摘要
针对国内农业机械领域存在的联合收割机诊断技术起步较晚、精度不足等问题,提出一种多目标蜣螂算法与概率神经网络融合的诊断方法。首先构建多目标蜣螂算法,通过Tent混沌映射初始化传统蜣螂算法的种群,同时引入了t分布扰动策略和Levy飞行机制,进而平衡算法的全局、局部搜索速率,增强对复杂问题的求解能力。然后通过多目标蜣螂算法对概率神经网络的平滑因子进行自适应调整,提高模型的泛化能力和稳定性,更好地表征特征与模式间的关系。最终将建立的MTDBO-PNN模型应用于联合收割机的故障诊断,经过与其他常用方法的对比验证,表明该方法在诊断时间和精度上均有一定的优势。
Aiming at the domestic agricultural machinery field,the diagnostic technology of combine harvester starts late and the accuracy is insufficient,this study proposes a diagnostic method of fusion of multi-objective dung beetle algorithm and probabilistic neural network.Firstly,the multi-objective dung beetle algorithm is constructed,and the population of the traditional dung beetle algorithm is initialized by Tent chaotic mapping,and at the same time,the t-distribution perturbation strategy and Levy flight mechanism are introduced,so as to balance the global and local searching rate of the algorithm,and to enhance the ability of solving the complex problems.Then the smoothing factor of the probabilistic neural network is adaptively adjusted by the multi-objective dung beetle algorithm to improve the generalization ability and stability of the model.Finally,the model is applied to the fault diagnosis of combine harvester,and after comparative validation,it shows that the method has some improvement in both diagnosis time and accuracy.
作者
杨小强
刘波
刘金仙
YANG Xiaoqiang;LIU Bo;LIU Jinxian(College of Intelligent Manufacture,Chongqing Creation Vocational College,Chongqing 402160,China;School of Automation,Chongqing University,Chongqing 400044,China;Chongqing Changan Automobile Co.,Ltd.,Chongqing 400023,China)
出处
《中国工程机械学报》
北大核心
2025年第5期845-850,共6页
Chinese Journal of Construction Machinery
基金
国家自然科学基金资助项目(62003059)
重庆市教委科技资助项目(KJQN202005401)。
关键词
联合收割机
蜣螂算法
t分布扰动
Levy飞行
概率神经网络
故障诊断
combine harvester
dung beetle algorithm
t-distribution perturbation
Levy flight
probabilistic neural network
fault diagnosis
