基于WRF-SOLRA数值模式的太阳总辐射预报性能分析

Performance Analysis of Total Solar Radiation Forecasting by WRF-SOLAR Numerical Model

基于分布于内蒙古具有区域代表性的8个辐射站观测资料,分析不同区域、时间段和天气条件下WRF-SOLRA数值预报模式对太阳总辐射的预报性能,并对预报结果进行订正.结果表明:WRF-SOLRA数值模式的太阳总辐射预报与实况的趋势基本一致,且太阳总辐射的预报误差、趋势具有显著的时空分布特征.误差日分布由东至西呈现上午为正-中午为负-下午为正、全时段为负和上午为正-下午为负的特征;所有站午间误差大,预报偏移幅度小,而傍晚时段与之相反.同时,随着云量、气温和PM2.5、PM10浓度的增加,均方根误差均呈增大趋势,模式预报性能降低,而平均风速、平均气压和能见度增加,预报性能转好.通过检验分析,针对性的对太阳能总辐射数值预报进行了应用订正,订正后预报与观测的相位差由0～2h减为0～1h内、预报均方根误差由36.29～117.38W/m2减为36.19～91.00W/m^2,达到了服务光伏电场功率预测的目标.

Based on the solar radiation observational data of 8 stations with region representation in Inner Mongolia,the forecast performance of WRF-SOLAR numerical model in different area,different time periods and different weather conditions was analyzed,and the forecast results were corrected.Conclusions were as follows:the trends of forecast were basically consistent with the observation data,and the prediction errors and trends of the total solar radiation both had significant spatial and temporal distribution characteristics.The characteristics of the errors diurnal variations took on three types of features.From east to west,the values of errors were positive in morning,negative in noon and positive in afternoon,negative in all period of time and positive in the first half of day,negative in another half,respectively.The errors were large and deviations were small at noon,while errors and deviations showed the reverse change trend in the evening.In the meantime,with the increase of cloud,air temperature,PM2.5 and PM10 concentration,the model forecast performance decreased,while with the average wind speed,average air pressure and visibility increase,the forecast performance became better.The numerical prediction results of total solar radiation were revised,the phases between forecast and observation data were reduced from 0~2 hto 0~1 h,and the RMSEs were reduced from 36.29~117.38 W/m2 to 36.19~91.00 W/m2,reaching the goal of serving the region’s photovoltaic power prediction.