中国逐日地表太阳直射和散射辐射数据

该数据集可用于太阳能领域的光伏、光热发电潜力仿真模拟，尤其适用于双面光伏组件的发电潜力计算，还适用于其他光伏工程应用、水文建模和地表建模等。数据构建过程中的算法说明
数据构建过程的算法仅在数据集完成前使用，公布的数据集已经是完整版本。
模型训练：基于机器学习模型，使用中国大量地表时数观测衍生的直射和散射辐射数据增强训练数据，训练估算直射和散射辐射的机器学习模型。
数据估算：将基于卫星数据反演的总辐射数据和气象数据（温度、湿度、气压）输入训练好的模型，得到地表直射和散射太阳辐射的估算值（即数据结构中的“中国地表水平面散射太阳辐射分量”和“中国地表水平面直射太阳辐射分量”)，此处输入的总辐射和气象数据（温度、湿度和气压）仅在此步骤作为估算的输入数据，在估算值和最终的数据集中没有这些数据的信息，在数据处理时也无需这些数据作为中间数据。
数据集制作：将直射和散射辐射的估算值按照经纬度和时间的顺序进行排序，每一年的直射和散射存储为400*700*365（366）的数组，直射保存为dir_年份.mat文件，散射保存为dif_年份.mat文件。

This dataset can be used for the simulation of photovoltaic (PV) and concentrated solar power (CSP) generation potential in the solar energy sector, and is specifically applicable to the calculation of power generation potential for bifacial PV modules. It is also suitable for other photovoltaic engineering applications, hydrological modeling, land surface modeling and other relevant scenarios.

Algorithm Description for Dataset Construction Process
The algorithms employed during the dataset construction phase are only utilized prior to the finalization of the dataset, and the publicly released dataset is already a complete version.

Model Training: A machine learning model for estimating direct and diffuse solar radiation is trained by augmenting the training dataset with direct and diffuse radiation data derived from a large volume of surface hourly observational data across China.

Data Estimation: The trained model is supplied with satellite-retrieved total solar radiation data and meteorological data (temperature, humidity, atmospheric pressure) to generate estimated values of surface direct and diffuse solar radiation, which correspond to the "China Surface Horizontal Plane Diffuse Solar Radiation Component" and "China Surface Horizontal Plane Direct Solar Radiation Component" in the data structure. The total radiation and meteorological data used as inputs here only function as estimation inputs during this step; no information from these data is included in the estimated values or the final dataset, and they are not required as intermediate data during data processing.

Dataset Production: The estimated direct and diffuse solar radiation values are sorted according to the sequence of longitude, latitude and time. The direct and diffuse radiation data for each year are stored as 400×700×365 (366) arrays, with direct radiation saved as `dir_year.mat` files and diffuse radiation saved as `dif_year.mat` files.

搜集汇总

数据集介绍

特点

中国逐日地表太阳直射和散射辐射数据集包含大量记录，通过机器学习模型估算得出，适用于光伏工程、水文建模等领域。数据按经纬度和时间排序，每年数据存储为特定格式的文件。