针对广域分布式新能源普遍缺乏新能源资源监测装置,而导致功率预测精度不足的问题,提出一种基于气象资源插值与迁移学习的广域分布式光伏功率预测方法。首先,基于地理信息和粗颗粒气象数据,对广域范围下的气象资源数据进行网格化插值;其...针对广域分布式新能源普遍缺乏新能源资源监测装置,而导致功率预测精度不足的问题,提出一种基于气象资源插值与迁移学习的广域分布式光伏功率预测方法。首先,基于地理信息和粗颗粒气象数据,对广域范围下的气象资源数据进行网格化插值;其次,依据插值结果对具有相同气象特征的光伏电站进行自组织映射(self-organizing maps,SOM)网络聚类,并对每一类中的光伏电站进行迁移学习的源域和目标域的划分,以保证预测精度;然后,结合长短期记忆(long short term memory,LSTM)网络,引入误差修正环节,建立源域至目标域的双迁移模型;最后,以浙江省绍兴市的分布式光伏电站为实例验证该方法的有效性。相比于对各个光伏电站单独建模,所提方法能将目标域光伏电站的训练速度提高10倍以上,且在预测精度方面也有显著提升,具有一定的推广应用价值。展开更多
Soil types, humus types and vegetation as well as their hypsometric variation were analysed in terms of sequences in the northern part of the high mountains of the Pirin National Park at altitudes between looo and 240...Soil types, humus types and vegetation as well as their hypsometric variation were analysed in terms of sequences in the northern part of the high mountains of the Pirin National Park at altitudes between looo and 2400 m a.s.1. The study area is characterised by a large variety of natural parameters like petrology （mainly marble and granite）, morphology （different slope deposits, exposition） and the orographic climate gradient. Statistical analyses using these parameters provided a basis for the soil group classification of the sites. Based on a Digital Terrain Model （DTM） and a geological map of the Pirin National Park, the results of these statistical analyses were used to generate a ＂map of potential soil groups＂ （regionalisation using GIS）. Six potential soil groups could be determined. The resulting map exhibits a confidence level of 68 % on 74.4 % of the covered area. Rendzic Leptosols, in combination with Folic Histosols and Histi-lithic Leptosols occur in the alpine and subalpine regions on calcareous substrates. With decreasing altitude they are replaced by a mosaic of Rendzic Leptosols, Phaeozems and an increasing occurrence of Cambic Umbrisols. Umbrisols found on silicatic substrates in the alpine region are replaced by Cambic Umbrisols with decreasing altitude as well. Hence, pedogenesis is characterised by increasing browning and depth of the soil profiles with decreasing altitude. The pH-level is slightly acidic to neutral in lower zones and on calcareous rocky bases. Acidification increases in the subalpine zone. Soil pH decreases down to 4 on silicate subtrates. Typical humic values in mineral topsoils are 10 to 12 %, and in organic layers of the soils above 2000 m a.s.1, they are even more. The C：N ratio closely ranges around 20 （median）.展开更多
文摘针对广域分布式新能源普遍缺乏新能源资源监测装置,而导致功率预测精度不足的问题,提出一种基于气象资源插值与迁移学习的广域分布式光伏功率预测方法。首先,基于地理信息和粗颗粒气象数据,对广域范围下的气象资源数据进行网格化插值;其次,依据插值结果对具有相同气象特征的光伏电站进行自组织映射(self-organizing maps,SOM)网络聚类,并对每一类中的光伏电站进行迁移学习的源域和目标域的划分,以保证预测精度;然后,结合长短期记忆(long short term memory,LSTM)网络,引入误差修正环节,建立源域至目标域的双迁移模型;最后,以浙江省绍兴市的分布式光伏电站为实例验证该方法的有效性。相比于对各个光伏电站单独建模,所提方法能将目标域光伏电站的训练速度提高10倍以上,且在预测精度方面也有显著提升,具有一定的推广应用价值。
文摘Soil types, humus types and vegetation as well as their hypsometric variation were analysed in terms of sequences in the northern part of the high mountains of the Pirin National Park at altitudes between looo and 2400 m a.s.1. The study area is characterised by a large variety of natural parameters like petrology （mainly marble and granite）, morphology （different slope deposits, exposition） and the orographic climate gradient. Statistical analyses using these parameters provided a basis for the soil group classification of the sites. Based on a Digital Terrain Model （DTM） and a geological map of the Pirin National Park, the results of these statistical analyses were used to generate a ＂map of potential soil groups＂ （regionalisation using GIS）. Six potential soil groups could be determined. The resulting map exhibits a confidence level of 68 % on 74.4 % of the covered area. Rendzic Leptosols, in combination with Folic Histosols and Histi-lithic Leptosols occur in the alpine and subalpine regions on calcareous substrates. With decreasing altitude they are replaced by a mosaic of Rendzic Leptosols, Phaeozems and an increasing occurrence of Cambic Umbrisols. Umbrisols found on silicatic substrates in the alpine region are replaced by Cambic Umbrisols with decreasing altitude as well. Hence, pedogenesis is characterised by increasing browning and depth of the soil profiles with decreasing altitude. The pH-level is slightly acidic to neutral in lower zones and on calcareous rocky bases. Acidification increases in the subalpine zone. Soil pH decreases down to 4 on silicate subtrates. Typical humic values in mineral topsoils are 10 to 12 %, and in organic layers of the soils above 2000 m a.s.1, they are even more. The C：N ratio closely ranges around 20 （median）.
