电力负荷与气象因子因果关系分析及预测研究

Analysis and Prediction of Causal Relationship between Power Load and Meteorological Factors

短期电力负荷预测是电力部门生产调度的重要参考,不同地区影响电力负荷的因素有所不同,因此,为探索电力负荷与气象因子之间的因果关系,以某变电站逐小时电力负荷和气象要素数据为基础,采用格兰杰因果检验分析气象要素与电力负荷的因果关系。采用ADF检验验证气象因子与电力负荷数据的平稳性,将通过ADF检验的平稳性变量进行格兰杰因果检验。结果显示,温度、相对湿度和风速均是电力负荷的格兰杰原因,温度和相对湿度对电力负荷变化的影响是实时的,风速对电力负荷变化的影响具有1 h以上的滞后性。采用灰色关联度和余弦相似度建立综合相似性指标,基于相似日法提出一种短期电力负荷预测模型,以2018—2020年的气象数据和电力负荷数据作为样本库,采用2021年的数据对模型预测准确性进行检验。经验证,模型在天气因子变化不明显或变化缓慢情况下预测准确率为90%以上,可作为电力部门生产调度参考。

Short term power load forecasting is an important reference for production scheduling in the power sector.The factors that affect power load vary in different regions.Therefore,in order to explore the causal relationship between power load and meteorological factors,based on hourly power load and meteorological factor data of a certain substation,Granger causality test is used to analyze the causal relationship between meteorological factors and power load.Using ADF test to verify the stationarity of meteorological factors and electricity load data,Granger causality test is conducted on the stationarity variables that pass ADF test.The results show that temperature,relative humidity,and wind speed are all Granger causes of power load.The impact of temperature and relative humidity on power load changes is real-time,and the impact of wind speed on power load changes has a lag of more than 1 h.A comprehensive similarity index is established using grey correlation and cosine similarity,and a short-term electricity load prediction model is proposed based on the similarity day method.The meteorological data and electricity load data from 2018 to 2020 is used as the sample library,and the prediction accuracy of the model is tested by the data from 2021.Verification results show that,the model has a prediction accuracy of over 90%when weather factors do not change significantly or change slowly,and can be used as a reference for production scheduling in the power sector.