以神经网络与MATLAB实现理论为依据,提出了一种新的滚动轴承振动预测方法。这种方法根据轴承的加工质量试验数据,建立轴承振动预测的BP网络试验模型,在MATLAB开发环境下输入训练样本数据矩阵和目标矩阵。经过训练后,网络误差达到要求,预报结果的最大相对误差小于10%。
Based on neural network and MATLAB realizing theory, a new method of forecasting vibration of rolling bearings was proposed. According to the test data of bearing processing quality, this method estabcishes the BP net trial model for bearing vibration prediction and the training sample data and the goal matrix are inputted in the MATLAB development environment. After.training, the network error is met matrix and the maximum of the relative errors of forecasting results is less than 10%.