Quantitative analysis method for the importance of stress influencing factors of a high arch dam during the operation period using SPA–OSC–PLS

Abstract

Stress analysis is an important part of the health monitoring and evaluation of a high arch dam. At present, a reasonable quantitative analysis method for the importance of stress influencing factors of a high arch dam during the operation period remains lacking. According to measured data of strain gauge groups, a quantitative analysis method is proposed integrating succession projection algorithm (SPA), orthogonal signal correction (OSC), and partial least squares (PLS). First, based on the stress calculation of the strain gauge group of a concrete dam, the statistical model of high arch dam stress is established. Second, independent variables are preprocessed using SPA and OSC to screen out effective influencing factors without noise signals. Finally, the principal component with the strongest explanation for the dependent variable is extracted using PLS, and the importance of stress influencing factors of a high arch dam during the operation period is quantitatively separated. Results show that the SPA–OSC–PLS method can effectively eliminate the effects of multiple correlations, information overlap, and noise among influencing factors. In addition, the predictive ability and interpretability of the SPA–OSC–PLS method are superior to the PLS and SPA–PLS methods, which can reasonably obtain the importance of stress influencing factors of a high arch dam. The proposed method can accurately reveal the change law and action mechanism of the stress of a high arch dam during the operation period.