A Time-of-Use pricing time division method based on DTW-C-FCM clustering and an improved quantile approach

•DTW–Canopy–FCM clustering enhances typical daily net load curve extraction.•Improved quantile method captures both load magnitude and fluctuation dynamics.•Bayesian framework evaluates TOU period segmentation rationality and validity.•Method reduces peak–valley disparity and boosts demand response...

Full description

Saved in:
Bibliographic Details
Published in:International journal of electrical power & energy systems Vol. 172; p. 111279
Main Authors: Wang, Liying, Dong, Houqi, Wang, Peng, Shi, Mengshu, Wang, Jiani
Format: Journal Article
Language:English
Published: Elsevier Ltd 01.11.2025
Elsevier
Subjects:
Bayesian evaluation
Dynamic time warping-Canopy–Fuzzy clustering means algorithm
Net load
Quantile-based segmentation
Time division
Quantile-based segmentation
Net load
Bayesian evaluation
Dynamic time warping-Canopy–Fuzzy clustering means algorithm
Time division
ISSN:0142-0615
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:•DTW–Canopy–FCM clustering enhances typical daily net load curve extraction.•Improved quantile method captures both load magnitude and fluctuation dynamics.•Bayesian framework evaluates TOU period segmentation rationality and validity.•Method reduces peak–valley disparity and boosts demand response potential. Accurate time division is essential for implementing peak–valley time-of-use electricity pricing. With the increasing penetration of renewable energy in new-type power systems, the peak–valley disparity has intensified, and traditional fixed time division methods—based solely on user load data—fail to address fluctuations caused by the randomness and uncertainty of both generation and demand. To tackle this problem, this study considers net load fluctuations from the supply and demand sides and proposes an improved fuzzy C-means (FCM) clustering method that integrates dynamic time warping (DTW) with a density-based Canopy mechanism to enhance the accuracy and robustness of typical daily load curve extraction. Load fluctuation rate and load magnitude are then used as key indicators to develop a time division model based on the improved quantile approach. Furthermore, a Bayesian evaluation method is employed to assess the rationality of the resulting time divisions. Simulation results indicate that the proposed method provides finer time granularity and effectively mitigates peak–valley disparity compared with conventional schemes. These findings demonstrate the method’s potential to support more adaptive and precise time-of-use pricing strategies in modern power systems.
ISSN:0142-0615
DOI:10.1016/j.ijepes.2025.111279