基于长短时记忆网络的结构动态载荷预测方法
Structural dynamic load prediction method based on long short-term memory network
樊昱玮 1郭腾博 1李哲 2洪良友 2刘超 1蒋东翔1
作者信息
- 1. 清华大学 能源与动力工程系,北京 100084
- 2. 北京强度环境研究所,北京 100076
- 折叠
摘要
[目的]针对传统代理模型无法处理具有时间依赖性的动态过程和异构数据的问题,提出一种基于长短时记忆网络(LSTM)的动态载荷代理模型方法.[方法]代理模型包含载荷特征编码和载荷响应解码2 个模块.首先,通过载荷特征编码模块的LSTM对动态外载荷时间序列进行特征提取;然后,将外载荷时序特征与结构参数特征进行融合,由载荷解码模块的LSTM进一步进行特征提取并生成最终输出,从而综合考虑动态外载荷时间序列和结构参数一维特征的异构数据输入,预测结构内力响应时间历程;最后,在有限元仿真数据集上对模型进行精度评估,并与其他代理模型方法进行对比.[结果]结果显示,该动态载荷代理模型的平均精度可达 98%,高于其他对比方法,且计算速度相较于有限元方法更快.[结论]所提方法可解决时序-非时序异构数据的代理模型问题,具有精度高、效率高的优点,在快速迭代计算场景下能够发挥较大作用.
Abstract
[Objective]To address the limitations of traditional surrogate models in handling time-dependent dynamic processes and heterogeneous data,this paper proposes a dynamic load surrogate model method based on a long short-term memory(LSTM)network.[Methods]The surrogate model is comprised of two mod-ules:the load feature encoder and load response decoder.First,the LSTM in the load feature encoder per-forms feature extraction on the time series of dynamic external loads.Next,the extracted load features are combined with the structural parameter features.The LSTM in the load decoder conducts further feature ex-traction and finally generates output while comprehensively considering the heterogeneous data input of the dynamic external load time series and one-dimensional structural parameter features in order to predict the time history of internal force responses.Finally,the model's accuracy is evaluated using a finite element simu-lation dataset and compared with other surrogate model methods.[Results]The results show that the aver-age accuracy of the dynamic load surrogate model can reach 98%,which is higher than that of other methods,and its calculation speed is faster than that of the finite element method.[Conclusions]The proposed meth-od addresses the issue of heterogeneous data involving both time-series and non-time-series features,and of-fers advantages such as high accuracy and efficiency,making it effective for fast iterative computation tasks.
关键词
结构优化/动态载荷/人工智能/代理模型/深度学习/长短时记忆网络Key words
structural optimization/dynamic loads/artificial intelligence/surrogate model/deep learn-ing/long short-term memory(LSTM)network引用本文复制引用
出版年
2024
国家科技期刊平台
NSTL
万方数据