基于并行禁忌神经网络和DS证据的飞机燃油系统故障诊断

Fault Diagnosis for Aircraft Fuel System Based on Parallel Tabu Neural Network and D-S Evidence Theory

摘要:

飞机燃油系统是一个由许多相互联系的子系统构成的复杂总体,因而易于发生各类故障,当故障发生时会造成严重影响,为此,设计了一种基于禁忌神经网络和DS证据的飞机燃油系统故障诊断方法;首先,建立了飞机燃油系统的故障诊断模型,然后,建立了3层的BP神经网络故障诊断模型,并采用禁忌优化算法对BP神经网络进行参数优化,得到多个并行诊断的禁忌神经网络,输入样本数据对其训练并利用BP反向传播算法再次调优；最后将测试样本数据输入各禁忌神经网络,并将诊断结果作为证据采用DS证据理论进行融合,得到最终的故障诊断结果;实验结果表明:引入DS证据理论的故障诊断方法能有效克服单一故障诊断方法无法精确诊断故障的不足,诊断精度高,具有较大的优越性。

Abstract:

Aircraft Fuel System is a compound system with many interconnecting sub-systems and easy emerging many kinds of faults, when the fault happening, a fault diagnosis model based on Tabu neural network and D-S evidence is proposed. Firstly, the model of diagnosis for aircraft fuel system is built, and the tabu algorism is used to optimize the parameters of the neural network and get concurrent tabu neural networks, the training sample data is input to the tabu neural networks to train them and the BP back propagation algorism is used to adjust the parameters of the network, finally, the test sample data is input to all the tabu neural networks to diagnose and the result is got as the evidence, then the DS evidence theory is used to fuse the data to get the final diagnosis result. The simulation experiment shows the method in this paper solve the problems of diagnose the fault accurately for single diagnosis method, has the high diagnosis accuracy, so it has big priority.