起重机液压系统的粒子群神经网络故障诊断

Fault Diagnosis Methods for Crane Hydraulic System Based on Particle Swarm Neural Network

为了研究汽车起重机回转液压系统故障诊断问题,提高诊断效率。为有效提高汽车起重机回转液压系统故障诊断的速度与精度,针对传统故障诊断算法收敛速度慢、容易陷入局部最优,导致诊断精度不高的问题,提出了一种根据小波包能量熵的粒子群神经网络汽车起重机回转液压系统故障诊断方法(EEPSONN)。首先依托汽车起重机回转液压系统实验平台,提取五种回转故障模式信号;然后利用小波包变换提取Shannon熵值,作为故障输入特征向量;最后利用粒子群优化算法提升BP神经网络,对故障进行建模识别。试验表明此法具有较高识别率,为汽车起重机回转液压系统故障诊断提供一种有价值的诊断方法。

The diagnostic efficiency should be improved in the study on fault diagnosis of the truck crane slewing hydraulic system fault diagnosis. For the problem of low efficiency of the traditional diagnosis method with slow con- vergence speed and easily getting into local optimal value, a novel method of hydraulic system fault diagnosis based on the wavelet packet Energy Entropy and the Particle Swarm Optimization Neural Network（EE-PSONN） was pro- posed to improve the efficiency and the accuracy of the hydraulic system fault diagnosis. First, relying on the exper-imental platform of truck crane slewing hydraulic system, the method extracts five kinds of signals of slewing fault mode, and then uses the wavelet packet transform to pick up the energy value of Shannon entropy as the feature vectors of the fault input, and finally uses the particle swarm optimization algorithm to promote the BP neural network to make identification about the fault. Experimental results show that the proposed method has a higher recognition rate and provides a valuable diagnostic method for the hydraulic slewing system fault diagnosis of the truck crane.