Research On Simulation And Control Of A Gun Load Simulator

The development of war requires a lot in artillery.In the future,the artillery must have mobility,rapidity and accuracy,and be able to aim at and hit targets quickly and accurately.In the process of firing and turning,the load end of gun will be acting on complex and variable forces,which has an impact on the accuracy of the gun control system.The electric load simulator can be used to simulate the fickle torque at the load end in real time,evaluate the performance of the system,shorten the development cycle and reduce the cost.However,there are many complicated non-linear links inside the load simulator,which lead to surplus torque.In-depth research on the control strategy of the load simulator can improve the accuracy of the gun control system,which has important theoretical significance and application value.The main contents of the thesis are as follows:(1)Introduce the structure and working principle of the electric load simulator of the gun control system,and establish it's mathematical model.The causes of excess torque are analyzed by transfer function.The nonlinear factors in the system are studied and analyzed.(2)Due to the uncertainties and complex non-linear factors existing in the electric load simulator,it is difficult to obtain an accurate model,so RBF neural network is used to identify the system,witch has strong approximation ability.Gradient descent method and particle swarm optimization needle were selected as learning algorithms of RBF neural network respectively.Finally,the simulation results show that the RBF neural network optimized by particle swarm optimization is more accurate than the former in system identification.(3)Considering the invariance feature of sliding mode variable structure control when faced with external interference and parameter perturbation,it is used to control the electric load simulator.Integrated switching gain and fuzzy neural network were used to adjust the switching term of sliding mode respectively to deal with the chattering problem.Finally,MATLAB simulation experiment shows that the system under the fuzzy neural network sliding mode control strategy has higher torque loading accuracy,rapidity and stability.(4)This paper introduces and builds a semi-physical simulation platform to verify the control strategies.The conclusion is that the fuzzy neural network sliding mode control strategy has a higher control accuracy and meets the performance requirements of the system.