A method for characterizing and analyzing the structural behavior of concrete dams in cold regions.

Aiming at the gross error and data missing in the monitoring sequence of concrete dam, the variational mode decomposition method and the gated recurrent unit depth learning algorithm are respectively used to extract the effective information of the monitoring sequence. On the basis of the research on the characteristics of the traditional concrete dam structural behavior characterization model, the paper explores the expression mode of the effect of cold wave, freeze-thaw, wintering layer and other influencing factors. In order to reflect the correlation between the monitoring measurement values, the space coordinate variable is introduced to establish the monitoring measurement change characterization model, so as to realize the characterization and analysis of the structural behavior changes of concrete dams in cold regions and the quantitative analysis of various influencing factors. Based on the research in this article, we can fully understand the operation status of the dam, identify hidden dangers, and carry out relevant risk investigation and reinforcement. It can reduce the risk of dam failure to a certain extent.