Probability theory provides a formal basis for quantifying risks that otherwise must be dealt with qualitatively by the use of engineering judgment. It also provides a sound basis for choosing a course of action in the face of incomplete information.
Arthur Casagrande was among the first to articulate formally the need to balance possible risk with cost while making decisions in geotechnical engineering. His concept of “calculated risk” in geotechnical engineering embodied two elements (Casagrande, 1965, p. 1).
The use of imperfect knowledge, guided by judgment and experience, to estimate the probable ranges for all pertinent quantities that enter into the solution of a problem;
The decision on an appropriate margin of safety, or degree of risk, taking into consideration economic factors and the magnitude of losses that would result from failure.
In the early 1970s, when probability theory was first introduced into geotechnical engineering, many geotechnical engineers expected it would cause widespread change in geotechnical engineering practice. At the time, the Soil Mechanics and Foundations Division of the American Society of Civil Engineers established its Committee on Reliability and Probability Concepts in Geotechnical Engineering Design. In 1977, the National Science Foundation sponsored a workshop at the Rensselaer Polytechnic Institute on probability theory and reliability analysis in geotechnical engineering. The following were among the conclusions presented in the proceedings of the workshop which reflected the consensus of participants (Grivas, 1977, p. 289).
The time is right for introduction of decision and risk analysis into geotechnical engineering practice. This will be analogous to the introduction of CPM (critical path method) and PERT (program evaluation and review technique) into construction practice. . . .
Programs should be initiated for the education of professionals, owners and planners in decision-making in geotechnical engineering. . . .
The Engineering Foundation or ASCE (American Society of Civil Engineers) should take the initiative for showing through continuing education and conferences how risk evaluation and decision analysis can be done.
Since then, several national and international workshops, symposia, and meetings on reliability in geotechnical engineering have been held, and an increasing number of papers on probabilistic approaches to geotechnical engineering have been published. For some types of problems—natural hazard risk assessment, rock engineering as applied to mining, and geotechnical engineering for offshore construction —there has been considerable use of probabilistic methods. These methods have also found application in the design of exploration programs. Some notable areas of application include rock joint models (e.g., Kulatilake, 1988), landslide risk assessments (e.g., Einstein, 1988), tunnel design (e.g., Einstein et al., 1989) and offshore foundation performances (e.g., Wu et al., 1989).
Given the variable nature of soil and rock, changeable environmental conditions, and the uncertainties in predicting field performance from available geotechnical models, coping with uncertainty is a hallmark of geotechnical engineering. In spite of previous efforts to promote the use of probabilistic methods, formal probabilistic methods have been used much less widely in mainstream geotechnical engineering than might logically be expected. While probabilistic methods have been introduced into some areas of practice by innovative engineers and prescribed by some perceptive owners and planners, the vision articulated in the proceedings of the Rensselaer Polytechnic Institute workshop has generally gone unfulfilled. Most areas of geotechnical engineering practice have not been influenced to any perceptible degree by probabilistic methods, and many geotechnical engineers remain skeptical of the value of probabilistic methods as applied to geotechnical engineering problems.
To explore the reasons for this apparent dichotomy between projected and actual use of probabilistic methods in geotechnical engineering, the Workshop on Reliability Methods for Risk Mitigation in Geotechnical Engineering was held at the National Academy of Sciences' Beckman Center in Irvine, California, on July 16–18, 1992. The workshop was organized by the Geotechnical Board of the National Academy of Sciences and sponsored by the U.S. Army Corps of Engineers, the Federal Highway Administration, and the National Science Foundation. Thirty eminent geotechnical engineering practitioners, researchers, and academics attended the workshop (further background and the statement of work for this study are given in Appendix A).
A few of the participants were almost pure traditional geotechnical determinists and a few were strong probabilists, but most had at least some experience in both deterministic and probabilistic approaches. The participants reviewed the state of practice in their experience and debated whether more-widespread use of probability methods in geotechnical engineering would benefit the profession, its clients, and the public. Although the workshop did not produce a firm consensus and participants agreed that opinions were mixed, the material generated at the workshop by working groups and in plenary sessions
(see workshop agenda in Appendix B) provides the basis from which the committee generated this report.
