基于自适应模糊神经网络的发动机故障诊断

发动机是车辆的核心部件,及时有效地发现并排除故障,对降低维修费用,减少经济损失,增加发动机工作时的可靠性,避免事故发生具有重大的意义。以某型号发动机为研究对象,运用测试技术、信号处理、小波分析、神经网络和模糊控制理论,提出了自适应模糊神经网络发动机故障诊断。首先建立了发动机故障信号采集试验台,在试验台上人工模拟四种工况,通过加速度传感器采集正常工况和异常工况的振动信号。再利用小波理论对采集到的振动信号进行消噪处理,提高信噪比,并提取出故障信号的特征值,作为网络训练和测试的样本数据。用样本数据训练和检测自适应模糊神经网络,即对发动机故障进行模式识别。通过仿真分析,取得了很好的诊断效果;同时与传统的BP神经网络故障诊断方法进行对比,无论在诊断精度上还是学习速度上,模糊神经网络在故障诊断中更具有优势。

Engine Fault Diagnosis Based on Adaptive Fuzzy Neural Network

The engine is the core component of the vehicle, timely and effective manner to discover and troubleshooting, reduce maintenance costs, reduce economic losses, increase the reliability of the engine at work to avoid accidents have great significance. Taking a model engine for the study, with testing techniques, signal processing, analysis, neural networks and fuzzy control theory, and come out adaptive fuzzy neural network fault diagnosis engine.The paper established a fault signal acquisition engine test stand and four kinds of artificial conditions, the vibration signal acquisition normal operating conditions and abnormal operating conditions by the acceleration sensor, and then using wavelet theory collected vibration signal de-noising process, improve signal to noise ratio and extract the fault characteristic value of the signal sample data as the network training and testing. Fuzzy neural network training and testing using sample data adaptive, that is pattern recognition engine failure, through simulation, and achieved good diagnostic results. Compared with the traditional BP Neural Network diagnostic methods, both in learning speed or accuracy of the diagnosis, Fuzzy Neural Network has more advantages in fault diagnosis.