可变流量燃气发生器压强自适应控制方法研究  

Researching on Pressure Adaptive Control Method for Variable Flow Gas Generator

作　　者：周景亮 陈雄[1] 柴金宝 周长省[1] ZHOU Jing-liang;CHEN Xiong;CHAI Jin-bao;ZHOU Chang-sheng(School of Mechanical Engineering,Nanjing University of Science and Technology,Nanjing Jiangsu 210094,China)

机构地区：[1]南京理工大学机械工程学院,江苏南京210094

出　　处：《计算机仿真》2020年第7期113-118,218,共7页Computer Simulation

基　　金：总装备部预先研究项目(404040301);国家自然科学基金(51606098);江苏省自然科学基金(BK20140772)。

摘　　要：固体火箭冲压发动机燃气流量控制系统较难实现对压强的精确闭环控制。为解决该问题,提出了一种可变流量燃气发生器压强自适应控制方法——基于RBF(Radial Basis Function)神经网络系统辨识的燃气发生器压强自适应免疫粒子群PID控制方法。针对系统参变的特点,首先利用RBF神经网络在线辨识系统传递函数。过免疫算法(Immune Algorithm,IA)与粒子群算法的结合(Particle Swarm Optimization,PSO)进行PID参数寻优。利用RBF-PSO-IA控制器对某锥阀式流量控制系统在设计工作点下(10.32MPa)进行PID参数寻优与全压强在线调节进行仿真,验证控制器的性能。仿真结果显示,在设计点处PSO-IA、PSO、传统PID整定方法得到的压强响应均没有超调;PSO-IA算法的响应时间最短,即PSO-IA优化的PID控制器能够显著提高压强响应速度;全压强调节下,RBF神经网络辨识器能够有效的辨识传递函数,RBF-PSO-IA算法能够提高某锥阀式流量控制系统的稳态精度和动态性能。It is difficult for the gas flow control system of the solid punching engine to realize the precise closed loop control of the pressure.In order to solve this problem,a kind of pressure adaptive control method for variable flow gas generator was proposed,that is,the adaptive immune particle swarm optimization PID control for gas generator pressure based on the identification of radial basis function(RBF)neural network system.In view of the parameter change of the control system caused by time change,the RBF neural network was used to identify the transfer function of the system online,and the parameters of the transfer function were sent into the controller which was combined by immune algorithm(IA)and particle swarm optimization(PSO)to obtain the PID parameters.The RBF-PSO-IA controller was used to simulate the off-line search parameters and the full pressure on-line regulation of a conical valve flow control system under the design working point(10.32 MPa).The simulation results show that the pressure responses of PSO-IA,PSO and the traditional PID setting method are not overshoot;the pressure response time of the PID parameter value obtained by PSO-IA off-line search is the fastest,which means that PSO-IA can significantly improve the pressure response speed.Under the full pressure regulation,the RBF neural network identifier can effectively identify the transfer function.For the conical valve type flow control system,the RBF-PSO-IA algorithm is able to improve the steady-state accuracy and dynamic performance.

关 键 词：变流量 燃气发生器 系统辨识 免疫粒子群算法 自适应控制 

分 类 号：V448.15+3[航空宇航科学与技术—飞行器设计]