基于TLBO-LIBSVM的联合收割机振动筛螺栓故障诊断  

Fault Diagnosis of Vibrating Screen Bolts in Combine Harvester Based on TLBO-LIBSVM

作　　者：李鹏程 顾新阳 梁亚权 章浩 唐忠 Li Pengcheng;Gu Xinyang;Liang Yaquan;Zhang Hao;Tang Zhong(School of Agricultural Engineering,Jiangsu University,Zhenjiang 212013,China)

机构地区：[1]江苏大学农业工程学院,江苏镇江212013

出　　处：《农机化研究》2025年第5期28-33,42,共7页Journal of Agricultural Mechanization Research

基　　金：浙江省农作物收获装备技术重点实验室开放课题(2021KY02);国家自然科学基金项目(52275253)。

摘　　要：联合收割机振动筛工作时的瞬时冲击与交变载荷易导致振动筛螺栓结构发生失效。为解决联合收割机振动筛螺栓故障诊断问题,提出了一种基于多元特征融合TLBO-LIBSVM的振动筛螺栓失效故障诊断方法,通过提取特征矩阵,分别将时域特征、频域特征、WOA-VMD能量熵特征组合归一化得到多元融合高维特征矩阵,导入经验参数LIBSVM模型,得到的成功率分别为64.44%、74.44%、81.11%、90%。结果表明:随着特征矩阵维数不断增加,失效特征信息不断完善,识别成功率不断提升,也验证了联合收割机振动筛螺栓频域特征敏感性高于时域特征。通过运用TLBO算法对LIBSVM模型超参数进行优化,得到最佳参数组合下的识别成功率为98.89%,完成了联合收割机振动筛螺栓失效故障的高精度识别,可为联合收割机振动筛螺栓故障的精确诊断提供参考。Instantaneous impact and alternating load of vibrating screen of combine harvester during operation easily lead to failure of vibrating screen bolt structure.To solve the problem of fault diagnosis of vibrating screen bolts in combine harvester,and proposed a fault diagnosis method of bolt failure based on multiple feature fusion TLBO-LIBSVM.By extracting feature matrices,the time-domain features,frequency-domain features,and WOA-VMD energy entropy features were combined and normalized to obtain a multivariate fusion high-dimensional feature matrix,which was imported into the empirical parameter LIBSVM model.The success rates were 64.44%,74.44%,81.11%,and 90%,respectively.With the increasing dimension of the feature matrix,the failure feature information was constantly improved,and the recognition success rate was constantly improved.At the same time,it also verified that the frequency domain feature sensitivity of combine harvester vibrating screen bolts was higher than the time domain feature.By using TLBO algorithm to optimize the hyperparameter of LIBSVM model,the identification success rate under the optimal parameter combination was 98.89%,and the high-precision identification of bolt failure of combine harvester shale shaker was completed.This provided a reference for the accurate diagnosis of combine harvester vibrating screen bolt fault.

关 键 词：振动筛螺栓 变分模态分解 鲸鱼优化算法 支持向量机模型 教与学优化算法 故障诊断 

分 类 号：S225.3[农业科学—农业机械化工程]