基于小波能量熵和邻域粗糙集的电磁阀故障诊断研究

Research on Fault Diagnosis of Solenoid Valve Based on Wavelet Energy Entropy and Neighborhood Rough Set

摘要:

液体电磁阀是航空变量柱塞泵系统的关键控制元件，准确诊断电磁阀的故障类型是系统安全运行的重要保证；为了对航空变量柱塞泵系统中电磁阀进行检测与诊断，提出了一种基于小波能量熵和邻域粗糙集相结合的方法；研究了液体电磁阀驱动端的电流故障特征；采用AMEsim多学科领域仿真软件建立仿真模型，分析驱动端电流与阀体位移的关系，搭建实验系统，采集正常状态和故障状态的电流信号，分析不同状态的电流特性；然后利用小波包分解重构信号，提取对应的频带能量作为特征向量；最后，引入变精度模糊邻域粗集约简算法简化数值属性，提高了系统的效率；简化的属性集用于构建决策树，经过迭代训练，该模型诊断准确率达90%，达到了预期效果，实现了对液体电磁阀的快速诊断。

Abstract:

The liquid solenoid valve is the key control component of the variable displacement piston pump system in aviation. Accurate diagnosis of the fault type of the solenoid valve is an important guarantee for the safe operation of the system; In order to detect and diagnose the solenoid valve in the aviation variable displacement piston pump system, a method based on wavelet energy entropy and neighborhood rough set is proposed; The current fault characteristics of liquid solenoid valve driver are studied; The AMEsim multidisciplinary simulation software is used to establish the simulation model, analyze the relationship between the drive end current and the valve body displacement, build an experimental system, collect the current signals in normal and fault states, and analyze the current characteristics in different states; Then the wavelet packet decomposition is used to reconstruct the signal, and the corresponding band energy is extracted as the feature vector; Finally, variable precision fuzzy neighborhood rough set reduction algorithm is introduced to simplify the numerical attributes, which improves the efficiency of the system; The simplified attribute set is used to build the decision tree. After iterative training, the diagnostic accuracy of the model reaches 90%, achieving the desired effect, and realizing the rapid diagnosis of the liquid solenoid valve.