各向异性层合阻尼结构的BP集成神经网络模型设计

Design on the mode of the BP integrated neural network based on the anisotropic laminated damping structure

本文以各向异性层合阻尼结构为研究对象,设计一种基于BP集成神经网络的智能分析模型。该模型中的集成神经网络由两个子系统神经网络并联融合而成,学习算法主要采用Sigmoid函数。同时,该模型设计针对各向异性层合阻尼结构参数的扰动性问题综合采用结构模式归类、学习算法的改进、小波分析方法予以处理。计算结果表明:该BP集成神经网络模型,较好地解决了各向异性层合阻尼结构参数的扰动性问题,并能有效量化结构参数的变化影响。

This paper studies the anisotropic laminated damping structure by a kind of BP integrated neural network, which is made up of two subsystems of neural network. However, the learning method adopts Sigmoid function. At the same time, the design adopts three methods of structure classification, improved arithmetic, and the wavelet analysis for resolving the mutual disturbance of structure parameters. The calculated result shows that the mode based on the BP integrated neural network can resolve the disturbance of structure parameters and quantitatively analyse the changes of structure parameters.