发电技术 ›› 2018, Vol. 39 ›› Issue (6): 566-573.DOI: 10.12096/j.2096-4528.pgt.18050

• 新能源 • 上一篇    下一篇

一种基于神经网络的硅基光伏组件运行温度在线软测量方法

于航1(),刘阳1,连魏魏2,朱红路2()   

  1. 1 龙源(北京)太阳能技术有限公司, 北京市 西城区 100034
    2 华北电力大学可再生能源学院, 北京市 昌平区 102206
  • 收稿日期:2018-09-10 出版日期:2018-12-31 发布日期:2018-12-28
  • 作者简介:于航(1976),男,硕士,高级工程师,研究方向为风能及太阳能利用技术, yuhang@clypg.com.cn|朱红路(1982),男,讲师,研究方向为光伏电站功率预测、智能运维及能量管理, hongluzhu@126.com
  • 基金资助:
    国家重点研发计划项目(2017YFB0902100)

An Online Soft Sensing Approach of Operating Temperature for Silicon-Based Photovoltaic Module Based on Neural Network

Hang YU1(),Yang LIU1,Weiwei LIAN2,Honglu ZHU2()   

  1. 1 Longyuan(Beijing) Solar Technology Co., Ltd., Xicheng District, Beijing 100034, China
    2 Renewable Energy School, North China Electric Power University, Changping District, Beijing 102206, China
  • Received:2018-09-10 Published:2018-12-31 Online:2018-12-28
  • Supported by:
    National Key Research and Development Program of China(2017YFB0902100)

摘要:

硅基光伏组件的运行温度对组件电气性能和发电效率具有显著影响,是光伏系统建模和性能评估的重要参数,它的精确计算对于光伏系统分析和最大功率跟踪等算法的应用等具有重要意义。通过对组件运行温度经典计算方法进行实例验证,发现该算法在不同季节、天气条件下的精度不一致,使得光伏电站发电量的计算与实际状况存在较大误差。针对这一问题,提出了一种基于多层反向传播(back propagation,BP)神经网络的硅基光伏组件运行温度在线建模方法,它分别以实测太阳辐照度、环境温度和输出功率作为模型输入,以组件运行温度作为模型输出,实现了组件运行温度的在线软测量。通过实际运行数据的对比表明上述方法是有效的,在条件允许时,也应该将风速作为模型输入之一。

关键词: 光伏发电, 光伏组件运行温度, 神经网络, 在线建模

Abstract:

The operating temperature of silicon-based photovoltaic (PV) modules has a significant impact on the electrical performance and power generation efficiency, which is an important parameter of PV system modeling and performance evaluation. Its precise calculation is very important for PV system analysis and maximum power tracking. Based on physics-based method of the operating temperature of silicon-based PV modules, it is found that the accuracy of the algorithm always changes with seasons and weather conditions, which resulted in a big error between the calculated and actual output of the PV plant. Aiming at the above problem, the paper proposes an online modelling method for the operating temperature of silicon-based PV modules, which adopts the multi-layer back propagation artificial neural network (BP-ANN) algorithm and uses the measured solar irradiance, ambient temperature, output power of PV modules as the inputs and the PV module operating temperature as the model output. And the online soft sensor of the operating temperature of the silicon-based PV module is realized. The proposed method is verified based on the actual operating data, whose comparative results show that the above method is effective. If conditions permit, wind speed should also be one of the inputs to the model.

Key words: PV power generation, operating temperature of PV modules, neural network, online modelling