CAJ | 학술논문

针对非线性、时变等缺陷导致传统的控制器控制效果较差、不适应电液伺服系统的现象，提出了用于电液伺服控制的基于粒子群优化算法对神经网络的权值进行学习训练的PSO-NN算法。结合电液伺服系统实例分析，用MATLAB仿真得到了输入阶跃信号和正弦信号时，PSO-NN算法的输出曲线以及适应度曲线；为了展示PSO-NN算法的效果，用BP算法仿真了对应输入阶跃信号和正弦信号的输出。仿真结果表明：在电液伺服系统的控制中，PSO-NN算法性能优于BP算法，系统输出具有更好的收敛性和对输入的跟随性，从而证明PSO-NN算法对于电液伺服系统的控制是合适并有效的。
침대비선성、시변등결함도치전통적공제기공제효과교차、불괄응전액사복계통적현상，제출료용우전액사복공제적기우입자군우화산법대신경망락적권치진행학습훈련적PSO-NN산법。결합전액사복계통실례분석，용MATLAB방진득도료수입계약신호화정현신호시，PSO-NN산법적수출곡선이급괄응도곡선；위료전시PSO-NN산법적효과，용BP산법방진료대응수입계약신호화정현신호적수출。방진결과표명：재전액사복계통적공제중，PSO-NN산법성능우우BP산법，계통수출구유경호적수렴성화대수입적근수성，종이증명PSO-NN산법대우전액사복계통적공제시합괄병유효적。
Considering the control nonlinearity and uncertainties and other defects of intelligent control electro -hy-draulic servo system , PSO-NN algorithm is proposed to use particle swarm optimization algorithm to train neural network weights .Combined with a specific instance of electro-hydraulic servo system , when the system input is step and sine signal , the output and fitness curves of PSO-NN algorithm are simulated by using MATLAB software .In order to demonstrate the result of PSO-BP algorithm, BP algorithm is simulated further .The simulation results show that PSO-NN algorithm is superior to the BP algorithm due to the better output convergence and input following per -formance of the electro-hydraulic servo system , and that PSO-NN algorithm is suitable and effective for electro-hy-draulic servo system control .