FSTPSO优化VMD及OMRDE特征在联合收割机装配质量检测中的应用研究  

Application Research of FSTPSO Optimization VMD and OMRDE Features in Combine Harvester Assembly Quality Inspection

作　　者：徐国夏 张家铭 马毅臻 轩梦辉 赵思夏 温金羽 XU Guoxia;ZHANG Jiaming;MA Yizhen;XUAN Menghui;ZHAO Sixia;WEN Jinyu(Henan University of Science and Technology,Luoyang 471003,China;Xinxing Jihua(Beijing)Intelligent Equipment Technology Research Institute Co.,Ltd.Beijing 100020,China)

机构地区：[1]河南科技大学,河南洛阳471003 [2]新兴际华(北京)智能装备技术研究院有限公司,北京100020

出　　处：《拖拉机与农用运输车》2025年第1期37-47,共11页Tractor & Farm Transporter

基　　金：河南省高等学校重点科研项目计划(25B460001)。

摘　　要：针对联合收割机在装配质量检测问题上缺乏有效的检测方法,提出一种基于模糊自整定粒子群算法(Fuzzy Self-tuning Particle Swarm Optimization,简称FSTPSO)优化变分模态分解(Variational Mode Decomposition,简称VMD)及最小二乘支持向量机(Least Squares Support Vector Machines,简称LSSVM)的故障诊断方法。采用优化多尺度反向离散熵(Optimized Multi-Scale Reverse Discrete Entropy,简称OMRDE)进行特征提取,并与时频域特征进行特征融合。建立FSTPSO-VMD-FSTPSO-LSSVM故障诊断模型,对比分析OMRDE、多尺度离散熵、模糊熵三种熵函数的特征提取效果,对比FSTPSO-VMD-DF、FSTPSO-VMD-DT、FSTPSO-VMD-SVM、FSTPSO-VMD-LSSVM、FSTPSO-VMD-KNN、FSTPSO-VMD-NBM的分类准确率,验证了本文所述故障诊断模型的有效性,试验结果证明本文提出模型对联合收割机装配质量检测的分类准确率可达99%,较现有模型具有更好的准确度与稳定性。In view of the lack of effective detection methods for assembly quality detection of combine,a fault diagnosis method based on fuzzy self-tuning particle swarm optimization(FSTPSO)variational mode decomposition(VMD)and Least squares support vector machine(LSSVM)is proposed.An optimized multi-scale reverse discrete entropy(OMRDE)is proposed for feature extraction,and the feature fusion is carried out with the time-frequency-domain features.The fault diagnosis model of FSTPSO-VMD-FSTPSO-LSSVM is established.The feature extraction effects of three entropy functions,OMRDE,multi-scale discrete entropy,and fuzzy entropy,are compared and analyzed.By comparing the classification accuracy of VMD-LSSVM,VMD-PSO-LSSVM,VMD-FSTPSO-LSSVM and Ensemble Empirical Mode Decomposition(EEMD)-FSTPSO-LSSVM,the effectiveness of the fault diagnosis model is verified,and the classification accuracy of the model for combine assembly quality detection is 93%.

关 键 词：联合收割机装配质量检测 模糊自整定粒子群算法 变分模态分解 优化多尺度反向离散熵 最小二乘支持向量机 

分 类 号：S225[农业科学—农业机械化工程]