基于修正晴空模型的超短期光伏功率预测方法

文中旨在提高晴空或有薄云这类小波动场景下的光伏功率超短期预测精度.虽然,太阳辐射的日周期性和年周期性使光伏功率序列具有确定性分量,但是电站和当地气象的详细参数随时间变化且难以获取.为此,提出一个仅依赖少量参数的改进晴空功率计算模型,并在此基础上构建了在线更新参数的预测算法,预测小波动天气下光伏电站未来4h的功率.采用中国吉林省某电站的数据进行了测试,结果表明:所提模型得到的晴空功率曲线可以较准确地拟合小波动天气下的电站出力,而基于在线更新参数的光伏预测结果可以使小波动天气第4个小时的预测误差降低到约3.78％,弥补了相邻晴天方法在一些场景下误差超过5％甚至达到10％的不足.文中所提方法不仅可以提高小波动天气下光伏功率超短期预测精度,也为复杂天气条件下的预测提供了更准确的基准值.

Ultra-short-term Photovoltaic Power Prediction Method Based on Modified Clear-sky Model

This paper aims to improve the accuracy of ultra-short-term photovoltaic power prediction in small fluctuation scenarios such as the clear sky or thin cloud. Although, the diurnal and annual periodicity of solar radiation makes the photovoltaic power series have deterministic components, the detailed parameters of power stations and local weather change with time and are difficult to obtain. To this end, this paper proposes a modified clear-sky power calculation model that only depends on a few parameters. On this basis, a prediction algorithm with online parameter updating is constructed to predict the power of a photovoltaic power station in the next four hours under the small fluctuation weather condition. The test is performed by using data from a power station in Jilin province. The results show that the clear-sky power curve obtained by the proposed model can accurately fit the power station output under the small fluctuation weather condition. The photovoltaic prediction results based on the online parameter updating can reduce the prediction error to about 3.78% in the fourth hour under the small fluctuation weather condition, which makes up for the deficiency that the error of adjacent sunny method is more than 5% or even 10% in some scenarios. The proposed method not only improves the accuracy of ultra-short-term prediction of photovoltaic power under the small fluctuation weather online, but also provides a more accurate benchmark for power prediction under complex weather conditions.