A method for electricity decomposition integrating ant colony optimization and time series proportions
ZHOU Haowei
CUI Yuxuan
ZHOU Ying
BAI Xuefeng
TIAN Chuanbo
ZHAO Weibo
Abstract：The efficient operation of power systems relies on precise electricity decomposition techniques.To address complex electricity demands and unforeseen events,this study proposes an electricity decomposition method integrating ant colony algorithms with time series ratios.First,the Balanced Iterative Reduction Clustering(BIRCH)method is applied to cluster electricity consumption data from a textile industry user in Fuzhou,Fujian Province,extracting typical electricity consumption curves.Subsequently,wavelet decomposition and artificial ant colony optimisation are employed to refine thresholds,thereby eliminating noise components and enhancing both signal-to-noise ratio and data accuracy.Time series decomposition(STL)is then applied to decompose the electricity data,isolating trend,fundamental,and residual components.The Shapley additive explanation(SHAP)method is subsequently utilised to quantify each component's contribution to electricity fluctuations.To enhance decomposition accuracy,a quantile mapping method is introduced to calibrate the results,yielding a significant improvement in precision.Experimental results demonstrate that post-correction decomposition errors are reduced,withMSE,MAE,RMSE,and overall error decreasing to 0.028%,0.016%,0.027%,and 2.3%respectively.This methodology effectively enhances electricity decomposition accuracy,providing critical support for refined management and optimised scheduling within power systems.
Keywords：time series proportional decomposition methodant colony optimisationwavelet transformquantile mapping methodartificial bee colony algorithm
Publication Date：2025-11-28
Online Publishing Date：2025-12-18(First online date of this platform, not the publication date of the document)
Pages：11( 32-42 )