基于神经网络滑模控制光伏系统最大功率点跟踪  

MAXIMUM POWER POINT TRACKING OF PHOTOVOLTAIC SYSTEM BASED ON NEURAL NETWORK SLIDING MODEL CONTROL

作　　者：阳同光[1,2] 桂卫华[2] 

机构地区：[1]湖南城市学院信息科学与工程学院,益阳413000 [2]中南大学信息科学与工程学院,长沙410083

出　　处：《太阳能学报》2016年第9期2386-2392,共7页Acta Energiae Solaris Sinica

基　　金：国家自然科学基金重点资助项目(61321003);国家自然科学基金(51237003)

摘　　要：在光照强度和环境温度变化的情况下,难以有效跟踪太阳电池的最大功率点。针对这个问题提出一种基于神经网络滑模控制技术的最大功率点跟踪方法。首先建立以太阳电池输出功率为状态量的数学模型,并选择实际输出功率、理想光照和温度下输出功率的差值构造滑模面。然后为消除时变和非线性不确定对控制系统的影响,利用RBF神经网络逼近滑模控制器的不确定部分,并通过Liyapulov稳定性理论求取RBF神经网络权值的自适应律。仿真和实验结果表明:该方法能同时实现光伏发电系统的最大功率点跟踪和变流器控制,具有良好的鲁棒性。Due to the varying environmental condition such as temperature and solar irradiation, it is difficuh to track the maximum power point of photovohaic cells effectively. Aiming at this problem, a method based neural network sliding mode control technology is implied to track the maximum power point. At first, a mathematical model with output power as the state variable of the photovohaic cells is established, and the error between actual output power with the output power under ideal solar irradiation and temperature is selected as the sliding mode surface. Then, to eliminate the time- varying and nonlinear uncertain impacts on control system, a RBF neural network is used to approximate the uncertain part of sliding mode controller, and the weights of RBF neural network are calculated by Liyapunov stability theory. The simulation and experimental results showed that this method can track the maximum power point of photovohaic power systems, control the converter effectively and has good robustness.

关 键 词：太阳电池 最大功率点跟踪 神经网络滑模控制 鲁棒性 

分 类 号：TK513.5[动力工程及工程热物理—热能工程]