基于机器学习与数值预报技术的电网短期临近气象预警模型设计

Design of short⁃term near⁃meteorological early warning model for power grid based on machine learning and numerical forecasting technology

电网短期临近气象预警是预防电网故障发生的重要手段之一,故对其进行及时、准确地预警可有效减小因灾害性天气带来的损失。针对电网短期临近预警问题,文中开展了基于机器学习与数值预报技术的电网短期临近气象预警模型设计的研究。该研究对收集到的风速、温度与湿度等数据进行预处理与三性分析,并将得到的数据转换为经纬度网格数据。同时根据气象模型设计预警流程,采用VC++6.0平台完成对雷达数据的处理,且通过小波融合算法提高了红外数据的准确度。利用交叉相关法进行时间序列分析以提高数据识别的准确率,进而解决了短时预警问题。实际算例结果表明,所提出的方法能够有效预警极端天气,且预测精度可达96.88%。

Short-term approaching meteorological early warning of power grid is one of the important means to prevent power grid faults.Timely and accurate early warning can effectively reduce the losses caused by disastrous weather.Aiming at the problem of short-term proximity early warning of power grid,this paper carries out the design and research of short-term proximity meteorological early warning model of power grid based on machine learning and numerical prediction technology,preprocesses and analyzes the collected data such as wind speed,temperature and humidity,and converts the obtained data into longitude and latitude grid data.The early warning process is designed according to the meteorological model,the radar data processing is completed by using VC++6.0 platform,and the accuracy of infrared data is improved by wavelet fusion algorithm.The cross-correlation method is used for time series analysis to improve the accuracy of data recognition,so as to solve the problem of short-term early warning.The results of practical examples show that the method proposed in this paper can effectively warn extreme weather,and the prediction accuracy is 96.88%.