基于提升小波的时间序列分析法的电力负荷预测  

Power Load Forecasting in the Time Series Analysis Method Based on Lifting Wavelet

作　　者：张帆[1] 张峰[1] 张士文[1] 

机构地区：[1]上海交通大学电子信息与电气工程学院,上海200240

出　　处：《电气自动化》2017年第3期72-76,共5页Electrical Automation

摘　　要：为了合理分配电能和规划电网运行,需要对居民小区电力负荷进行精确预测。提出了一种基于提升小波的时间序列分析法进行电力负荷预测,采用提升小波对居民小区电力负荷进行主要特征量提取,避免了用电量数据随机性和波动性的干扰;运用时间序列法对经过提升小波去噪后的电力负荷序列,求取自相关和偏相关系数,确立相应数学模型,预测未来时刻的用电量。最后,利用上海嘉定某小区的历史负荷大数据进行了比较和分析,验证了方法的有效性。For the purpose of rational distribution of electric energy and planning of power grid operation, it is necessary to make precise forecasting on power load of residential quarters. Here, we propose a time series analysis method based on lifting wavelet to make power load forecasting. Lifting wavelet is used to extract main characteristic quantities of power load in residential quarters, thus avoiding interference from randomness and volatility which occur if electricity data is used. The time series method is used to calculate autocorrelation and partial correlation coefficient of the power load series which has been denoised through lifting wavelet, establish corresponding mathematical models and forecast future electricity consumption. Finally, comparison and analysis is made by use of the big data of past toad of a certain residential quarter in Jiading District, Shanghai, and the effectiveness of the proposed method is verified.

关 键 词：电力负荷预测 提升小波 时间序列法 数学模型 大数据 

分 类 号：TM714[电气工程—电力系统及自动化]