采用L-M优化算法的设备状态预测

Predicting in device status based on L-M optimization algorithm

设备状态预测是设备预防性维护的重要组成部分。针对传统方法处理设备状态数据的不足,采用L-M优化算法进行设备状态预测。通过原始样本数据学习和训练BP神经网络,并用测试数据进行设备状态预测,实验证明该方法不仅在误差分析精度和收敛速度方面具有优良的性能,而且还证明了该算法的有效性。

Predicting device status is the important component of preventive maintenance. Specific to shortage in dealing with de- vice status data with conventional methods,it is appropriate to adapt L-M optimization algorithm to predict device status. Through learning and training neural network by sample data, and predicting device status by diagnostic data, the experimental result shows：the method not only owns excellent performance in error analysis accuracy and convergence rate , but also it is effective to forecast device status.