基于BP-PID算法的柴油机调速复合控制策略的应用研究

BP Neural Network and PID for the Application of the Engine Electronic Speed Control

随着电子控制技术的不断发展,在柴油机电子调速过程当中,其被控系统逐渐趋向于复杂化和非线性化,进而导致了传统PID控制表现出一定的不适应性。我们提出了一种误差反向传播算法(BP神经网络)与传统PID结合的复合控制策略,利用BP神经网络的自适应能力和自学习能力,对PID控制器的三个参数进行在线实时整定。并以D6114柴油机为控制对象,基于dSPACE半实物仿真平台,分别对传统PID与BP-PID控制进行了仿真和配机试验。通过结果的对比性分析,验证了BP-PID复合控制策略在非线性、时变性复杂工况下,能够有效地降低负载变化对转速的影响,其抗干扰性、稳态性能更好,可以用于实际柴油机转速控制系统研究开发。

With the continuous development of electronic control technology, traditional PID control is no longer adapted to the diesel engine speed control because of its gradually tend to be complex and nonlinear. A composite control strategy of the error back propagation algorithm （BP neural network） combining with the traditional PID has been developed by this paper, which means the three parameters of the PID controller can be tuned online real-time depending on the adaptive capabilities and self-learning ability of BP neural network. In addition, the machine test and simulation of the traditional PID and the BP-PID have been developed respectively based on the dSPACE semi-physical simulation platform and D6114 diesel engine. It＇s verified that the BP-PID hybrid control strategy has better interference immunity, steady state performance and can effectively reduce the impact on the speed of load changes as well as its practicality on the research and development of diesel engine.