The workshop addressed four questions relevant to the seemingly slow acceptance of probabilistic methods in geotechnical engineering and to the appropriate role of these methods in geotechnical engineering practice:
What factors related to the training and experience of geotechnical engineers, or to the expectations of their clients, have inhibited the use of probability theory in the professional practice of geotechnical engineering?
In what areas of geotechnical engineering does probability theory have potential for more-widespread use and application? What might be the attendant benefits for the profession, its clients, and the public?
What is the appropriate role of probabilistic methods in codes and regulations that govern geotechnical engineering practice?
How should probabilistic methods be taught in civil engineering curricula in universities? Are the requirements of geotechnical engineering so unique that special considerations are required for this area? How should probabilistic methods be introduced to practicing geotechnical engineers who have had little or no formal education in the underlying theory?
In the remainder of this report, the Committee on Reliability Methods for Risk Mitigation in Geotechnical Engineering examines the four questions defined above. Chapter 2 first presents the objectives of a probabilistic analysis and its use in engineering, which should aid the reader who is not familiar with the subject. This is followed by a section, titled “Factors That Have Inhibited Use,” that is specifically aimed at answering Question 1. Question 2 is addressed in the subsequent section of Chapter 2, titled “Areas That Have Potential for Widespread Use,” by means of a series of example applications. Question 3, concerning codes and regulations, is addressed in Chapter 3, and Question 4, concerning education of students and practitioners, is dealt with in Chapter 4. The report focuses primarily upon the potential use of probabilistic methods in mainstream geotechnical engineering while drawing upon applications to a range of problems, including problems beyond the mainstream, to illustrate potential benefits from application of such approaches.
The committee decided to include a substantive appendix, “Appendix C: Basic Concepts of Probability and Reliability,” because of a concern that such brief introductory materials are not available in the literature in a single place. While it is not necessary to master the material in the appendix to understand this report, this appendix will serve to enhance understanding of the terms and procedures that are used in the case examples in Chapter 2. A selected bibliography on geotechnical reliability is given in Appendix D.
Casagrande, A. 1965. Role of the “calculated risk” in earthwork and foundation engineering. Journal of Soil Mechanics and Foundations Division, American Society of Civil Engineers 91(SM4): 1–40.
Einstein, H.H. 1988. Landslide risk assessment procedures. Pp. 1075–1090 in Proceedings of the Fifth International Symposium on Landslides, July 10–15, Lausanne, Christophe Bonnard, ed. Rotterdam: A.A. Balkema.
Einstein, H.H., G.F. Salazar, Y.W. Kim, and P.G. Ioannou. 1989. Computer based decision support for underground construction. Pp. 1287–1308 in Proceedings of Rapid Excavation and Tunnelling Conference, June 11 –14, Los Angeles, California, R. Pond and P. Kenny, eds. Littleton, Colorado: Society for Mining, Metallurgy and Explorations, Inc.
Grivas, D. A. (ed). 1977. Probability Theory and Reliability Analysis in Geotechnical Engineering . Report of an NSF-Sponsored Workshop at Rensselaer Polytechnic Institute (RPI). New York: RPI.
Kulatilake, P.H.S.W. 1988. State-of-the-art in stochastic joint geometry modelling. Key Questions in Rock Mechanics. Pp. 215–229 in Proceedings of the 29th U.S. Symposium on Rock Mechanics, June 13 –15, Minneapolis, Minnesota, P.A. Cundall et al., eds. Rotterdam: A.A. Balkema.
Wu, T.H., W.H. Tang, D.A. Sangrey, and G.B. Baecher. 1989. Reliability of offshore foundations—State-of-the-art. Journal of Geotechnical Engineering, American Society of Civil Engineers 115(2): 157–178.