基于BP神经网络的GD-1高压共轨系统的建模研究

Study on BP Neural Networks Modeling for GD-1 High Pressure Common Rail System

基于GD-1高压共轨燃油喷射系统,运用BP神经网络理论对GD-1系统高压油泵及共轨管进行建模,在Matlab平台上利用实际测得的数据对所建的神经模型进行训练,利用Simulink工具将训练好的高压油泵及共轨管模型与GD-1控制策略连接在一起进行闭环仿真,仿真结果表明设计的神经网络能很好地模拟共轨管内实际油压变化。

Based on the GD - 1 high pressure common rail fuel injection system, Model was built for GD - 1 High Pressure Fuel Pump and Common Rail in BP Neural Networks Types. The BP neural networks model is trained with practical data gained in experiments. Finally in the Simulink platform we connected the well - trained BP Neural Networks with GD - 1 fuel injection control strategy for closed loop simulation. The simulation result showed that the well - trained Neural Networks Model really simulates the characteristic of practical fuel pressure changes in common rail.