考虑变形效应量间随机效应的水工混凝土结构安全监控模型OA

In-depth analysis of deformation monitoring data of hydraulic concrete structures is one of the most important means to assess their real-time operational state.However,due to the complex environmental conditions under which these structures typically operate,nonlinear coupling relationships often exist among deformation monitoring points,making direct quantification challenging.To address this issue,an adaptive X-means clustering algorithm is first employed to group multiple deformation monitoring points of a given hydraulic concrete structure.Then,based on stochastic analysis theory,the synergistic relationships among various monitoring points within strongly correlated groups are extracted.The statistical characteristics of these correlations are incorporated into the traditional set of deformation-influencing factors.Using the measured values of one point within a correlated group as the input variable,a multi-kernel least squares support vector regression(MK-LSSVR)model is applied to capture the nonlinear relationships between input and output variables.This approach establishes a deformation monitoring model for hydraulic concrete structures,and its effectiveness in improving monitoring accuracy is validated through relevant evaluation metrics.