforcement layers provided was found to be reasonable. There was variation in the grade of geo-grid increasing with depth. The soil appeared to be in good condition as required from an RE wall point of view. In fact the failure plane had not gone beyond a meter from the wall face which was the boundary of the drainage zone of granular material. The soil behind the failed zone was standing near vertical indicating the sufficiency of the reinforcement layers and good quality of compaction. It was reported that the PHI value for the soil was found to be more than 34 degrees. In such a situation it was difficult to evaluate the reasons for failure. A few theories were proposed by engineers and local people for the cause of failure. One particular theory was that a heavily loaded truck rammed against the crash barrier. Really no such sign was visible. All the theories were discounted and it was indicative that most of the reasons were only to divert the attention and hence search started for a realistic one. 3.

INVESTIGATION

A closer examination of the detailing and the site condition of failure indicated that the failure has initiated from snapping of upper layers of geo-grid/ connections. This was indicative from large elongations of the upper layers. Further the crash barrier and

11

the friction slab are rigid and continuous where as the pavement beyond the friction slab is flexible. The integrity between them cannot be established by a thin bitumen pavement on the top. The wheel loads will transmit the load and vibrations differently in the two systems. In rigid pavement portion the vibrations are transmitted down into the granular medium whereas the same is damped/ dissipated in the flexible pavement. The transmitted vibration will tend to compact the granular soil resulting in the loss of of the friction slab and soil. This will in turn makes the friction slab to transmit the reaction due to the wheel load to the facing element. As the facing element assembly is not designed to transmit the axial loads the reaction will tend the assembly of facing element to BUCKLE causing the snapping of the cleats/ ts between the facing and the geo-grid. This is the primary reason for the failure. Figures A and B indicate the schematic representation of the wall and details of the analysis of failure.

4.

REHABILITATION SCHEME

To reconstruct the wall, the following scheme was proposed and implemented and the same has been presented in the drawings. The essential steps were as follows. 1. the failed surface with sand bags and provide to the exposed crash barrier using wall form jacks. 2. Providing a suitable staging, remove the sand bags to a depth of 1.65 meters. Trim the failure plane to remove loose materials. 3. Drive 6 meter long 20 TOR bars at a spacing of 550 mm c/c (three rows) in both directions. 4. Fasten 50x50x2.8 mm MS grid reinforcement vertically along a defined line which can facilitate fixing of the permanent RE wall s and horizontally along the lowest layer of nails. 5. Fill the void space beyond the MS grid with 50 mm aggregates as densely as possible. 6. Provide 50mm thick shot-crete on the weld mesh as per relevant IS code. 7. Repeat the steps 2, 3, 4, 5 and 6 to reach the full depth. 8. Install the RCC pre-cast facing for one row and fasten the same suitably with the projected nails. 9. Back fill the gap between the shot-creted surface and the Pre-cast with self compacting coarse sand. 10. Repeat the steps 8 and 9 to reach the top level.

12

FGE 2009, Bangalore

CRASH BARRIER WITH FRICTION SLAB (RIGID)

FLEXIBLE PAVEMENT BITUMIN WEARING SURFACE

GEOGRID LAYERS RCC FACING

GRANNULAR DRAINAGE MEDIUM about 1 meter thick

FIGURE A

13

FGE 2009, Bangalore

CRASH BARRIER WITH FRICTION SLAB (RIGID)

FLEXIBLE PAVEMENT

BITUMIN WEARING SURFACE

REACTIONS

RCC FACING

GEOGRID LAYERS

GRANNULAR DRAINAGE MEDIUM about 1 meter thick

FIGURE B

14

FGE 2009, Bangalore

ROAD SURFACE

SAND BAGS TO A SLOPE OF 1:1

WALL FORM JACKS TO THE CRASH BARRIER

FAILED SOIL SURFACE

Figure1. The failed reach is protected using sand bags to a slope of 1:1 and ing the crash barrier with wall form pipe jacks

15

FGE 2009, Bangalore

20 TOR BARS OF LENGTH 0.7H DRIVEN AT SPACING OF 550 C/C IN BOTH THE DIRECTIONS

Figure 2. The sand bags are removed to a depth of 1.65 meters and three rows of nails are driven in to the back fill material.

