Study On Safety Comprehensive Assessment Of Hydro-Power Construction Sites

Safety management has always been the major theme of the management of a hydroelectric project. Adopting safety evaluation has changed safety management from remedy after the accident to prevention beforehand, which also improves the enterprise's ability to control occurring of accident. Current safety evaluation of construction is based on safety checklist. Safety checklist method is easy to be carried out, but it is a static evaluation method and the flexibility is poor. Researchers tend to use Analytic Hierarchy Process (AHP)–Fuzzy approach to solve the problem in safety evaluation. AHP-Fuzzy is easy to use, and the result's accuracy depends on weight vector of indexes and the fuzzy relation matrixes. Considering the complexity of hydraulic and hydroelectric construction, establishing multiple index systems to aim at a variety of situation cause poor flexibility and operability, which needs multiple models to meet requirements. However, establishing an index system containing all factors necessary isn't fit for overall different situations, which resulting in the problem of incomplete data, and the accuracy of Fuzzy will be decrease. Neural network model is applicable for the problem.With the analysis about the fact of hydro-Power construction sites, this study establishes a comprehensive evaluation index system. The weights of indexes are determined by the method of AHP. Study use Fuzzy Comprehensive Evaluation to train the sample. Fuzzy Comprehensive Evaluation is also used to train the BP neural network in order to get a stable safety evaluation model. Finally, this study consider the influence on the final result when lack data of each indexes. The Comparative Results tell that result coming from Multiple Imputation and Fuzzy evaluation method has a larger Error than the result coming from Neural network,and the larger weights it has the More obvious this error can be found.