基于不确定理论的光伏出力预测研究

Research on the Prediction of PV Output Based on Uncertainty Theory

预测光伏出力对光伏发电系统的并网运行具有重要意义。目前很多预测方法中只能得到单一预测结果,缺乏对影响因素不确定性的分析。考虑到太阳辐射和温度是光伏出力的两个最大的影响因素,本文建立了基于不确定理论的太阳辐射值预测模型。首先分析了云量的模糊性和云遮系数的双重随机性,采用双重随机理论计算云量对应云遮系数的期望值和关键值,然后以无云天气计算太阳辐射值的REST模型为基础,用云遮系数对其进行修正,得到太阳辐射值的预测结果,最后得到光伏出力在不同云量下的预测期望值和满足不同置信水平的预测区间。模型选取美国BMS光伏电站的数据,验证了改进后模型的有效性。预测结果表明,该模型能够提供较丰富的信息,具有较好的实用性。

The forecast of photovoltaic output is important to the grid-connection of the photovoltaic power generation system. Most of the current method of forecasting PV output can only obtain a single prediction results, lacking of the analysis of uncertainty factors. Taking into account that solar radiation and temperature are the two most important factors, a model is presented based on the uncertainty theory in the paper. After analyzing the fuzziness of the cloud cover and the birandomness of cloud cover index, the expectations and critical values of the cloud cover index are calculated corresponding to forecasted cloud cover by using the birandom theory. By correcting the radiation values got from the cloudless weather REST model with the cloud cover index, the forecast solar radiation can be got, and ultimately the PV output under different cloud cover andthe prediction interval under different confidence levels can be got. The model selects data from BMS PV plant in the U.S. to verify its validity. Prediction results show that the model can provide richer information andhas good practicability.