基于改进反距离加权的山东风资源高分辨率网格化数据集生成方法  

A High-resolution Gridded Dataset Generation Method for Wind Resources in Shandong Province Based on Improved Inverse Distance Weighting

作　　者：王姝 靳双龙 王勃 杨虎 刘晓琳 刘畅 王红庆[2] WANG Shu;JIN Shuanglong;WANG Bo;YANG Hu;LIU Xiaolin;LIU Chang;WANG Hongqing(State Key Laboratory of Renewable Energy Grid-integration(China Electric Power Research Institute),Haidian District,Beijing 100192,China;College of Science,Beijing Forestry University,Haidian District,Beijing 100083,China)

机构地区：[1]可再生能源并网全国重点实验室(中国电力科学研究院有限公司),北京市海淀区100192 [2]北京林业大学理学院,北京市海淀区100083

出　　处：《电网技术》2025年第3期1166-1175,共10页Power System Technology

基　　金：国家电网有限公司科技项目(5100-202255337A-2-0-SY)。

摘　　要：风力发电在我国能源结构中占比逐年攀升。对风能资源进行准确全面的评估是提升风电出力水平和消纳能力的先决条件。基于空间插值方法建立的高分辨率网格化风资源数据集,可对风资源进行大范围、格点化和精细化的有效评估。为提高风资源数据集的准确性,文章提出了一种基于K-means++自适应的改进反距离加权插值方法(K-means++adaptive inverse distance weighted interpolation method,K-means++AIDW)。使用该方法对山东地区2022年全年109个国家级气象观测站点的风速实测数据进行处理,构建空间分辨率为9km×9km的网格点,使用风速实测数据逐小时对网格点进行插值填补,得到高分辨率的网格化风资源数据集。将插值后的结果与原始观测数据进行比较发现,与传统反距离加权法(inverse distance weighting,IDW)和Kriging插值方法相比,所设计的K-means++AIDW插值方法平均绝对误差较IDW方法降低了5.4%,较Kriging方法降低了7.8%;均方根误差较IDW方法降低了5.9%,较Kriging方法降低了8.1%,显示出其在整体误差控制上的优势。与空间分辨率0.25°×0.25°的再分析回算数据集ERA5(Fifth Generation of European Centre for Medium-range Weather Forecasts Atmospheric Reanalysis of the Global Climate)的风资源要素相比,所设计的K-means++AIDW插值数据集平均绝对误差和均方根误差平均降低了11.95%和10.07%,验证了所设计插值方法的准确有效性,以及生成的高分辨率网格化数据集的精准可靠性,可作为评估山东省的风能资源潜力的可靠数据基础,为风能资源管理和风电场选址等领域提供准确的数据支持。The proportion of wind power generation in China's energy mix has increased yearly.Accurate and comprehensive assessment of wind resources is a prerequisite for improving wind power output and consumption capacity.A high-resolution gridded wind resource dataset based on the spatial interpolation method can effectively provide an assessment of wind resources in a wide range,gridded,and refined way.To improve the accuracy of the wind resource dataset,this paper proposes an improved K-means++Adaptive Inverse Distance Weighted interpolation method(K-means++AIDW).Adopting this method,the wind measurement data derived from 109 nation-level meteorological observation stations in Shandong province across the whole year of 2022 are processed and interpolated on every constructed grid point in the constructed grid with a spatial resolution of 9km×9km,hour by hour,to obtain a high-resolution gridded wind resource dataset.According to statistical comparison between interpolation and original observation,we found that the K-Means++AIDW interpolation method proposed in this paper outputs an MAE decreased by 5.4%compared with that of the traditional Inverse Distance Weighting(IDW),decreased by 7.8%compared with Kriging interpolation;as well as an RMSE decreased by 5.9%compared with that of the traditional IDW,decreased by 8.1%compared with Kriging interpolation,which shows its advantage in overall error control.Compared with ERA5(Fifth Generation of European Centre for Medium-range Weather Forecasts Atmospheric Reanalysis of the Global Climate)with spatial resolution of 0.25°×0.25°,The K-Means++AIDW interpolation dataset’s MAE and RMSE decreased by 11.95%and 10.07%respectively on average,which verifies the accuracy and effectiveness of the interpolation method designed in this paper and the accuracy and reliability of the generated high-resolution gridded dataset.It can be used as a reliable data basis for evaluating the wind energy resource potential of Shandong Province and provide accurate data support for wind farm si

关 键 词：高分辨率风资源网格化数据集 风能资源评估 K-means++自适应 改进反距离加权插值 

分 类 号：TM614[电气工程—电力系统及自动化]