Optimization of Seismic Performance of Structures Based on Machine Learning Algorithms
Year: 2025
Type: conference-paper
Abstract: With the acceleration of urbanization and the frequent occurrence of natural disasters, the seismic performance of building structures has become an important issue for the protection of people’s lives and properties. Traditional seismic design methods usually rely on empirical formulations and static analysis, which are difficult to comprehensively consider the structural response under complex seismic effects. To address this challenge, this study proposes an optimization method for structural seismic performance based on machine learning, ML algorithm. Specifically, we first collect a large amount of historical seismic data and structural response data to construct a comprehensive dataset. Then, ML algorithms such as convolutional neural network (CNN) and recurrent neural network (RNN) are used to train the dataset and establish the complex mapping relationship between seismic actions and structural responses. Through repeated iterations and optimization, the model can accurately predict the structural response under different seismic actions. In this study, the structural response under different design parameters is simulated and quickly evaluated with ML model to screen out the design solutions with the best seismic performance. In particular, the frame-shear wall hybrid structure shows the best performance in terms of optimized maximum displacement reduction rate and maximum acceleration reduction rate, which are 19.1% and 15.7%, respectively. The structural seismic performance optimization method based on ML algorithm proposed in this study not only overcomes the limitations of the traditional method, but also provides a new method and new ideas for seismic design of building structures.
Source: Procedia Computer Science
Author: Zhidan Lin
Language: English
Subfield: Civil and Structural Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W4411176152