摘要
基于AMEsim建立了牵引-制动型液力变矩器液压系统仿真模型,通过注入不同故障信息,分析变矩器进口压力不稳定的故障原因,生成了仿真故障样本。通过试验,采集待检测的液压故障样本,基于遗传神经网络以仿真故障样本为训练样本,对待检测故障样本进行了模式识别。试验结果表明,故障仿真模型可以较好的模拟液压系统故障,将其与遗传神经网络相结合可以实现故障的智能诊断。
The simulation model of hydraulic system of tractor-brake hydraulic torque converter was established based on AMEsim. By injecting various fault information into the model, characteristics of fault was calculated through simulation. At last, normal state and some fault state of were tested in test-bed. The results of test show that results of fault simulation model is accord with fault characteristics of tractor-brake hydraulic torque converter and based on simulation results, genetic neural network can recognized faults efficiently.
出处
《流体传动与控制》
2009年第4期1-4,共4页
Fluid Power Transmission & Control
基金
总装"十一五"预研基金资助项目(项目编号:40402060202)
总装重点实验室建设基金资助项目(项目编号:9140C3405070805)
