Calibrating Engineering Judgment in Geotechnical Risk Analysis

点击翻译按钮获取中文摘要

This research utilizes a case-study approach to develop improved protocols for quantifying engineering judgment in geotechnical engineering risk analysis. The concept is based on four converging trends in geotechnical engineering: (1) Subjective judgment is viewed as important to geotechnical practice; (2) important organizations are using geotechnical risk analysis; (3) current geotechnical risk analyses involving quantified expert judgment tend to be practical, but a theoretical basis is needed; and (4) current risk analyses case studies are a real life laboratory for furthering our understanding of how engineering judgment is quantified, how it is validated and calibrated, and how robust the resulting risk analyses are.The research employs two major geotechnical risk analysis studies of dam safety to understand the process of validating and calibrating probabilities assigned by experts. As a starting point for the case studies, the research uses theory from subjective probability. The research furthers this body of knowledge, building on unique aspects of the use of geotechnical judgment in complex risk analyses. It applies the developed approaches in two major geotechnical risk analysis projects undertaken by large dam-owning organizations.The research is organized in four tasks: case studies, theory and method development, tested applications, and information dissemination. The two case studies employed use an extensively documented geotechnical risk analysis of the safety of large embankment dams. Each study relied heavily on the quantified judgment of consulting experts. The project integrates historical engineering case studies with theoretical developments in other disciplines (e.g., cognitive psychology, management science, and inductive reasoning), and with practical tested applications on ongoing projects. The intellectual merit consists in bring multidisciplinary insight to a practical, emerging, critical area of geotechnical engineering practice.The broader impact of the research will be enhanced by several factors: The University of Maryland provides a diverse research environment; strong efforts will be made to recruit women and minority students, both graduate and undergraduate; the outcomes of the work will appear in undergraduate research, graduate and professional courses, and in a large multi-disciplinary course; technological risk analysis has become pervasive in the management and regulatory activities of government at all levels. The use of professional judgment, increasingly quantified as probabilities, pertains to most of these uses of risk analysis, and the results of the project can inform all these endeavors. In particular, the security of our civil infrastructure against natural hazards and man-made threats has come to dominate national debate. Risk analysis and the necessary role of quantified expert judgment are integral to our ability to address these hazards and threats.