基于数字孪生的区域气象关联风电预测模型

Regional Meteorology Correlated Wind Power Prediction Based on Digital Twin

摘要:

新型电力系统建设下，风电清洁能源得到大力开发。由于风电的发电功率与风速、风向、气温等气象因素强关联，具有波动性，且新建的风电场缺乏历史气象和发电数据，难以被精确预测。因此，文中提出了一种基于数字孪生的区域气象关联风电预测方法。首先，将风电场系统物理实体与数据驱动模型相结合，实现了数据同步和实时更新；然后，通过灰色关联分析方法筛选出对风电场发电功率影响作用较大的多元气象因素，使用XGB算法对优选后的气象关联数据集进行训练，结合天气预报对风电场发电功率进行预测；最后，利用有历史数据风电场的训练模型对无历史数据新建风电场进行预测。算例对四川某区域风电场数据进行了分析和预测，验证了所提方法的有效性与合理性，能够获得比传统预测方法更准确的预测结果。

Abstract:

With the construction of new power system, clean energy has a great development. Because the wind power has strong correlation with meteorological factors such as wind speed, wind direction and temperature, it has the strong volatility. And there lacks historical data of weather and power generation in the newly-built wind farms,so it is difficult to accurately predict the wind power. Therefore, a regional wind power prediction method correlated with meteorology based on digital twin is proposed. Firstly, the physical entity of wind farm system is combined with the datadriven model to realize data synchronization and realtime update. Then, the multivariate meteorological factors that have a great impact on generation power of wind farm are screened out by the gray correlation analysis method, and the optimized wind power meteorological data set is trained by extreme gradient boosting（XGB） algorithm and the future generation power of wind farm is predicted in combination with the weather forecast. Finally, the training model of wind farms with historical data is used to predict the newly-built wind farms without historical data. Cases are given to analyze and predict the wind farm data of a certain region in Sichuan province, which verifies the effectiveness and rationality of the proposed method and can obtain more accurate prediction results than the traditional prediction methods.