16

FGE 2009, Bangalore

50x50x2.8 mm MS WELD MESH FASTENED TO THE NAILS AND VOID SPACE BEHIND FILLED CAREFULLY WITH AGGREGATES 50 mm THICK SHOTCRETE AS PER IS 9812

Figure 3. MS weld mesh of size 50x50x2.8 mm is fastened to the nails and at the top of the sand bags. The gap between the soil and the weld mesh is carefully packed with aggregates of suitable size. 50mm thick shot-crete is provided on the weld mesh. Repeat the procedure of removal of sand bags, driving of nails, fastening of weld mesh, backfilling with aggregate and shot-creting to reach the full depth of the wall as indicated in Figures 4 and 5.

17

FGE 2009, Bangalore

Figure 4

18

FGE 2009, Bangalore

Figure 5

19

FGE 2009, Bangalore

PRECAST RCC FACING S ASSEMBLED FROM BOTTOM

THE FACING IS CONNECTED TO THE NAIL SYSTEM AND THE GAP IS BACK FILLED WITH SAND

Figure 6 The pre-cast RCC facing s as per the original designs shall be assembled along the original alignment starting from the ground level. The s are connected to the nails using suitable MS cleats and fastening system. The gap between the shot-crete surface and the RCC pre-cast shall be backfilled with good self compacting material like sand. The above sequence shall be followed till the full height is covered.

20

FGE 2009, Bangalore

INTERACTION BETWEEN APPROACH EMBANKMENT ON SOFT GROUND AND ROAD OVER BRIDGE FOUNDATIONS – A CASE STUDY M.R.Madhav1, B.Vidyaranya2 & G.B. Rajendra Prasad3 Professor Emeritus, 2Research Scholar & 3Asst. Executive Engineer 1 JNTU College of Engineering, 2Osmania University, Hyderabad & 3R&B Amalapuram Division [email protected] & [email protected] 1

ABSTRACT The paper presents an interesting instance of approach embankment construction affecting the already constructed Road Over Bridge (ROB) and its foundations. The ground consists of very thick layer of very soft to soft soil extending to a depth of nearly 15 to 16 m underlain by dense sands. As such, the bridge was constructed on well foundations sunk into strong and stiff underlying strata. In spite of this fact, the bridge suffered extensive damage due to the construction of nearly 4.5 m high embankment. The lateral squeezing of the soil is believed to be the cause of the failure/damage to bridge foundations. The case history highlights the need for an integrated approach for the design and construction of a bridge and its approaches.

lined with concrete s to the surrounding soil. Even the s moved in to the canal by about half a meter.

1. INTRODUCTION The design of structures on or over soft soils needs special care and concern. The interaction between structures adjacent to each other is an important aspect that should not be ignored. The paper presents a case study of a Road over Bridge which was constructed prior to the construction of the approach embankments. The embankment was constructed subsequently. However, the effect of possible squeezing of soft soil in the longitudinal direction was not considered leading to a major geotechnical problem. The paper presents the details of the bridge, damage caused to it and the results of soil investigations at the site.

The plan of ROB is shown in Fig. 1. The ROB consists of a single span of 16.0 m long RCC ‘T’ beam slab. The abutments were founded on well foundations two under each. The wells, 4.50 m in diameter, were plugged at depth between 18.30 m to 18.60 m below the ground level. Wing and return walls (Fig. 2) were made in Mass CC M-20 grade up to ground level and Mass CC M-30 grade above ground level with 40 mm grade hard granite stone aggregate.

2. DESCRIPTION OF THE SITE The site under investigation is on a by road constructed to facilitate free movement of traffic over a railway line and avoid the use of a level crossing. The proposed road connecting two towns es over an irrigation canal and a railway track approximately 400 meters apart. The irrigation canal runs parallel to the railway track. An embankment of height 6.5 m was to be constructed to connect the ROB and the bridge crossing the canal. The site is surrounded by paddy fields all-round. The ground water table was at a depth of about 1.5 m below the existing ground level at the time of investigation but is known to rise close to GL following monsoon rains.

