A multicriteria fuzzy approximate reasoning approach for risk assessment of dam safety

The benefits of dams are unquestionable, with the water impoundment upstream and flood-control downstream usually being sufficient to increase investments and promote growth. However, their presence brings risks to the downstream population and the environment. This paper employs a fuzzy approximate reasoning for dam safety risk assessment, initially using a FAHP to weight the importance of the indicators suggested by a dam expert and then adopts linguistic variables in a two-stage risk assessment. In the first stage, we compute the risk of dam collapse, hereinafter called dam break, through indicators such as design premises and safety procedures. In the second stage, we evaluate the combined effect between the risk of a dam collapse and the socioeconomic and environmental impact on the surrounding area, hereinafter called potential risk. The danger scale of dam break and the hazard scale of potential risk were constructed. The risk assessment of the Simplicio Hydroelectric plant, located in Brazil’s Southern region, showed that the risk classification of the proposed method attached great importance to the downstream consequences. The accuracy of the evaluation results was proved through a study of the Oroville Dam, whose spillway faced a failure event. These findings indicate the importance of extracting the knowledge of experts and incorporating it in the model specification through fuzzification of variables and the rule-based construction.