Design And Control Algorithm Of Servo Control System For A Rocket Launcher

Rocket artillery is the main weapon of the modern Army which could attack enemy massive clusters in the long distance. The high fire accuracy and the quick reaction rates of rocket launcher is always the criterion of high-performance rocket artillery and it’s always the goal of many researchers. The performance of the servo system of rocket launcher can have a big influence on the fire accuracy and the reaction rates of rocket launcher. So the study of a high precision servo system is very important. This paper is major in it and it’s mainly about the identification and the control strategy of the high precision AC servo system which is used in a new rocket launcher.The working principle of rocket AC servo system is deeply analyzed in the paper. The thesis also derived the transfer functions of permanent magnet synchronous motors and servo control system. A detailed analysis of nonlinear and uncertainties that exist in the AC servo system was gave at the end of the second chapter.The paper introduces the system off-line identification scheme. Off-line identification use diagonal recurrent neural network (DRNN) as the algorithm. Genetic algorithm and differential evolution algorithm are used to optimize the initial weights of DRNN which could improve the convergence precision and convergence speed of DRNN. Through simulation experiments, we found that DE-DRNN has a better performance in the off-line identification study.Furthermore, the paper uses indirect neural network model reference adaptive control based on DRNN as the control strategy of the high precision AC servo system. Simulation result shows that this control strategy has high control accuracy and it has strong robustness. Outside interference has little effect on the system performance.Finally, the paper builds a semi-physical simulation test bench and conducted simulation experiment on it. Simulation result proves the correctness of the control strategy and the simulation studies.