自动弹仓虚拟样机修正的故障定位与诊断  

Fault location and diagnosis for the modification of the virtual prototype of an automatic shell magazine

作　　者：赵抢抢[1] 侯保林[1] 

机构地区：[1]南京理工大学机械工程学院,江苏南京210094

出　　处：《哈尔滨工程大学学报》2017年第3期440-445,477,共7页Journal of Harbin Engineering University

基　　金：国家安全重大基础研究计划项目(6132490102)

摘　　要：针对自动弹仓定位精度超差的故障的情况,本文提出了一种基于标准状态实验与虚拟样机修正相结合的故障诊断方法。对自动弹仓在运行过程中的弹筒位移及主动轮转角进行测试,同时建立自动弹仓的虚拟样机模型;以测试结果为标准,对建立的虚拟样机模型进行了修正;对弹仓定位精度超差的故障机理进行了分析,以修正后的模型进行大量仿真,利用主元分析法对仿真结果和测试结果进行分析,将分析所得故障特征用于极限学习机(ELM)的训练以产生自动弹仓的诊断机。诊断结果说明,此方法切实有效,为自动弹仓的故障诊断提供了一种新的技术方法。Aiming at the situation that the location accuracy of an automatic shell magazine is often out of tolerance, a fault diagnosis method was proposed by combining the tests under standard state with the modification of a virtual prototype. The displacement of the magazine and the rotation angle of the driving wheel during the running process were tested,and a virtual prototype model of the automatic shell magazine was built. T he model of the virtual proto-type was modified by the test data. The fault mechanism of out- of-tolerance location accuracy was analyzed. Several simulations were carried out for the modified model. T he simulation results and testing results were analyzed by principal component analysis （PCA ）. Finally,the fault eigenvalues from PCA were used to train the extreme learn-ing machine （ELM ） to generate a diagnosis machine for the automatic shell magazine. T he diagnosis results show that this method is effective and satisfactory,providing a new method for the fault diagnosis of an automatic shell magazine.

关 键 词：自动弹仓 模型修正 故障诊断 主元分析 极限学习机 虚拟样机 

分 类 号：TJ307[兵器科学与技术—火炮、自动武器与弹药工程]