Figure 1 Plan of the Road over Bridge

A road runs parallel to the canal on the side away from the ROB. The bridge structure over the canal was founded on a raft 15 m wide. The bridge settled by about 0.5 to 1.0 m and the piers tilted leading to severe cracking of the deck, piers and the abutments, the bottom of the canal heaved due to the construction of the approach embankment to the bridge on the canal. The sides of canal ing the road were

Fig 2 Section of Return Wall

6

Foundation provided up to 4.0m below ground level consisted of 2.0 m sand filling well compacted in layers of 300 mm and RCC raft foundation of 0.50 m followed by 1.50 m sub-structure with 1:1 slope on one side and straight face on other side. Figs 1 through 4 show the plan, sectional details of the abutments and the return walls in detail.

block. The crack had widened to a width of about 25 mm since then. The top of the abutment moved by about 15 to 20 mm towards the backfill. Earthwork on the rear side of the abutment No. 2 was carried out to a height of 4.4 m from the well cap level. Horizontal cracks 25 mm wide were noticed at 3.60 m 4.80 m and 6.0 m from the top of the slab on wing walls No.4. Figs 5 through 10 shows the tilt generated in the wing walls due to movements caused by embankment construction and Table 1 summarize the movements reported.

Figure 5 Sectional Elevation of Abutment with wing walls

Figure 3 Section of Abutment

Figure 6 Sectional Elevation of Wing Wall 1

Figure 4 Section of Wing Wall The embankment was constructed using locally available soil. A layer of geogrid was laid at the bottom of the fill. The soil used for the embankment appeared to be heavy clay, possibly with slightly higher percentage of fine sand and silt. Sandy soils were used for the central core while clay was used for the remaining portion of the backfill. The wing walls got settled and heavy cracks were generated. The construction of the bridge was completed in Jan’ 01 but the construction of the embankment was delayed till year 2005. The work was recommenced with work on Abutment No.1 side to a height 6 m from the top of the well cap and to 1.2 m from the bottom of the bed block. A crack was observed at a level of 1.20 m below the top of bed


7

FGE 2009, Bangalore

Table 1 Details of Movements of Abutments Settlement at top

Lean Toward Fill

W1

270

340

W2

400

290

W3

Nil No formation

75

Loca tion

Figure 8 Sectional elevation of Abutment with Wing Walls

W4

120

Horizontal Movements Well Bed Cap Block 120 20 Butting 100 against abutment 75 40 20

Nil

3. SITE INVESTIGATIONS The site was investigated by several agencies prior to the construction of the bridge. The soft clay layer was reported to extend up to a depth of 18.5 m. The water table was close to GL. The site conditions needed fresh investigations to evaluate the soil properties and to examine the reasons for the movements. Fresh boreholes were advanced to a depth a 25- 30 m at three locations close to the ROB structure. The soil profile indicated black soft clay up to 15 to 18 m depth. Medium to fine sand underlies the soft clay. SPT ‘N’ values were negligibly low and some times equal to zero. The variations of the SPT values with depth are shown in Fig. 12. SPT N values were slightly high (4 to 5) near the top (2-3 m) possibly due to desiccation but reduce to 2 to 3 until a depth of 15 in BH.1, 16 m in BH.2 and 18.0 m in BH.3. N values increased to about 20 to 27 at 18 to 20 m depth range. N values in the lower sand stratum were greater than 40.


The liquid limit values ranged from 72 to 85 while the plasticity index values were of order 45 to 55. The natural moisture content increases with depth over the top 4.0 to 5.0 m and lie in the range of 75 to 90. The liquidity index values are greater than 1.0 at few locations. The shrinkage limit values measured over selected depths are 8.0 for depths less than 11.0 m and about 10 at a depth of 16.0 m. Low shrinkage limit and high plasticity index attest to the fact that the soil is of expansive type and the predominant clay mineral could be montmorillonite. The soils at the site are within the flood plains of the river Godavari. The river over a major part of its length traverses through the Deccan Plateau wherein basaltic formations exist. These rocks normally weather in to expansive clay, locally called Black Cotton clay. 4. IN SITU TESTING Standard Penetration Tests were conducted in the three boreholes. Results from other agencies indicate N values of 2 to 3 up to a depth of 12.5 m and increase to about 10 to 12 at 17 to 18 m. The N values


8

FGE 2009, Bangalore

in the sand stratum were in excess of 30. Another investigation conducted in 2001 also indicated N values of 45 to 60 in the depth range of 19 to 22.

but reduce to 2 to 3 until a depth of 15 in BH.1, 16 m in BH.2 and 18.0 m in BH.3. N values increase to about 20 to 27 at 18 to 20 m depth range. N values in the lower sand stratum were greater than 40. Fig. 12 shows the variation of the SPT values conducted by various consulting agencies.

