区域风电场群集中式功率预测系统设计与应用
摘要:风电功率预测系统大多遵循一个风电场配置一套预测系统的“一场一测”开发模式，此种模式下区域内相邻风电场的数据不能汇聚，预测模型无法充分考虑场站间的时空依赖关系来提升预测精度，大量独立的预测系统也无法得到统一、高效的维护。为此，提出了一种区域风电场群集中式功率预测系统。首先，该系统引入了多源数值天气预报数据作为模型输入；接着，在预测系统中进行了风机级和站级的数据清洗；最后，在预测系统中设计了风电场群时空联合预测模型。该模型可以捕捉区域内各风电场间的时空依赖性，能够同时为区域内所有风电场提供短期、超短期预测结果。所设计的系统应用于由10个风电场构成的风电场群的区域风电集控中心，在线运行结果表明，该预测系统符合区域电网评估要求，10个风电场在评估中排名靠前，运行期内预测考核扭亏为盈。
Abstract：Wind power forecasting system mostly follows the development mode of "one wind farm, one forecasting system" , in which one wind farm is equipped with one set of forecasting system. In this mode, the data of adjacent wind farms within the region cannot be aggregated, the prediction model cannot fully consider the spatio-temporal dependence between the stations to improve the prediction accuracy, and a large number of independent prediction systems cannot be maintained uniformly and efficiently. Therefore, a regional wind farm cluster Chinese power prediction system is proposed. Firstly, multi-source numerical weather forecast data are introduced as the input of the model. Then, the data of fan level and station level are cleaned in the prediction system. Finally, the spatio-temporal combined prediction model of wind farm group is designed in the prediction system. The model can capture the temporal and spatial dependence of wind farms in the region and provide short-term and ultra-short-term prediction results for all wind farms in the region simultaneously. The designed system is referenced in the regional wind power centralized control center of the wind farm group composed of 10 wind farms. The online operation results show that the prediction system meets the requirements of regional power grid evaluation. The 10 wind farms rank top in the evaluation, and the prediction assessment turns from loss to profit during the operation period.
标题：区域风电场群集中式功率预测系统设计与应用
title:Design and Application of Centralized Power Forecasting System for Regional Wind Farm Cluster
作者：阎 洁, 张永蕊, 张 浩
authors:YAN Jie, ZHANG Yongrui, ZHANG Hao
关键词：风电,集中式功率预测,时空预测,系统设计,
keywords:wind power,centralized power prediction,spatio-temporal prediction,system design,
发表日期：2022-04-08
- 文件大小:
- 7.92 MB
- 下载次数:
- 60
-
高速下载
|
|
