《基于神经网络的汽车故障诊断系统》-毕业论文（设计）.doc

摘 要 随着汽车电子控制技术的快速发展，汽车故障诊断也不断变得智能化，电子控制系统由各种传感器，执行器和控制单元（ECU)组成，其中执行器的运行状态的好坏，直接影响到整车的工作状况。各种执行器的故障诊断在整个汽车的故障诊断中占有很重要的地位。课题在对目前国内外汽车执行器故障诊断方法进行广泛研究的基础上，把神经网络技术引入了汽车执行器故障诊断过程中。同时在汽车几种执行器工作原理的基础上，分析该执行器的故障诊断参数，从执行器电压、输出波形等信号中提取故障征兆，再将这些故障征兆归一化处理后作为输入神经网路的特征向量，利用事先训练好的BP神经网络对该执行器进行故障诊断，通过实例仿真，证明了BP神经网络在车用执行器故障诊断中具有较大的理论意义和应用价值。最后对几种典型的执行器的典型故障的实际测试，得到预期结果，验证了该诊断系统的性能，说明了所建立的故障诊断网络的有效性和实用性。 关键词：汽车；执行器；故障诊断；BP算法 ABSTRACT As the car the rapid development of electronic control technology, cars have been becoming intelligent fault diagnosis Electronic control system consists of a variety of sensors, actuators and control unit (ECU) components. Actuator operation in which the state directly affect the working conditions of the vehicle.A variety of actuator fault diagnosis fault diagnosis of the vehicle plays an important role. Issues in the implementation of the current domestic and international car fault diagnosis method based on extensive research, the neural network technology into the car during the implementation of fault diagnosis. Some in the car while working principle of the actuator based on the analysis of the fault diagnosis of the parameters of the actuator, through the implementation of voltage from the waveform to extract the signal failure and other symptoms, symptoms of these faults as input normalized neural networks Feature vector, the use of pre-trained BP neural network for fault diagnosis of the actuator.The BP neural network in motor fault diagnosis of the actuator has great theoretical significance and application value through simulation examples. To validate the diagnostic performance of the system ,we can have several typical actuator faults typical actual test for the system, the established fault diagnosis network is effective and practical. Key Words：Engine; Actuators ; Fault diagnosis; BP algorithm 目 录 TOC \o 1-4 \u 1 绪 论 PAGEREF _Toc6369 1 1.1课题研究的目的、意义 PAGEREF _Toc22640 1 1.1.1课题研究的目的 PAGEREF _Toc6334 1 1.1.2课题研究的意义 PAGER