基于频域分量法的多模型日最大负荷预测

Multi Model Daily Maximum Load Forecasting Based on Frequency Domain Component Method

日最大负荷是表征日用电负荷的重要特征之一,对日最大负荷的预测可为合理安排日发电计划提供重要的支持。为提高预测的准确性,文中提出用最大相关最小冗余特征选择算法(mRMR特征选择)分析负荷与其影响因素之间的关系,筛选出影响负荷变化的主要因素;并运用小波分析的原理对日最大负荷进行频域分解,得出低频分量和高频分量,针对不同分量的特点构建相应的预测模型进行预测,最后把各分量的预测结果进行重构作为最终的预测结果,实验证明,与其他预测方法相比,文中所用方法能取得较好的预测结果。

The maximum load is one of the important characteristics of the daily electricity load, and the forecast of the daily maximum load can provide important support for the reasonable arrangement of the daily power generation plan. In order to improve the accuracy of the forecast, the relationship between the load and its influencing factors is analyzed by using max-relevance&min redundancy algorithm(mRMR feature selection), and the main factors that affect the load change are screened out;and the wavelet analysis is used to decompose the maximum load in the frequency domain, and the low frequency component and the high frequency component are obtained, the forecasting model is constructed according to the characteristics of different components. Finally, the forecasting results of each component are reconstructed as the final forecast result, experiments show that compared with other forecasting methods, the method used in this paper can achieve better forecasting results.