摘要
液压系统不同程度泄漏故障发生原因多样,特征十分相似,难以正确诊断。针对此问题,提出了改进的EEMD方法与GA-SVM结合的液压系统泄漏故障诊断方法。首先,在EEMD方法的基础上提出改进,抑制模态混叠和端点效应对振动信号分解的影响,保证信号分解的真实性。运用改进的EEMD方法将液压振动信号分解成若干个IMF分量,计算各IMF分量能量并归一化处理提取振动信号特征向量。然后运用遗传算法对SVM进行参数优化,将提取到的特征向量输入优化后SVM分类诊断,判断液压系统泄漏故障类型和严重程度。实验结果表明,该方法能够有效地应用于液压系统泄漏故障诊断。
The hydraulic system leakage fault characteristics are very similar in different degrees and difficult to correctly diagnose the various reasons of fault. A 1 improved EEMD method and GA-SVM is proposed eakage fault diagnosis method for hydraulic system based on the to solve this problem. We put forward the improvement on the basis of EEMD method to restrain the effect of mode mixing and end effect to the decomposition of vibration signal and ensure the authenticity of the signal decomposition. The improved EEMD method is used to decompose hydrau- lic vibration signal into several IMF components, and the IMF component energy is calculated and normalization to the energy to feature extraction of vibration signal vector is presented. Then the genetic algorithm is used to optimize the parameters of SVM. The feature vector is input to the optimized SVM for diagnosis and classification to deter- the type and severity of the hydraulic system leakage fault. The experimental results show that the method can be effectively applied to fault diagnosis of hydraulic system leakage
出处
《液压与气动》
北大核心
2014年第9期32-38,共7页
Chinese Hydraulics & Pneumatics
基金
国家自然科学基金资助项目(51175511)
