KRIGING RISK MODEL OF DAM FAILURE BASED ON HL-RF ALGORITHM

Majority of common dam risk analysis methods based on reliability theory are conducted under an assumption that the limit state equation is known. However, in complex dam engineering, the explicit function relationship between performance functions and base variables is generally unknown. That is, the risk rate analysis model cannot be predetermined when the performance function of the dam structure is implicit or nonlinear. To solve this problem, Latin hypercube sampling is adopted to select the sample point, and a Kriging model is used to reconstruct the dam performance function. An HL-RF optimization algorithm is established, and a risk analysis model under a single-dam failure mode is proposed. An engineering example shows that the proposed method proposed can meet engineering requirements, thereby validating its effectiveness and precision. The proposed method provides a new way of solving a risk rate with a complex performance function.