神经网络自适应PID控制在柴油机齿条位置控制中的应用

Application of PID Based on BP Neural Networks to the Control of Rack Displacement for Diesel Engine

提出了基于BP神经网络的PID控制算法在柴油机齿条位置控制系统中的应用,把传统的PID控制和BP(误差反向传播)神经网络控制相结合,采用"先离线学习,后在线控制"的思想实现了齿条位置闭环的自适应控制。初步实验结果表明,该控制算法能满足齿条位移控制的性能指标,且其稳态特性较好,克服静摩擦的效果令人满意。

The rack displacement has serious nonlinear and temporal characteristics in the working process of fuelinjection pump in diesel engine. A digital proportinintegraldifferential(PID) control algorithm based on backpropagation (BP) neural networks was developed. The control idea of 'studying firstly, then controlling on line' was adapted to rack positional closedloop system and adaptive control was carried out effectually. Experimental results prove that the control algorithm can resolve the performance of rack position control, improve the stability and overcome the static friction satisfactorily.