基于SOM算法和免疫神经网络的飞机燃油系统故障诊断

Fault Diagnosis for Aircraft Fuel System Based on SOM Algorism and Immune Neural Network[HS)]

摘要:

针对传统的飞机燃油系统故障诊断方法如硬件冗余方法和系统模型检测方法存在的飞机重量限制和难以建立精确数学模型的问题,设计了一种基于SOM算法和BP神经网络的故障诊断模型；首先,建立了系统故障诊断模型并对诊断原理进行了描述,然后,对故障征兆数据进行预处理,即先采用SOM算法进行连续属性离散化处理,再通过粗糙集互信息方法进行属性降维,以减少数据量和提高诊断效率；最后,建立了基于BP神经网络的故障诊断模型,为了进一步提高故障诊断精度,在采用免疫优化算法对BP神经网络故障诊断模型中的各参数即权值和阈值等进行优化的基础上,进一步采用BP反向传播算法进行参数调整,从而得到最终的故障诊断模型。通过飞机燃油系统故障诊断实例仿真实验证明了文中方法能较为精确地实现故障诊断,且与其它方法相比,具有较高的诊断精度和诊断效率,具有较大的优越性。

Abstract:

Aiming at the traditional fault diagnosis method has the defects such as the aircraft weight constraint and the difficulty for establishing the mathematical model,a fault diagnosis model based on SOM algorism and BP neural network was proposed. Firstly,the model of diagnosis for aircraft fuel system was given,then fault symptom data was preprocessed,namely,using the SOM algorism to discrete the data and rough set mutual information method to reduce the dimensions of the data,finally,the immune optimizing method firstly the BP error counter propagation secondly were used to optimize the parameters in BP neural network fault diagnosis model,therefore,the final fault diagnosis model was obtained. Through the Aircraft fuel system fault diagnosis was simulated,and the result shows the method in this paper can accurately realize fault diagnosis,and compared with the other methods,it has the high diagnosis accuracy and efficiency. It is proved has big priority.