GA-BP神经网络在液压缸故障诊断仿真中的应用

Application of GA-BP Neural Network in Hydraulic Cylinder Fault Diagnosis Simulation

大型AGC伺服液压缸结构复杂、价格昂贵、维修成本高,故障模拟代价巨大。为解决大型液压缸实际工作过程中故障数据难收集难处理的问题,提出利用仿真模拟液压缸模型,从中提取故障数据,通过遗传算法(Genetic Algorithm,GA)来优化BP神经网络处理数据的故障诊断方式。运用仿真软件AMESim建立对应的液压缸仿真模型,通过改变仿真参数模拟出液压缸5种故障类型,获取故障数据。分别用传统的BP网络和经过遗传算法优化后的BP网络进行训练和测试。测试结果表明,GA-BP神经网络比传统BP神经网络测试误差小、预测精度高,能够准确实现故障诊断。该方法也为大型液压缸故障诊断提供了一种解决思路和方法。

Large AGC servo hydraulic cylinder has complicated structure,expensive price and high maintenance cost,so the cost of fault simulation is huge.In order to solve the problem that the fault data is difficult to collect and process in the actual working process of large hydraulic cylinders.A fault diagnosis method of extracting fault data by simulating hydraulic cylinder model and processing data with BP neural network optimized by genetic algorithm(GA)was proposed.AMESim was used to establish the corresponding simulation model of hydraulic cylinder,and the five fault types of the hydraulic cylinder were simulated by changing the simulation parameters to obtain fault data.The traditional BP neural network and BP neural network optimized by genetic algorithm were used for training and testing.The test results show that GA-BP neural network has smaller test error and higher prediction accuracy than traditional BP neural network,and can accurately realize fault diagnosis.This method also provides a solution and method for fault diagnosis of large hydraulic cylinders.