Bayesian decision analysis of dam safety

Inspections and safety assessments are germane elements of proper dam maintenance programs. To aid geotechnical engineers and/or dam owners in developing such programs, a Bayesian methodology is presented. It offers four main features: (1) the engineer can formally combine his judgment based on expertise with factual information obtained from geotechnical tests, (2) uncertainty about the state of the dam can be quantified explicitly, (3) the owner can specify the desired risk preference, and (4) an optimal geotechnical test plan and remedial action can be determined so as to take account of uncertainty, risk preference, and economic losses. The methodology is illustrated by a case study of an earth dam.