宜昌市电力负荷对气象因子的响应及气象预报模型的建立

利用2015-2017年宜昌市逐小时电力负荷资料及对应时段地面气象观测站数据,分析宜昌电力负荷的变化特征,研究气象敏感负荷与气象因子的关系,基于主要气象敏感因子通过逐步回归法建立宜昌电力负荷预报方法。结果表明:宜昌电力负荷呈逐年增长的趋势,夏季和冬季是一年中电力负荷高峰期,年最大电力负荷出现在夏季,年均增幅达11.8%,年最小电力负荷出现在春节期间;气温对气象敏感负荷影响最大,随着日平均气温T升高逐日气象负荷率先减小后增加,当T为17℃时,气象负荷率最小,从而划分了4个变化阶段:17℃≤T<26℃、T≥26℃、7℃≤T<17℃、T<7℃,基于各阶级主要气象敏感因子分别建立电力负荷回归预报方程,经检验,在实际应用中预报相对误差绝对值为3.8%,基本能够满足电力部门负荷预测的精度要求。后期可结合人工智能算法,进一步提高宜昌电力气象负荷预测的稳定性和准确性。

Response of Power Load to Meteorological Factors and its Weather Forecast Model in Yichang

Based on the hourly power load data and meteorological data of Yichang from 2015 to 2017, the characteristics of power load and the relationship between meteorological sensitive load and meteorological factors are analyzed. Stepwise regression method is adopted to get the prediction equations of load based on the main meteorological sensitive factors. The results show that power load is increasing year by year in Yichang, and that summer and winter are the two peak periods of power load in a year. The maximum power load occurs in the summer, with the annual growth rate reaching 11.8% on average, and the minimum power load occurs during the Spring Festival. Temperature is the major influence factor of meteorological sensitive load. With the increase of daily average temperature(T), the daily meteorological sensitive load rate decreases at first and then increases. When the daily average temperature is 17 ℃, the meteorological sensitive load rate is the lowest. Therefore, the change can be divided into four stages: 17 ℃≤T < 26 ℃, T≥26 ℃, 7 ℃≤T < 17 ℃, T< 7 ℃. Based on the main meteorological sensitive factors, each class of power load regression prediction equation is established respectively. The forecast test results show that the absolute relative error of forecast in practical application is 3.8%, which can basically meet the requirements of load forecasting in power sector. Combined with artificial intelligence algorithm, the stability and accuracy of meteorological sensitive load forecast of Yichang will be improved in the future.