Research On Identification And Suppression Technique About Superfluous Torque On Servo Load Simulator In A Weapon

Servo system dynamic performance is the main factors of a weapon, which impacted of heavy artillery (rockets) mobility, automation exercise aiming and shooting rapidity. Servo system load simulator can dynamically in real-time simulation load of heavy artillery fire turned the force during load changes in the assembly of weapons before the servo system dynamic performance test. However, the superfluous torque in load simulator can cause great disturbance on the load torque. Therefore, the formation mechanism of the superfluous torque, identification and suppression technology researched to improve the system load simulator servo torque loading precision, has important theoretical significance and application value.The main research work of this dissertation focuses on the following aspects:1. The structure and working principle of load simulator is deeply analyzed of servo system. Hydraulic valve motor servo system to perform part of the transfer function, the transfer function of the load torque and the superfluous torque transfer function can be derived by application servo valve flow continuity equation and motors, load moment equilibrium equation, and the electrical part of the mathematical equation, respectively. Moreover, the state space model of servo load simulator is established, which providing theoretical support for the designed of load torque controller.2. According to the hydraulic servo control system structure diagram, on the basis of checking component parameters, load simulator physical model can be obtained by using the AMESim simulation software platform, provides an experimental platform for the identification and prediction of superfluous torque. The key technologies of co-simulation between in the Matlab/Simulink and AMESim are studied, and the establishment of a joint servo load simulator simulation model to load torque controller is designed to provide an experimental platform.3. Neural network based on error saturation prevention function, T-S fuzzy model based on cluster analysis and grey forecast model are studied in this paper for the identification and prediction of superfluous torque. Combined with the test data and evaluation indicators, the three kinds of methods identification results and prediction accuracy are compared and summarized. The results show that the simple structure of gray prediction model and a small amount of calculation can be used online to predict superfluous torque.4. Load torque controller designs for the problems faced in the analysis of fuzzy logic and conventional PID control system structure and characteristics, based on in-depth study of the dynamic fuzzy neural network GD-FNN and gray prediction variable universe fuzzy adaptive PID controller the working principle, structure and parameter learning algorithm is designed based on GD-FNN controller and a load torque on gray prediction variable universe adaptive fuzzy PID controller load torque, combined with AMESim simulation platform for the suppression simulation about superfluous torque.5.The servo control software of load simulator is designed and the experiments are conducted for the servo system of a weapon, two load torque controller for the load test results show that the proposed GF-PID controller is able to suppression the magnitude of the superfluous torque from the load torque, can be used to the servo load system. Subsequently, the construction of the "Universal Servo System Load Simulation Test System" as the basis for a certain type of weapon servo system with dynamic and static performance evaluation conducted to verify. The results show that the static turned precision, constant tracking error and tracking MSE equivalent sine meet the design requirements.