The investigations carried out by JNTU corroborate the above studies. SPT N values were slightly high (4 to 5) near the top (2-3 m) possibly due to desiccation JNTU 0

10

20

N 30

Professional Agency 1 40

50

60

0

10

20

30

40

50

0

60

0

0

5

5

5

10

10

D epth, Z

15

20

N

40

60

Depth, Z

0

10 Depth, z

Professional Agency 2

N

15

15

20

20

25

25

30

30

20

25

Series1

.

30 BH - 1

BH - 2

Series2

BH - 3

Fig. 12 Comparison of SPT N Values from Different Agencies 5. RESULTS AND DISCUSSION 0

20

Water Content 40 60

80

100

0

The laboratory tests were conducted on the undisturbed soil samples to determine the Atterberg limits, the natural moisture content and the undrained strength of the soil at various depths. Fig 13 shows the variation of the moisture content with depth for the three boreholes investigated. The water content is close to 60% in the top 3 to 4 m depth and increase to 80% in the soft clay layer. The values reduce to about 20 to 25 at depth 20 m within the sandy soils underlying the soft clay. Fig. 14 shows the variation of the Liquidity index with depth for the three boreholes investigated.

5

10

Depth

15

20

25

30

35

BH-1

BH-2

BH-3

Figure 13 Variation of Natural Moisture Content with Depth.

9

FGE 2009, Bangalore

-0.5

LI

0.5

1.5

2.5

0

z

10

20

30

BH-1

BH-2

BH-3

Figure 14 Variation of Liquidity Index with Depth The displaced soil induced lateral pressures on the surrounding structure as a result of which the wing walls settled and cracked. The damage to the ROB structure is severe due to large lateral pressures generated by the embankment fill. 6. ANALYSIS The undrained strength of the in situ soft soil is around 10 to 15 kPa. The ultimate bearing capacity of the embankment would be of the order of 55 to 65 kPa. An embankment of 4.5 m height generated a vertical stress of 83 kPa (=4.5x18.5 kN/m3). The construction of the embankment induced plastic flow and failure of the soft soil. Since the soil is highly plastic clay and has low bearing capacity, the soil was squeezed laterally due to embankment loading. The lateral movements were less across the embankment possibly because of the berms and basal reinforcement provided. However, there was no such restraint in the longitudinal direction towards the wing walls and the abutments. Hence the substructure moved significantly. 7. CONCLUSIONS The case study illustrates an interesting interaction between an already constructed bridge abutments and embankment constructed subsequently by another agency with no communication between the two. The stability of the embankment was ensured in the transverse direction by the provision of berms and basal geogrid reinforcement. The interaction between the embankment and the abutment in the longitudinal direction was not considered resulting in large unacceptable movements of the bridge structure rendering it dysfunctional.

10

FGE 2009, Bangalore

Case Study 12 AN INVESTIGATION INTO THE COLLAPSE OF A PORTION OF SOIL NAILED WALL Prof. T.G. Sitharam, Professor of Geotechnical Engineering, and Chairman, Centre for infrastructure, Sustainable Transportation and Urban Planning Indian Institute of Science, Bangalore – 560012. Email: [email protected]

SUMMARY The excavation for a site, which has three basements, the soil nailing has been adopted for retaining the earth so as to have a vertical cut temporarily until a permanent concrete retaining wall is constructed. On March 12th, west side of the soil nailed wall was collapsed (see Fig 2) for a length of 20m and road on the adjacent property was also damaged. A temporary restoration scheme (See Fig 4) consisting of earthen embankment fill and sand bags fill was already adopted as an immediate measure. Geotechnical report clearly indicates that the top soil (is about 0.5m) and lateritic red soil (up to a depth of 2.75m to 3.5m) from the ground level. Very close to the failure area, the lateritic red soil exists up to 3m depth below which disintegrated rock is present. Soil gradation also clearly indicates presence of uniform sand (less than 2mm particle size), silt and clay sizes non presence of any gravels or boulders in the area. N values are invariably greater than 15 in insitu condition. Driven soil nailed walls have been constructed as a temporary measure all along the boundary to retain the earth for a depth of about 10m. 20mm dia TOR steel nails of length 5m long at spacing of 450 mm c/c were driven in to the ground as shown in cross section. 50mm thick shotcrete of 1:2:2 grout with 50x50x2.6mm wire mesh was adopted (see Fig 2 for the final finished wall surface). Nailing was done in stages of 1.5m deep at every stage. This was reported to be repeated up to a depth of hard rock. Typical cross section of the soil nailed wall is shown below in fig 1. The soil nailing has been reported to be completed on 14th October 2008. On March 12th, west side of the soil nailed wall was collapsed for a length of 20m and road on the adjacent property was also damaged. The typical diagram clearly indicates that the failure (yellow colored area) is only in the top 3m of the soil mass which corroborates well will the geotechnical borehole data. The fig 2 clearly indicates there is no failure below 3m from the ground level (Ground RL is at 99.85 m). The failure surface does not through the disintegrated rock mass. Figure 3 shows the collapsed area in plan. 39

