沙尘天气对太行山东麓地面太阳辐射影响及预报订正研究

Impact of Sand-dust Weather on Surface Solar Radiation at the Eastern Foothill of Taihang Mountains and Forecast Revision

摘要:

利用太行山东麓涞源、曲阳、平山、涉县4个光伏电站2021年1月—2023年12月太阳辐照度实况和辐照度、总云量预报数据，逐小时总云量综合实况判别数据和环境监测站PM₁₀小时平均浓度实测数据，分析沙尘天气过程期间太阳辐照度时间和频次特征，基于辐射比与PM₁₀小时平均浓度的晴天订正模型，进行订正效果检验。结果表明：（1）2021—2022年沙尘天气过程较多，单次沙尘天气过程最长影响时间超过63 h，当PM₁₀小时浓度均超过1 588 μg/m³时，太阳辐照度降低到晴天的26.3%；（2）太阳总辐射辐照度日、月变化特征明显，以12—13时BT为中心呈正态分布，以6月为峰值中心呈单峰形分布，3、7、11月受沙尘、降水和大雾等影响，太阳总辐射辐照度总体偏低；（3）沙尘天气辐照度模型对太阳辐照度预报订正有正效果，MAE、RMSE分别降低34.3、31.0 W/m²。订正效果与PM₁₀小时平均浓度预报质量密切相关，PM₁₀预报效果较好时，订正效果较明显，否则订正效果较差。

Abstract:

Utilizing solar irradiation and irradiance, total cloud forecast data, hourly total cloud comprehensive data, and environmental monitoring station PM₁₀ hourly average concentration measured data from four photovoltaic power stations in Laiyuan, Quyang, Pingshan and Shexian at the eastern foothill of Taihang Mountains from January 2021 to December 2023,this paper analyzed the time and frequency characteristics of solar radiation intensity, studied the clear sky correction model based on radiation ratio and hourly average concentration of PM₁₀, and tested its correction effect. The results indicate that: (1) There were numerous sand-dust weather processes from 2021 to 2022 with a maximum impact time exceeding 63 hours. When PM₁₀ hour concentrations exceeded 1 588 μg/m³, solar irradiance decreased to 26.3% of sunny days. (2) Daily and monthly variations of total solar radiation were evident with a normal distribution centered around 12-13 o'clock BT and a single peak distribution centered around June. In March, July, and November, the solar radiation intensity was lower than the overall trends due to sand-dust, precipitation, and heavy fog. (3) The sand-dust weather irradiance model positively impacted the overall revision of solar irradiance forecast with MAE decreasing by 34.3 W/m² and RMSE decreasing by 31.0 W/m². The correction effect is closely related to the forecast quality of PM₁₀ hour average concentration. When the PM₁₀ forecast effect is good, the correction effect is obvious; otherwise it is poor.