Typical characteristic mining and load evaluation for the difficulties of power supply–demand balance regulation

• Published in: IET generation, transmission & distribution transmission & distribution, 2024-12, Vol.18 (23), p.3945-3958
• Main Authors: Li, Junwei, Yu, Yang, Mi, Zengqiang, Wu, Jian
• Format: Article
• Language: English
• Subjects: data mining; demand side management; load management; load regulation; power system planning; supply and demand
• ISSN: 1751-8687; 1751-8695
• DOI: 10.1049/gtd2.13323

Power systems face significant challenges in maintaining power balance because of the stochasticity of sources and loads. This unpredictability makes it difficult to characterize the typical demand for grid power balance regulation, which results in a lack of clear objectives for evaluating the contributions of demand‐side users to power balance. To bridge these gaps, in this paper, a new approach is proposed on the basis of the inherent difficulties in grid power balance regulation. First, a method for portraying the change in power balance regulation demand is presented, which emphasizes the trend characteristics of time‐varying regulation demand through the weighting of trend segments. Second, demand periods and curves are calculated on the basis of a regulation capacity of 5% of the maximum load, and their distribution during the monthly cycle is analysed. Finally, to elucidate the contribution of random samples to typical demand features, a sample weighting method utilizing clustering categories is proposed and a distance‐minimum optimization model is constructed to estimate typical features by leveraging the “group effect” concept. Actual power data from a region in China are selected for verification, confirming that the proposed typical value calculation method is more representative of random sampling situations. Moreover, considering the power grid's time‐varying nature, the evaluation of the contribution of demand‐side users to the power grid is improved. This paper proposes a load curve matching method that assigns weights to trend features. It emphasis the fluctuating demand trends in power balance regulation during the demand time period. This paper illustrates the complex distribution of monthly demand characteristics within the power system. It highlights the challenge of mining typical demand characteristics. To elucidate the contribution of stochastic demand feature samples to typical values, this paper proposes an integrative distance minimization approach for estimating typical values based on clustering‐category weighting, which is underpinned by the “group effect” principle. This paper evaluates the contribution of demand‐side resources in alleviating the supply‐demand imbalance in the power grid during the monthly cycle.