Fig 1. Cross Section of the soil nailed wall and Plan view

4

Fig 2. Typical Cross-section of the failed slope

Fig 3. View of the soil nailed wall after collapse (left fig) and restored slope with sand bags (right photo) 40

FGE 2009, Bangalore

Fig 4. Temporary scheme adopted Considering the adopted temporary restoration scheme with embankment (see photograph in Figure 3 and Fig 4), the non-availability of working space on the west side road, and noise/vibration due to piling, it was decided to work within the property of the client from inside their land initially to restore the slope.

The following scheme (as shown in Figure 5) is suggested as a remedial measure for permanent restoration of the failed section of the wall (about 20m long). The scheme consists of construction of a retaining wall at the toe of the temporary embankment, grouting of the pervious embankment to take the applied pressures from the bottom approach ramp and grouting of the sand bags/filled up soil in the failed slope section. As a precautionary measure, even the temporary embankment and the adjacent road re-constructed using sand bags has been grouted by drilling holes using clean cement grout, so that it can take the load of the approach ramp uniformly.

41

FGE 2009, Bangalore

The whole area was reported with oozing water on the face of the wall (See photo in Fig 2). This might be due to the puncturing of the live manhole at several points due to soil nailing. Further, the failure is attributed to not providing the deisgned length of the nails in the execution. The actual length of nails provided were shown through analysis shorter than what has been proposed to have a factor of safety less than 1. The present paper highlight the scheme of restoration suggested by the consultant rather than the cause of failure. Consultant was involved only after the failure in the month of April 2009. This presentation describes the scheme of restoration. The scheme involves a retainng wall away from the boundary (inside the clients area) along with a backfill and a retaing wall at the toe of temporary embankment coupled with constructing approach ramp in stages. As this scheme did not cause any disturbance to the neighbouring areas, the scheme was accepted and executed. The disadvantage of the scheme is that the client lost a small area of about 6 to 8 car parks.

Fig 4. The final view of the restored retaining wall with the approach slab

42

FGE 2009, Bangalore

CASE STUDY 13

THE CASE OF A TILTED CHIMNEY – A FORENSIC ANALYSIS RAO.V.V.S. Nagadi Consultants Pvt. Ltd., New Delhi.

THE PROBLEM: A R.C. Chimney of 45m.height having a diameter of 5m. at GL was constructed in a sugar factory in the year 2001. By about June 2002 the chimney was observed to have tilted by 450mm. at the top. To correct the tilt, sand bags were piled up on the diametrically opposite side of the direction of tilt. The chimney regained the original plumb, but it developed wide vertical cracks at three locations on the chimney throughout its height. To develop an appropriate repair solution, detailed investigation was conducted to determine the cause of tilt. DATA COLLECTION: From the design as well as “as constructed” drawings, it was observed that the chimney had been constructed as per designs. Detailed borehole investigations had been conducted and proper design parameters had been used. However, there was no borehole data at the location of the chimney. Hence bore hole investigations were performed around the chimney. The sub-soil profiles were uniform. The profile is shown in fig. 1.

