基于PSO-RBF的航空相机电源板故障诊断技术研究

针对机载航空相机电源板故障率高，传统故障诊断方法技术不足而造成的相机维护难度大的实际问题，提出一种基于PSO-RBF神经网络的电源板故障诊断方法。考虑选取RBF网络训练算法中隐含层节点数和中心参数的难题，采用实用粒子算法约简了RBF神经网络，设计了航空相机电源板故障诊断系统方案，并给出了诊断系统的软件模块和实现方法，实现了从计算机仿真到工程应用的电源电路故障诊断。仿真与实际检测实验结果均表明，系统在不依赖任何标准设备和附加测点时，可对航空相机电源板进行实时、全自动化故障检测，其故障现象的检测覆盖率为100%，故障诊断平均可靠性可达到97.73%，故障器件定位率可达到96.89%。

PSO-RBF Based Fault Diagnosis Technology of Aviation Camera Power Supply

Aiming at the high fault rate of supply board of airborne aerial camera, the performance of traditional fault diagnosis is not good, which make camera maintenance difficult, so the PSO- fault diagnosis method to power board based on RBF neural network is presented in this paper. Considering the problem to select the hidden layer node num-ber and the center parameters of RBF network training algorithm,through simplifying the RBF neural network by prac-tical particle algorithm ,the system scheme of fault diagnosis system is designed that includes the software module and the implementation method of this system, which makes it became reality that is from the computer simulation to the engineering application. The simulation and actual test results shows that the system can make a real-time, automatic faults detection, which does not rely on any standard equipment and additional measurement point. The coverage rate of detected faults is 100%,the average reliability of fault diagnosis can reach 97.73%,the location rate to fault devices can reach 96.89%.