基于多特征融合多核学习支持向量机的液压泵故障识别方法  

Fault Identification Method for Hydraulic Pumps Based on Multi-feature Fusion and Multiple Kernel Learning SVM

作　　者：刘志强[1,2,3] 姜万录[1,2] 谭文振 朱勇[1,2] 

机构地区：[1]燕山大学河北省重型机械流体动力传输与控制重点实验室,秦皇岛066004 [2]先进锻压成形技术与科学教育部重点实验室,秦皇岛066004 [3]唐钢高强汽车板有限公司,唐山063002

出　　处：《中国机械工程》2016年第24期3355-3361,共7页China Mechanical Engineering

基　　金：国家自然科学基金资助项目(51475405);国家重点基础研究发展计划(973计划)资助项目(2014CB046405);河北省自然科学基金资助项目(E2013203161)

摘　　要：提出基于多特征融合多核学习支持向量机的液压泵故障识别方法。该方法首先对原始信号进行集总经验模态分解,然后分别用AR模型和奇异值分解两种特征提取方法提取故障特征,最后将不同类型的特征分别用相应的核函数进行映射,用多核学习支持向量机来识别液压泵的工作状态和故障类型。实验结果表明该方法显著地提高了故障诊断的准确性。A hydraulic pump fault identification method was put forward based on multiple feature fusion and multiple kernel learning SVM. Firstly, the original signals were processed by the ensemble empirical mode decomposition. Then, the feature vectors of hydraulic pump faults were obtained by using the autoregressive model and the singular value decomposition. Through different types of features mapped by corresponding different kernel functions, the hydraulic pump working conditions and fault types might be finally identified by multiple kernel learning SVM. The experimental results show that the approach improves the accuracy of fault diagnosis significantly.

关 键 词：多核学习 多特征融合 支持向量机 故障识别 液压泵 

分 类 号：TP277[自动化与计算机技术—检测技术与自动化装置]