Fig.1.SUB - SOIL PROFILE BH 3 E.G.L

-0.0 FILL

-1.0 -2.0 -3.0

DEPTH IN METRES

-4.0 -5.0 -6.0 -7.0 -8.0 -9.0 -10.0 -11.0 -12.0 -13.0 -14.0 -15.0

……… ……… ……… ……… ……… ……… ……… ……… ……… ……… ……… ……… ……… ……… ……… ……… ……… ……… ……… ……… ……… ……… ……… ……… ……… ……… ………

-5

FILLED UP STRATA (GREY SANDY SILT/SAND WITH BRICK BATS AND DEBRIS)

-4

GREY CLAYEY (4-12%) SANDY (9-22%) SILT (65-82%) WITH GRAVEL (1-10%)

-10

-16

-22

-20

GREY SAND(90-99%) WITH TRACES OF SILT (1-10%) -26

-28

-37

SOIL PROFILE LEGEND :

5 Observed 'N' Values

Water Table

E.G.L.: Existing Ground Level

The foundation raft was founded at a depth of -3m. The soil below consists of medium dense sand.

During the site visit, it was found that a single storey shed had been built very near to the chimney on the tilted side. Upon enquiry, it was told that the shed was constructed after the factory went into operation, without consulting the design or geotechnical consultants. The crack was observed a few weeks later. The cross-section of the construction was as shown in fig.2.

Fig.2.

In plan, the section at GL was as shown in fig.3.

VIEW OF A CRACK

VIEW OF CHIMNEY AND SHED

CONCLUSIONS: The additional loads acting on diametrically opposite ends of the foundation raft resulted in inducing splitting action of the chimney resulting in cracks.

CASE STUDY 14

INFLUENCE OF VIBRATIONS ON INSTALLATION OF BORED PILES Dr. N. Santosh Rao Nagadi Consultants Pvt. Ltd., New Delhi 1.

INTRODUCTION

This is a case study of a forensic geotechnical investigation to assess the possible cause/s which prevented the satisfactory construction of large diameter bored piles using the bailer technique. The investigation involved first identifying the possible causes, carrying out field and laboratory tests where necessary to assess the same, analysis of the test results and finally arriving at the conclusion. 2.

PROJECT DETAILS

The project in this case study is an elevated corridor through the Kishanganj town, Bihar, India, located parallel to the existing road as also the railway line abutting the road. The elevated corridor had been proposed to be constructed in reinforced cement concrete with the deck consisting of box girders ed by two column piers resting on 1m diameter 16m long bored piles. 3.

PROBLEM DEFINED

The piling contractor had been unable to ensure satisfactory construction of the large diameter bored piles using the bailer technique, primarily because of frequent collapse of the sides of the bores made for the purpose of construction of the bored piles. As per the conditions of the contract, bentonite slurry was not to be used in the pile bores and use of liner/casing had also not been envisaged on of the nature of the subsoil as indicated by the report of the geotechnical investigations carried out earlier for this project. 4.

POSSIBLE CAUSES

An evaluation of the above problem indicated that one or both of the following could be the possible causes : 1. Subsoil being unsuitable for construction of the bored piles under the constraints put up by the contract. 2. Vibrations transmitted from the trains moving on the nearby railway tracks and vehicle movement on the adjacent road.

5.

INVESTIGATION OF POSSIBLE CAUSES

a.

Subsoil Conditions

To check whether the subsoil conditions at the site of the proposed elevated corridor are similar to those considered in the design and included in the tender documents, a few boreholes had been carried out at select locations along the length of the elevated corridor. A comparison of the typical subsoil profile indicated by these boreholes with the typical subsoil profile indicated by the geotechnical investigation report included in the tender documents is given below:

Depth (m)

Geotechnical Investigation Report Soil Description

0–5 5 – 10

10 – 15 Silty Fine Sand / Sandy Silt

10 – 20 20 – 25

Range of Nvalues

15 – 20 20 – 25

-

-

New Boreholes Soil Description Fine Sand with Silt Medium to Coarse Sand Coarse Sand Coarse Sand with Gravel

Range of Nvalues 10 – 14 12 – 20 25 – 35 > 50

The above comparison clearly shows that the actual subsoil conditions differ significantly from that indicated by the geotechnical investigation report. While an uned pile bore may remain intact in a silty sand / sandy silt soil the same will invariably tend to collapse in a medium to coarse sandy soil. Therefore, while based on the results of the geotechnical investigation report, an uned pile bore can be considered feasible, the same will be considered difficult to infeasible when the actual subsoil conditions as indicated by the new boreholes are taken into . b.

