基于BPNN的发动机停机相位预测研究  

Forecast research of engine stop phase based on BPNN algorithm

作　　者：姚国仲[1] 徐小鸿 王贵勇[1] 邓冬荣 路璐[1] YAO Guozhong;XU Xiaohong;WANG Guiyong;DENG Dongrong;LU Lu(Yunnan Provincial Key Laboratory of Internal Combustion Engine,Kunming University of Science and Technology,Kunming 650500,China)

机构地区：[1]昆明理工大学云南省内燃机重点实验室,云南昆明650500

出　　处：《传感器与微系统》2023年第9期52-55,60,共5页Transducer and Microsystem Technologies

基　　金：国家自然科学基金资助项目(52066008)。

摘　　要：为避免柴油发动机启动阶段因建立喷油时序而在寻找信号特征齿上花费过多时间,基于反向传播神经网络(BPNN)建立了发动机停机相位预测模型。以某双缸柴油机停油时转速和负荷为输入,单片机和增量式编码器为核心计算的原有曲轴位置传感器“失信点”后的相对相位变化为输出,建立了BPNN停机相位预测模型。预测结果表明:模型决定系数和修正决定系数均大于0.91,平均相对误差为5.9%,模型对于发动机停机过程中不可信点至静止期间转过的相对角度具有预测性,可作为发动机再启动阶段相位快速同步和判缸的首选方式。In order to avoid spending too much time on finding signal characteristic teeth due to the establishment of fuel injection timing in the diesel engine starting phase,an engine shutdown phase prediction model is established based on back propagation neural network(BPNN).Taking the speed and load of a two-cylinder diesel engine as input when the oil is stopped,and the relative phase change as output after the original crankshaft position sensor“distrust point”which the TC275 microcontroller unit(MCU)and incremental encoder are the core,the BPNN is established to predict the phase of engine stop position.The prediction results show that the model determination coefficient and the modified determination coefficient are both greater than 0.91,and the average relative error is 5.9%.The model is predictive for the relative angle from the unreliable point in the engine shutdown process to the rotation period during the stationary period,which can be used as the first choice for quick phase synchronization and cylinder judgment in the engine restart phase.

关 键 词：反向传播神经网络 位置同步 增量式编码器 停机相位 

分 类 号：TP212[自动化与计算机技术—检测技术与自动化装置] TK427[自动化与计算机技术—控制科学与工程]