基于CVFOA-GRNN的飞机液压系统的故障诊断研究  

Research on Fault Diagnosis of Aircraft Hydraulic System Based on CVFOA Optimized GRNN Network

作　　者：齐鹏[1] 柴佳佳 靳小波[2] 

机构地区：[1]中国飞行试验研究院,陕西西安710089 [2]航空工业北京长城航空测控技术研究所,北京100022

出　　处：《测控技术》2017年第12期43-47,共5页Measurement & Control Technology

基　　金：国防基础科研计划资助项目(JCKY 2016205A004)

摘　　要：针对飞机液压系统故障具有随机波动性和非线性的特点,基于仿真获取故障和正常数据建立液压系统故障诊断模型。因果蝇优化算法(FOA)寻优易陷入局部最优,改进果蝇优化算法的初始值散列方式和寻优步长,构建混沌变步长果蝇优化算法。通过改进的果蝇优化算法优化广义回归神经网络(GRNN),提高GRNN的非线性学习能力,最终构建CVFOA-GRNN(chaotic variable step size fruit fly optimization algorithm GRNN)模型。实验表明相比FOA-GRNN、GRNN和BP模型,本文模型在性能上更稳定、收敛更快,应用于液压系统故障诊断准确度更高,具有实用价值。Aiming at the characteristics of random fluctuation and non-linearity of aircraft hydraulic system fault, a fault diagnosis model of hydraulic system is established based on fault data and normal data obtained by simulation. The hash method of the initial value and the steps of the search of the fruit fly optimization algorithm( FOA) are optimized. The optimization algorithm of chaotic variable step size fruit fly( CVFOA) is established. The generalized regression neural network(GRNN) is improved by the improved FOA, and the nonlinear learning ability of GRNN is improved. Finally, CVFOA-GRNN model is established. The experimental results show that compared with FOA-GRNN, GRNN and BP models, the model is more stable and faster convergence, and the hydraulic system fault diagnosis is more accurate. It has practical value.

关 键 词：液压系统 故障诊断 果蝇优化算法 广义回归神经网络 

分 类 号：TH137[机械工程—机械制造及自动化]