摘要
分析了柴油机气缸盖系统激励和响应信号的非线性特性,阐述了基于BP神经网络的柴油机气缸压力识别方法.对测试的振动响应信号进行时域统计平均和低通滤波后,训练BP神经网络,利用自适应梯度下降算法,自适应调节学习速率,提高网络精度,识别柴油机气缸压力.结果表明,恢复出来的缸内压力信号和实测信号十分接近,该方法对柴油机的实时在线控制、监测和故障诊断有重要的应用价值.
The nonlinear characteristics of the excitation and vibration response of cylinder cover are analyzed in this paper. A diesel engine cylinder pressure recovery method based on BP neural network is expounded. It is proposed to process the vibration acceleration signal induced by combustion pressure with denoising method of time statistic average, and remove high frequency noise using a lowpass filter, then train BP neural network using the adaptive grads algorithm, adaptively adjust the learning rate, improve the precision of the network, finally identify the cylinder pressure. The results show that the recovered cylinder pressure is close to the raw pressure signal. This method is valuable for real time controlling, condition monitoring and fault diagnosis of diesel engines.
出处
《武汉理工大学学报(交通科学与工程版)》
2007年第2期209-211,共3页
Journal of Wuhan University of Technology(Transportation Science & Engineering)
基金
国家自然科学基金项目资助(批准号:50275109)
关键词
自适应
神经网络
气缸压力
adaptive
neural network
cylinder pressure