Vibrations from Rail and Road Traffic

To assess the level of vibrations caused by the rail and road traffic in close proximity to the site of the proposed elevated corridor, vibration measurements had been carried out using geophones (i.e. velocity pick-ups with a sensitivity of 28.3 mv/mm/s) connected either to a storage oscilloscope or to a computer based data acquisition system. The locations of the vibration measurement points have been indicated in the schematic layout plan of the area given in fig.1. The recorded vibrations from different types of trains as also from trucks plying on the road have been presented in figs. 2 to 6. The absorption coefficient for the subsoil at the site has been determined from the vibrations recorded during the age of a Rajdhani express train by locating the geophones in a line perpendicular to the railway tracks on the opposite side of the road under the proposed elevated corridor. From the recorded vibrations and the absorption coefficient determined, the values of the acceleration ratio at various distances from the tracks will be :

Distance (m) :

6

Acceleration Ratio : 0.03

10

14

18

22

0.013

0.009

0.005

0.002

As per Tscheboterioff (1965), dynamic forces caused by pile driving operations having high accelerations in the range of 0.01g to 0.001g have been shown to have caused significant soil settlements to cause distress to the nearby structures and substructures. As the levels of vibrations caused by the rail traffic are in the above range, the vibrations caused by the rail traffic can be considered be a major contributory factor in the collapse of the pile bores. 6.

CONCLUSIONS

The investigation of the possible cause/s indicates that both the considered possible causes have contributed to the frequent collapse of the pile bores in that: 1. The subsoil profile given in the tender documents as indicated by the geotechnical investigation report is wrong. 2. The actual subsoil conditions indicate that the subsoil at the site is unsuitable for use of the bailer technique. 3. The agency that carried out the initial geotechnical investigation is the main culprit in having provided wrong data about the subsoil conditions leading to the inappropriate choice of method of construction of the bored piles. 4. The engineers of the highways authority should have properly verified the geotechnical investigation report before going in for tendering of the work and insisting on bored piles with restrictions on method adopted for the same. 5. Vibrations from the rail and road traffic are significant enough to have contributed to the frequent collapse of the pile bores. 6. Use of full length liner would be the only method by which bored piles could be constructed in such subsoil conditions where significant vibrations are generated by rail and road traffic.

Fig.1 : Schematic Layout Plan

Fig.2 : Express Train on Track 3 in E-W direction

Fig.3 : Rajdhani Express Train on Track 3

Fig.4 : Truck Plying on the Road

Fig.5 : enger Train on Track 1

Fig.6 : Goods Train on Track 3

ANNEXURE

COMPOSITION OF TC 40: FORENSIC GEOTECHNICAL ENGINEERING Chairman:Dr.V.V.S.Rao, India [email protected] Secretary: Prof. G.L. Sivakumar Babu, India, [email protected] Core : 1. 2. 3. 4. 5. 6. 7. 8. 9.

Prof. Yoshi Iwasaki, Japan, [email protected] Dr. Richard Hwang, Moh & Associates, Taiwan, [email protected] Mr. David Starr, Golder Associates, Australia, [email protected] Prof. K.K. Phoon, Singapore, [email protected] Mr. P.W. Day, South Africa, [email protected] Prof. J. Mesci, Hungary, [email protected] Dr. Jan Hellings, U.K., [email protected] Mr. D.S. Saxena, USA. [email protected] Prof. Mihail e Popescu [email protected]

General : 1. Prof. L.G.De Mello, Brazil [email protected] 2. Dr. Dianging Le, China [email protected] 3.Prof. Zeliko Arbanas, Croatia [email protected] 4.Mr. Jack Pappin, Hong Kong [email protected] 5. Prof. Lingwei Kong, China [email protected] 6. Mr. W.K.Pun, Hongkong [email protected] 7. Dr. R.Szepeshazi, Hungary [email protected] 8. Prof. S. Hamsin, Kazakstan 9. Dr.W.Cichy, Poland [email protected] 10. Prof. R. Gilbert, USA [email protected] 11. Dr. Chan S. F. Malaysia [email protected] Corresponding : 1. Mr. Marco Uzielli Italy, [email protected] 2. Mr. Gil Yoon, South Korea [email protected] 3. Mr. Masahiro Shirato, Japan, [email protected] 4. Mr. Makoto Suzuki, Japan, [email protected] 5. Mr. Limin Zhang, Hong Kong, [email protected] 6. Mr. Chris Basile P.E. (USA) [email protected] 7. Mr. Grey Stephan P.E. (USA) [email protected] 8. Dr.Phili Pells [email protected] 9. Mr. Michael Marley, [email protected] 10. Strath Clarke 11. Prof. Mihail Popescu , [email protected] THE SUBCOMMITTEES SC 1: Characterization of distress. Convener: Mr. P.W. Day SC2: Diagnostic Tests. Convener: Mr. David Starr SC 3: Back Analysis. Convener: Dr. Hwang SC 4: Instrumentation. Convener: Prof. Yoshi Iwasaki SC 5: Development of failure hypothesis: Convener: Dr. J.Hellings SC 6: Reliability Checks. Convener: Prof. K.K.Phoon SC 7: Legal Issues: Convener: Mr. D.S.Saxena SC 1: CHARACTERIZATION OF DISTRESS. Convener: Mr. Peter W. Day [email protected] : 1. Prof. Luiz de Mello [email protected] 2. Jack Pappin [email protected] 3. Prof. Bob Gilbert [email protected] 4. Dr. Jan Hellings [email protected] 5. Nick Shirlaw [email protected]

