基于BP-ANN的光伏发电功率超短期预测方法  

Ultra-short-term Forecasting Method for Photovoltaic Generation Output Power Based on BP-ANN

作　　者：常学飞[1] 杨薏霏 王志煜[3] 李德鑫[1] 袁野[1] 

机构地区：[1]国网吉林省电力有限公司电力科学研究院,长春130021 [2]国网长春供电公司,长春130021 [3]国网通化供电公司,吉林通化134001

出　　处：《吉林电力》2014年第3期24-26,40,共4页Jilin Electric Power

摘　　要：针对光伏发电的间歇性和波动性问题,采用基于BP-ANN建立光伏发电系统输出功率超短期预测模型,利用输出功率的历史值、过往及预测日气象信息,对输出功率进行预测,并提出适用的预测流程及预测误差评估方法。实际应用以及与实时监测数据对比,表明该方法方法误差较小,合格率较高,能够满足应用的要求。Aiming at the intermittent and fluctuation issue in photovoltaic （PV） generation, an uhra-short-term prediction model based on BP-ANN is proposed to forecast the output power of PV system. Historical output data , the past and forecasted meteorology information have been adopted to forecast the output power. Applicable forecasting process and error evaluation method have been proposed. Practical application and comparison with real-time monitoring data proves that this method shows low error, higher qualification rate , which could meet the application requirements.

关 键 词：光伏发电 功率预测 超短期 BP-ANN 

分 类 号：TM615[电气工程—电力系统及自动化] TM715.1