A deep learning-based damage state prediction method for self-centering shear walls

Abstract

The application discloses a self-resetting shear wall damage state prediction method based on deep learning, and relates to the technical field of machine learning. The method comprises the following steps: acquiring self-resetting shear wall structure parameters and establishing a numerical model, obtaining capacity characteristics and response characteristics through loading analysis, and determining characteristic damage states according to material states and spatial positions; constructing a wall damage state sample and dividing a training set, a verification set and a test set; respectively performing standardization, feature-by-feature attention processing and independent coding on structure characteristics, capacity characteristics and response characteristics to obtain structure implicit representation, capacity implicit representation and response implicit representation; performing gate fusion on the three types of implicit representations to obtain fusion representation, further obtaining a shared damage score, and calculating a cumulative probability through an ordered damage threshold; and determining the characteristic damage state of the self-resetting shear wall according to the cumulative probability.

Images