CAJ | 학술논문

针对炮控系统电动负载模拟器存在的摩擦、间隙、弹性形变、对象参数时变和位置扰动等复杂非线性,传统的控制方法难以得到良好的动静态性能指标.结合电动负载模拟器系统组成和工作原理,建立了加载数学模型,利用炮控系统位置控制信号进行前馈补偿,设计了RBF神经网络控制器,并采用改进遗传算法对控制器的权值、节点和中心矢量等参数进行优化.实验结果表明:该控制策略能够有效抑制多余力矩,保证了系统静、动态加载时的控制精度和稳定性.
침대포공계통전동부재모의기존재적마찰、간극、탄성형변、대상삼수시변화위치우동등복잡비선성,전통적공제방법난이득도량호적동정태성능지표.결합전동부재모의기계통조성화공작원리,건립료가재수학모형,이용포공계통위치공제신호진행전궤보상,설계료RBF신경망락공제기,병채용개진유전산법대공제기적권치、절점화중심시량등삼수진행우화.실험결과표명:해공제책략능구유효억제다여력구,보증료계통정、동태가재시적공제정도화은정성.
For the complex nonlinearities of friction,clearance,elastic deformation,time-varying performance of the target parameters and position disturbance are existed in electric-driven load simulator of the gun control system,the conventional con-trol method can′t achieve the good static and dynamic performance indexes. In combination with the system composition and working principle of the electric-driven load simulator,the loading mathematical model was established. The RBF neural net-work controller(RBFNNC)was designed by using the position control signal of the gun control system to conduct with feedfor-ward compensation. The parameters of the controller′s weight,nodes and center vector are optimized by the improved genetic al-gorithm. The experimental results show that this control strategy can restrain the extra torque effectively,and ensure the control precision and stability when the system is loading in static or dynamic state.