最小发汗量冷却系统的PID进化控制策略和实现算法  

PID Evolutionary Control Strategy and Implementing Algorithm to Minimize Coolant Consumption in Transpiration Cooling System

作　　者：朱聪超[1] 秦世引[1] 杨学实[1] 

机构地区：[1]北京航空航天大学自动化科学与电气工程学院,北京100083

出　　处：《宇航学报》2008年第4期1336-1340,共5页Journal of Astronautics

摘　　要：针对发汗冷却系统中不可压缩发汗剂对热层温度场的优化控制问题,利用径向基函数神经网络建立了被控温度场的预测模型,构造了使发汗剂消耗量和预测温度超调量极小化的泛函性能指标,提出一种基于遗传算法的PID进化控制策略和实现算法。同时,该神经网络预测器可以通过在线学习随时跟踪温度场的变化,以实现对温度场的滚动优化和进化控制。仿真结果表明,这种基于神经网络预测和遗传算法优化的PID进化控制器能够实现在几乎无烧蚀的情况下发汗剂消耗量的最小化。而且,该算法还适合于多样化的性能指标,便于求解和实现,实用性强。Aiming at optimal temperature field control problems with uncompressible coolant in transpiration cooling system, a temperature predictive model based on RBF neural network is built and a performance index functional of coolant consumption and temperature overshoot is constructed, then a kind of PID evolutionary control strategy based on GAs is proposed firstly and its corresponding implementing algorithms are investigated in detail. Meanwhile, the neural network predictor has been learning by itself on line and tracking temperature in real time. As a result, a successively roiling optimization and evolutionary control can be carried out so as to realize the global control objective. A series of numerical simulations indicate that this kind of PID controller with prediction estimation and evolutionary optimization minimizes the coolant consumption while maintaining the porous material with nearly little ablation. Furthermore, this kind of strategy and algorithms can qualify to deal with some complex problems with diversifted performance indices and have strong practicability with their convenience and efficiency for solution and realization.

关 键 词：遗传算法 RBF神经网络 PID 发汗冷却 最优控制 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]