摘要
利用神经网络的非线性映射,及其高度的自组织和自学习能力,将改进的BP网络应用于柴油机的故障诊断。应用夹持式传感器获得柴油机喷射系统的燃油压力波形,对波形时域分析和特征提取,再根据所取得故障信息及其对应的故障类型来构造网络结构,应用附加动量的BP算法,从而实现对故障的分类。通过Matlab仿真理论表明,该方法可以有效地对故障进行识别分类。
The improved BP neutral network is introduced to the fault diagnosis by virtue of its nonlinear, high self-organize and learning abilities. Diesel engine fuel pressure wave shape is measured using clamp-on pressure transducer. By analysis of its shape and its characters, features are extracted from the fuel pressure wave, Based on the features and the kinds of faults, the network is constructed. An improved BP algorithrn is applied to the diesel fault diagnosis. The simulation result shows that this method is entirely practicable and reliable for the diesel fault diagnosis.
出处
《控制工程》
CSCD
2007年第5期518-521,共4页
Control Engineering of China
基金
国家自然科学基金资助项目(60374021)
关键词
柴油机
故障诊断
神经网络
燃油系统
diesel engine
fault diagnosis
neutral network
fuel oil system
