基于FPCA与DEELM的弹药协调机械臂性能故障诊断

为了实现弹药协调机械臂定位精度超差的性能故障诊断,提出了一种基于函数型主成分分析(FPCA)与差分进化极限学习机(DEELM)结合的故障诊断方法.建立了协调机械臂的动力学解析模型,进行了标准状态下协调过程的仿真分析,同时对协调机械臂实验台架进行了相同状态的协调过程测试,二者输出的支臂角位移曲线吻合较好;利用协调机械臂的...

Performance Fault Diagnosis of Ammunition Transfer Arm Based on FPCA and DEELM

In order to realize the performance fault diagnosis that the positioning accuracy of the ammunition transfer arm is out of tolerance,a fault diagnosis method based on the combination of functional principal component analysis(FPCA)and differential evolution extreme learning machine(DEELM)was proposed. The dynamic analytical model of the transfer arm was established,and the simulation analysis of coordination process under the standard state was carried out. At the same time,the coordination process test under the same state was performed on the transfer arm experimental bench. The arm angular displacement curves of their output are relatively consistent. The critical value ranges of two fault parameters,the gas spring initial pressure and the angular displacement measurement error of the arm,were obtained by using the dynamic analytical model of the transfer arm,and different fault types were defined accordingly. Sample data of arm angular displacement were obtained through sampling simulation and simulated failure experiment within the parameter value ranges corresponding to different fault types. The arm angular displacement data were analyzed from the perspective of function,and expressed as smooth function curves. Functional principal component scores of the sample data were calculated using FPCA as feature parameters. The feature parameters extracted by FPCA and the corresponding classification labels were used as input and output information to train DEELM. Diagnosis tests were performed on simulation samples and experimental samples using the trained DEELM. The diagnosis accuracy rate is 98.10%. It shows that this method can realize the effective diagnosis of the performance fault of the transfer arm.