2024 Volume 80 Issue 13 Article ID: 23-13100
In this study, we attempted to construct a damage discrimination model by machine learning of response acceleration time histories obtained from acceleration sensors installed in buildings, as a method for efficiently and quickly identifying damage states of wooden houses after an earthquake. First, we performed a number of seismic response analyses of a wooden house model to calculate the response acceleration time histories and response plasticity ratios, which were used as training data for machine learning. Then, Long Short-term Memory (LSTM) network and a one-dimensional convolutional neural network (1D CNN) were employed, and the obtained results were compared to construct an optimal discriminant model.
