线性振动的神经网络建模

Neural Network Modeling of Linear Vibration

本文利用人工神经网络对两自由度线性振动系统进行了神经网络建模,并通过所建立的神经网络模型对该系统进行了预测。分别利用MATLAB和BP网络作为平台和训练工具。以两自由度悬臂梁的受迫振动为例,将一段时间内的激励力作为网络的输入参数,对应于该段时间内由振动产生的挠度作为网络的输出参数,然后利用BP网络进行训练。将网络模型预测结果与精确解进行对比,误差甚小。该结果表明:所建立的神经网络模型合理、有效,可利用其对该类问题进行预测并应用于工程实践中。

Using artificial neural network to get two degrees linear vibration system modeling, the neural network by the neural network model for this system is forecasted. MATLAB and BP network have respectively been as platform and training toolsl For example, with two degrees of the cantilever beam forced vibration, a time of basic will be as input parameters and corresponding to that period of deflection by vibration as network output parameters, then using BP neural network was trained, comparing the network model prediction results with the exact solutions, there is a very small error. The results show that： the neural network model is reasonable and effective, which can be used to predict these problems and application in engineering practice.