CAJ | 학술논문

针对传统光伏系统MPPT控制算法的局限性及无法自适应外部复杂情况等诸多问题，提出一类以神经网络、模糊控制器以及PID控制器组成的神经网络模糊PID控制器。利用光照幅度、环境温度等参数的离线训练后的权值作为整个三层前馈神经网络的优化参数；通过预测最大功率点与实际的工作电压进行比较，运用模糊推理对PID相关参数进行最佳调整。仿真结果表明：与传统PID控制器、模糊控制器相比，本系统能增强消除系统误差能力，稳态性能有了明显提高，同时可获得更高的控制精度。
침대전통광복계통MPPT공제산법적국한성급무법자괄응외부복잡정황등제다문제，제출일류이신경망락、모호공제기이급PID공제기조성적신경망락모호PID공제기。이용광조폭도、배경온도등삼수적리선훈련후적권치작위정개삼층전궤신경망락적우화삼수；통과예측최대공솔점여실제적공작전압진행비교，운용모호추리대PID상관삼수진행최가조정。방진결과표명：여전통PID공제기、모호공제기상비，본계통능증강소제계통오차능력，은태성능유료명현제고，동시가획득경고적공제정도。
In this paper, aiming at the limitations of traditional photovoltaic system Maximum Power Point Tracking (MPPT) control algorithm and the problems of its no external complex adaptability,a class of neural networks fuzzy proportion integration differentiation (PID) controller was proposed,which comprise neural network,fuzzy controller and PID controller. The off-line trained weights of the light amplitude, environmental temperature and other parameters are used as the optimized parameter of the whole three layer feedforward neural networks;the fuzzy inference is used to optimally adjust the PID parameters by comparing the predicted maximum power point with the actual ones. The simu-lation results show that the system has better steady state and control precision, it does better in eliminating the sys-tem error than the traditional PID controller and fuzzy controller.