基于日类型及融合理论的BP网络光伏功率预测

来源期刊:中南大学学报(自然科学版)2018年第9期

论文作者:夏向阳 冉成科 杨明圣 ZHANG Zhen-zhen(张真真) 李延和 曾小勇 黄海 滕欣元 蔡昱宽 曹伯霖

文章页码:2232 - 2240

关键词:光伏功率预测;日类型;相关系数;信息融合;误差修正

Key words:photovoltaic power forecast; daily type; correlation coefficient; information fusion; error correction

摘    要:针对光伏发电受到外界环境的制约、发电功率波动较大、很难保证高比率接入等对电网的安全运行和调度造成一定影响的问题,提出一种基于日类型及融合理论的BP网络预测方法,将不同的天气大体分为3种日类型即晴天、多云天、雨天,并进行分类预测,充分考虑制约光伏发电的5个最主要因素即光照强度、环境温度、组件温度、风速和相对湿度,找出其与发电功率之间的相关系数,通过信息融合理念将其融合成1个综合影响因子λ。以BP神经网络模型为构架进行功率预测,通过动态变换隐含层层数提高预测精度,并利用改进的粒子群算法对其参数进行优化,同时通过光伏历史功率输出波动特性对预测误差进行修正,最后在某县光伏电站进行实际验证。研究结果表明:该网络预测方法可行有效,且精度较高。

Abstract: Considering that photovoltaic power generation is constrained by the external environment, and the fluctuations in power generation is large, and it is difficult to ensure high-rate access, which affect the safe operation and dispatch of the power grid, a BP network prediction method was proposed based on daily types and fusion theory. It divided different weathers into three daily types, and classification predictions were made for the three daily types. Five most important factors that restricted photovoltaic power generation were fully considered, and the correlation coefficient between the factors and power generation was found out, and information was integrated into a comprehensive influencing factor λ. The power forecast was made based on the BP neural network model. and the prediction accuracy is improved by dynamically changing the number of hidden layers. The parameters were optimized based on particle swarm algorithm and the prediction error is corrected by the historical power output fluctuation characteristics of the photovoltaic. Finally, the method was verified at a photovoltaic power station .The results show that the network prediction method is feasible and effective, and the accuracy is high.

相关论文

  • 暂无!

相关知识点

  • 暂无!

有色金属在线官网  |   会议  |   在线投稿  |   购买纸书  |   科技图书馆

中南大学出版社 技术支持 版权声明   电话:0731-88830515 88830516   传真:0731-88710482   Email:administrator@cnnmol.com

互联网出版许可证:(署)网出证(京)字第342号   京ICP备17050991号-6      京公网安备11010802042557号