摘要
针对模拟光照强度随外界环境快速变化而逐级下降的情况,由于电导增量法的步长固定,无法快速响应并完成对光伏电池最大功率点的跟踪,因此提出了神经网络跟踪最大功率的算法。该算法改进了传统电导增量法,优化了最大功率点跟踪(MPPT),将光照强度与温度作为输入变量,并为神经网络训练设置2层隐藏层,输出光伏最大功率下的参考电压。仿真结果表明,该算法的跟踪精度高达98.5%以上,能显著提高变化光照强度下光伏系统的能量转化效率。
In view of the situation that the light intensity decreases rapidly with the external environment,due to the fixed step size of the conductance increment method,it is impossible to respond quickly and track the maximum power point of photovoltaic cells,so the algorithm of tracking the maximum power of neural network is proposed.The algorithm improves the traditional conductance increment method,optimizes the maximum power point tracking(MPPT),takes the light intensity and temperature as the input variables,and sets two hidden layers for the neural network training,and outputs the reference voltage under the maximum photovoltaic power.The simulation results show that the tracking accuracy of the algorithm is more than 98.5%,which can significantly improve the energy conversion efficiency of the photovoltaic system under the changing light intensity.
作者
崔競文
吉宇
CUI Jingwen;JI Yu(Nanjing Institute of Technology,Nanjing,Jiangsu 211167,China)
出处
《自动化应用》
2024年第13期146-149,153,共5页
Automation Application
