基于多分支BP神经网络的结构系统辨识

STRUCTURAL SYSTEM IDENTIFICATION BASED ON MULTI-BRANCH BP NEURAL NETWORK

  • 摘要: 结构动力学系统是一个多输入多输出系统,建立能精确体现结构动力特性的辨识模型是实现结构高品质振动控制和关系到控制是否有效的关键。提出一种多分支BP神经网络辨识模型,将影响结构动力反应的结构状态变量和地震动输入分别作为模型的分支输入来进行辨识,提高学习效率及预测精度,并利用训练好的模型预测结构在不同地震波输入下的动力反应,验证模型的泛化能力。数值分析结果表明,用所提出的多分支BP网络模型对结构动力学系统进行动力特性辨识时能达到较高的精度,而且预测精度也很高。

     

    Abstract: A structural dynamic system has multiple inputs and outputs.To establish an identification model that can precisely reflect structural dynamic characteristics is the key factor to realize valid vibration control with high quality.In this paper,a kind of multi-branch BPNN identification model is proposed.To enhance training efficiency and prediction precision,structural status variable and earthquake inputs,which affect structural dynamic response,are treated as respective branches and inputted into the model.The trained model is used to predict structural dynamic response under different earthquake in order to validate the generalization ability of the model.Numerical analysis shows that the proposed multi-branch BBNN model can identify the dynamic characteristics of a structural dynamic system and predict the structural dynamic response with high precision.

     

/

返回文章
返回