【摘要】 提出了一种基于飞轮转速波动信号结合神经网络进行内燃机失火故障诊断的方法。对电涡流位移传感器拾取的飞轮位移信号直接进行时域采样，通过软件快速处理可以获得足够测量点数和测量精度的内燃机转速波动信号。ＢＰ神经网络以此作为输入向量，可以快速准确地对内燃机失火气缸进行识别和定位。试验结果表明，该方法具有良好的效果和工程实用性，抗噪声干扰能力和工况适应性很强，并能满足实时诊断的要求。更多还原

【Abstract】 A unique technique has been developed for fault diagnostics of internal combustion engines by flywheel speed fluctuations and neural networks.The speed fluctuation signals are simply measured with an eddy current displacement sensor opposite the flywheel gear teeth.Using such signals as the input vector,the BP (backpropagation) neural networks can easily give the information on engine faults location and extent.Experimental results in a real world engine are presented and verify the theoretical developments.It is shown that this technique is simple to implement and may provide practical utility for real time engine diagnostics.更多还原