Abstract: In order to diagnose the faults of aircraft hydraulic system effectively, a method based on entropy weight and ABC-BP neural network is proposed, which is according to the signal of hydraulic system pressure. In this model，extraction of eigenvalues of aircraft hydraulic system pressure signal is the first step, and then, it calculates eigenvalue information entropy according to entropy weight method, the bigger of results as the input of the neural network, and in this paper, BP neural network is optimized by artificial bee colony through replacing the artificial bee colony fitness with the error function of BP neural network, finally, selecting the best fitness individual parameters as the weights and thresholds of the neural network ,this method not only reduces the input dimension of the model, but also improves the diagnostic accuracy.The simulation model of the landing gear retractable control system is established. The simulation results show that the diagnosis model has better fault diagnosis effect, and provides a new idea for the faults diagnosis of aircraft hydraulic system.

Key words: aircraft hydraulic system, entropy weight method, information entropy, artificial bee colony, Back Propagation（BP） neural network, fault diagnosis

摘要： 为有效诊断飞机液压系统故障，根据液压系统压力信号采用了熵权ABC-BP神经网络的故障诊断模型。模型先提取飞机液压系统压力信号的特征值，根据熵权法计算特征值信息熵，选取熵权值较大的作为神经网络的输入，同时利用人工蜂群优化BP神经网络，将BP神经网络的误差函数作为人工蜂群的适应度，选择适应度最优的个体参数作为神经网络的权值和阈值，不仅降低模型输入维度，还提高了诊断精度。最后建立了飞机起落架收放系统仿真模型进行仿真研究，结果表明该诊断模型具有较好的故障诊断效果，为飞机液压系统故障诊断提供一种新思路。

关键词: 飞机液压系统, 熵权法, 信息熵, 人工蜂群, 反向传播（BP）神经网络, 故障诊断