基于智能优化算法的门式起重机结构地震易损研究OA

Given the important role of portal cranes in cargo handling,accurately predicting their post-earthquake damage is essential.150 ground motion entries and 4 portal cranes were selected to establish Data Set A and Data Set B,and a damage prediction model of portal crane based on support vector machine was constructed based on Data Set B.After comparing and analyzing the differences of models established with these four different kinds of kernel functions,Data Set A was applied to this model,and the performance of the model was assessed by comparing the vulnerability curves.In addition,the model was optimized by adopting intelligent optimization algorithm,followed by an assessment of the performance of the model after optimization.The findings reveal that the model utilizing the Gaussian kernel function outperforms others in predictive accuracy.This model demonstrates robust performance on Data Set A,with its predicted vulnerability curves closely aligning with those derived from the Incremental Dynamic Analysis(IDA)method.Furthermore,the model optimized by the sparrow search algorithm exhibits enhanced performance on Data Set A.In both scenarios,the optimized models yield vulnerability curves that more closely match the IDA-based curves compared to the non-optimized models,thereby validating the effectiveness of this optimization approach.