摘要
在某系列天然气发动机点火控制系统基础上,提出利用神经网络在线优化点火提前角,从而解决天然气发动机在使用过程中因磨损、老化等造成最佳点火提前角发生变化的问题,达到提高天然气发动机的动力性和燃气经济性的目的。首先通过Matlab/simulink建立的LNG发动机模型,编写标定程序,得出初始点火提前角,然后设计神经网络。优化策略为在稳定工况时,利用自适应神经网络在线修正,得出实时最佳点火提前角;在工况变化较快时直接读取MAP图,得到点火提前角。最后利用GT-POWER搭建天然气发动机模型进行仿真并与实测数据进行比较。
It proposes an online optimum neural network based on a series of nature gas engine electronic spark control system.It optimizes the ignition advance angle because the optimum ignition timing always change due to wear,aging etc.for improving the power and economy performance of natural gas engine.It firstly builds the LNG engine model by Matlab/simulink,writes calibration program,gets the initial ignition advance angle,then designs neural network for online optimum.The optimization strategy is using adaptive neural network online correction to get the optimum ignition advance angle in a stable condition,when conditions change rapidly,control system directly read the MAP figure to get ignition advance angle.It finally tests through GT-POWER and compare the real and simulation data.
出处
《机械设计与制造》
北大核心
2016年第6期57-60,共4页
Machinery Design & Manufacture
基金
山西省科技重大专项项目(20111101035)
关键词
LNG
发动机
点火提前角
神经网络
优化
仿真
LNG
Engine
Spark Advance Angle
Neural Network
Optimization
Simulation