针对发动机在实际运行中经常会出现单缸或者多缸失火这种典型的故障现象,从发动机瞬时转速的角度出发,对发动机气缸受力与发动机曲轴瞬时转速之间的关系进行了理论研究,建立了其数学模型,利用Matlab仿真软件对瞬时转速曲线进行了仿真,确定并提取出了特征参数,比较了正常情况下和单缸不点火情况下瞬时转速曲线及其特征参数的变化,并把这些变化作为下一步诊断发动机失火故障的判断依据;最后,建立了一个BP神经网络并对其进行训练,同时把仿真数据和实例数据输入到网络中进行了理论验证和实例验证。研究结果表明,发动机的失火故障与发动机瞬时转速之间存在一定的关系,利用瞬时转速可以对发动机失火故障进行诊断;利用BP神经网络方法诊断发动机失火故障具有速度快、效率高的特点。
Aiming at the typical phenomenon of the engine misfire fault in the actual operation, the research was started from the instantane- ous crankshaft speed of the engine. The relationship between the engine misfire fault and instantaneous crankshaft speed was briefly de- scribed. A mathematical model was established and the feature parameters were determined. Then, the Matlab was used to simulate it and BP neural networks were used to detect and analyse the engine misfire fault. The results indicate that there is a certain relationship between the engine misfire fault and the instantaneous crankshaft speed of the engine. Using instantaneous crankshaft speed can diagnose the engine misfire fault. Using BP neural networks can also diagnose the engine misfire fault and it has the features of fast and efficiently.