森林消防车发动机电控系统故障诊断方法研究

Research on Fault Diagnosis Method of Electronic Control System of Forest Fire Engine

针对森林消防车发动机电控系统故障诊断问题,本文设计了基于数据融合的森林消防车电控发动机故障诊断方法。该方法诊断模型由基于反向传播神经网络(Back Propagation,简称为BP)的数据融合方法、基于概率神经网络(Probabilistic Neural Network,简称为PNN)的分类方法和基于DS(Dempster/Shafer,简称为DS)证据理论的决策融合方法组成。首先,根据原始数据样本训练BP神经网络以达到原始数据融合的目的,而后分别对经过数据融合的原始数据和未经数据融合的原始数据进行基于KL(Karhunen Loéve)变换的特征提取。然后,使用特征提取后的数据训练概率神经网络,并使用训练好的网络验证测试样本。最后,采用DS证据理论对初步诊断结论进行决策级融合。研究结果表明,基于数据融合和概率神经网络的方法可有效地提高森林消防车电控系统的故障诊断精度。

Aiming at the problem of fault diagnosis of electronically controlled engine of forest fire truck,this paper designed a fault diagnosis method of electronically controlled engine of forest fire trunk based on data fusion.The diagnosis model of this method consisted of data fusion method with BP neural network,classification method by probabilistic neural network and decision fusion method based on D-S evidence theory.First of all,the BP neural network trained by the data samples was to achieve the purpose of original data fusion.Next,the data with and without data fusion was performed a feature extraction operation by K-L transform.Then,the data after feature extraction was used to train the probabilistic neural network,and the trained network was used to verify the test samples.Finally,the D-S evidence theory was used to fuse the preliminary diagnosis conclusion at the decision level.The simulation results showed that the method based on data fusion and probabilistic neural network can effectively improve the fault diagnosis accuracy of the electronic control system for forest fire truck.