SC 2: DIAGNOSTIC TESTS. Convener: Mr.David Starr [email protected] : 1. Dr Philip Pells, Pells Sullivan & Meynink (PSM) [email protected] 2. Mr Michael Marley, Golder Associates [email protected] 3. Strath Clarke (he works for PSM ) SC-3: BACK ANALYSIS. Convener: Dr. Mihail Popescu [email protected] Co-Convener:Dr.RichardHwang. [email protected] : 1.Dr.StephenButtling [email protected] 2.Mr.NiclauseShirlaw [email protected] 3.Dr.TingWenHui [email protected] 4.Prof.FusaoOka [email protected] 5.Dr. Chung Tien Chin [email protected] SC 4: INSTRUMENTATION Convener: Dr. Yoshinara Iwasaki [email protected] : 1.Mr. Craig A. Davis (USA) [email protected] 2. A. WADA (Singapore) : [email protected] SC 5: DEVELOPMENT OF FAILURE HYPOTHESIS. Convener: Dr. Jan Hellings. [email protected] : 1. Bill Grose. Ove Arup & Partners (UK). [email protected] 2. Rab Fernie. Cementation Foundations Skanska Ltd [email protected] 3. Kjell Karlsrud. NGI (Norway). [email protected] 4. Dirk Luger. GeoDelft (The Netherlands). [email protected] 5. Mahmoud Mahmoud. GES Geotech (Canada). [email protected] 6. Darren Page. OTB (UK). [email protected] 7. Roger Thomson. EDGE Consultants UK Ltd. [email protected] SC6: RELIABILITY CHECKS. Convener: Prof. Kok-Kwang Phoon (Singapore,) – [email protected] : 1. Prof. Sivakumar Babu (India) - [email protected] 2. Mr. Robert Gilbert (USA) - [email protected] 3. Mr. Dian-qing Li (China) - [email protected] 4. Mr. Marco Uzielli (Italy) - [email protected] 5. Mr. Gil Yoon (South Korea) - [email protected] 6. Mr. Masahiro Shirato (Japan) - [email protected] 7. Mr. Makoto Suzuki (Japan) - [email protected] 8. Mr. Jack Pappin (Hong Kong) 9. Mr. Limin Zhang (Hong Kong) - [email protected] SC 7- LEGAL ISSUES INVOLVING JURIPRUDENCE SYSTEM Convener: Mr. Dhirendra S Saxena P.E. (USA) [email protected] : 1. Mr. Christopher C Basile P.E. (USA) [email protected] 2. Mr. Greg A Stephan P.E. (USA) [email protected]

Forensic Geotechnical Engineering n5x3v

Overview 26281t

More details 6y5l6z

Related Documents 3h463d

Forensic Geotechnical Engineering n5x3v

Forensic Geotechnical And Foundation Engineering, 2nd Ed 6e2o49

Geotechnical Engineering v2p6y

Geotechnical Engineering Handbook3 765q

Advanced Geotechnical Engineering 3h5w48

Geotechnical Engineering Lecture Notes 4c1x2i

More Documents from "Georg Vatsellas" 3i2n4g

Geotechnical Engineering Handbook3 765q

Geotechnical Engineering Handbook2 2f4d6

Forensic Geotechnical Engineering n5x3v

Geotechnical Engineering Handbook1 p4n1x

Audio 7/93 Testbericht Harman/kardon 7625 Cd-player 4s3b6f

4p345f