摘要
为了预测一定负载下,液压机械无级变速器(简称HMCVT)湿式离合器由于液压系统压力、流量脉动使得主从动轴长时间产生较大转速差而一直处于滑摩状态,导致湿式离合器换段不理想的问题。通过分析液压系统故障形成机理,利用AMESim搭建湿式离合器液压系统,模拟注入故障类型,进行时域特征提取,提出量子粒子群(QPSO)优化BP神经网络权值、阈值的算法对湿式离合器液压系统进行故障诊断,提升该系统的诊断效率与诊断精度,并结合台架试验验证该算法的准确性。研究结果表明:QPSO算法优化BP神经网络的故障诊断算法故障识别率高,算法具有更高的收敛精度及收敛速度,研究结果为设计拖拉机HMCVT的故障自诊断系统提供理论参考以及工程开发思路。
Aiming at the problem that under a certain load,due to the pressure and flow pulsation of the hydraulic system of the hydraulic mechanical continuously variable transmission(HMCVT)wet clutch,and the winner of the driven shaft for a long time to produce a large speed difference and being in the state of sliding friction,these lead to the wet clutch change is not ideal.By analyzing the fault formation mechanism of wet clutch hydraulic system,modeled hydraulic system failure using AMESim structures,and simulating the fault injection and extracting the time domain feature,the quantum particle swarm optimization(QPSO)algorithm of BP neural network weights and threshold of wet clutch hydraulic system fault diagnosis are put forward to improve the diagnostic efficiency and diagnostic accuracy of the system,and finally validating the accuracy of the algorithm though the bench experiment.The results show that the QPSO algorithm has higher fault recognition rate,higher convergence accuracy and higher convergence speed.The research results provide theoretical reference and engineering development ideas for the design of tractor HMCVT fault self-diagnosis system.
作者
王正幸
鲁植雄
陈元
WANG Zhengxing;LU Zhixiong;CHEN Yuan(College of Engineering,Nanjing Agricultural University,Nanjing 210031,China)
出处
《机械科学与技术》
CSCD
北大核心
2022年第6期849-856,共8页
Mechanical Science and Technology for Aerospace Engineering
基金
国家重点研发专项(2016YFD0701103)
南京农业大学科研启动基金项目(689-RCQD2003-0603)
拖拉机动力系统国家重点实验室开放课题项目(SKT2022006)。
