基于GA-BP算法激光设备故障预测技术研究

针对激光设备非计划停机次数较多的问题，提出了一种基于遗传算法（Genetic Algorithm,GA）优化BP神经网络，建立激光设备故障预测模型的方法。利用激光设备的历史数据训练和调整预测算法，对激光设备采集的实时数据进行分析，按照算法模型预测故障发生概率，提前维护激光设备，减少非计划停机次数，提高激光设备的有效运行时间。通过测量各种情况下激光设备在切割零件时的数据变化，利用GA优化BP神经网络算法，建立激光设备故障预测模型。选取各种情形的切割零件的数据进行仿真预测和验证，以切割过程中各种情况的气体压力、激光功率、切割速度、加速度、各轴的温度和计算之后各轴的跟随误差作为模型输入，以粗糙度作为模型输出。结果表明，经过GA优化的模型在预测效果和预测精度上优于未经GA优化的模型，且模型经GA优化后，其粗糙度的预测精度和收敛速度得到了提升。

Research on Fault Prediction Technology of Laser Equipment Based on GA-BP Algorithm

Aiming at the frequent unplanned shutdown of laser equipment,a method based on genetic algorithm (GA) is proposed to optimize the BP neural network and estalbish the fault prediction model of laser equipment.The historical data of the laser equipment is used to train and adjust the prediction algorithm,analyze the real-time data collected by the laser equipment,predict the probability of fault according to the algorithm model,maintain the laser equipment in advance,reduce the number of unplanned shutdown,and improve the effective running time of the laser equipment.By measuring the data changes of the laser equipment when cutting parts under various conditions,the GA is used to optimize the BP neural network algorithm to establish a fault prediction model of laser equipment.The data of cutting parts in various situations are selected for simulation prediction and verification.The gas pressure,laser power,cutting speed,as well as the calculated following error,acceleration,and temperature of each axis in various situations during the cutting process are used as the input of the model.The roughness is used as the output of the model.The results show that the prediction effect and prediction accuracy of the model optimized by GA are better than that of the model without optimization by GA,and after GA optimization the prediction accuracy and convergence speed of the models roughness are improved.