利用石家庄地区2007~2019年最热月气温、空气湿度、风速、大风、降水与社会用电量数据，采用线性趋势、相关系数、多元回归和对比分析等统计方法，分析了最热月相关气象要素与社会用电量的变化关系。结果表明：近年来石家庄夏季社会用电量呈明显增长趋势，最热月增长趋势最为明显；以市区及其周边地区为高用电量中心，向周边各县市逐渐减少，东南部较西北部偏高。全市和各区域最热月平均气温、最高气温、最低气温与气象用电量呈明显正相关，正相关性市区较各区域更为明显；东部平原最热月平均风速与气象用电量呈明显正相关。利用SPSS软件对全市和各区域逐年最热月相关气象要素数据与气象用电量进行多元逐步线性回归分析，得到气象用电量的回归预测方程。通过全市和各区域逐年最热月预测社会用电量与实际社会用电量的对比误差分析，表明预测社会用电量平均误差在3%以内。
By using the data of air temperature, air humidity, wind speed, high wind, precipitation and social electricity consumption of the hottest month from 2007 to 2019 in Shijiazhuang, and using statistical methods such as linear trend, correlation coefficient, multiple regression and comparative analysis, the relationship between meteorological factors related to the hottest month and social electricity consumption was analyzed. The results show that, in recent years, the social electricity consumption in summer in Shijiazhuang showed an obvious growth trend having the most obvious growth trend in the hottest month. With the urban area and its surrounding areas as the center of high electricity consumption, it gradually decreased to surrounding counties and cities, and the southeast was higher than the northwest. The average temperature, maximum temperature and minimum temperature of the hottest month in the whole city and all regions were positively correlated with meteorological electricity consumption, and the positive correlation was more obvious in urban areas than in other regions. The average wind speed of the hottest month in the eastern plain was positively correlated with the meteorological electricity consumption. SPSS software was used to carry out multivariate stepwise linear regression analysis of meteorological elements data and meteorological electricity consumption in the hottest month of the whole city and each region year by year, and the regression prediction equation of meteorological electricity consumption was obtained. Through the comparison error analysis of predicting social electricity consumption and actual social electricity consumption in the hottest month of each year in the whole city and each region, it showed that the average error of predicting social electricity consumption was within 3%.
