中文摘要：

为了研究振动信号与系统所受载荷的耦合关系,提出了BP神经网络的识别方法.以拉格朗日方程为基础,建立了机械转子轴承系统的弯扭耦合方程,得到了基于系统位移和扭转角响应的系统运算方程.在MATLAB/simulink环境下建模,使用信号发生器模拟加载并对其进行仿真,构建自适应算法的BP神经网络模型,利用振动位移对实验的载荷进行识别.研究结果表明,在对转子系统进行载荷识别时,自适应BP神经网络方法相对简单且有更高的识别精度,平均相对误差为2.15％.

英文摘要：

In order to study the coupling relationship between the vibration signal and the load of the system,the method of BP neural network is proposed.A lateral torsional coupling vibration equation of machine rotor bearing system was set up based on Lagrange equation,and the system operation equations were obtained based on the system displacement and torsion angle response.Under the environment of MATLAB/simulink to realize the modeling,using signal generator analog loading and simulation,the self-adaptive algorithm of BP neural network model was built,using vibration displacement to applying the experiment load identification.The identification result shows that the self-adaptive BP neural network method is relatively simple and has higher recognition accuracy when load identification of a medium test bench with the rotor system,the average relative error is 2.15％.