Consumer Segmentation and Pricing Strategy for Indian Low-Voltage Distribution Networks Using Machine Learning

The primary objective of this study is to develop a fair and efficient electricity pricing mechanism using a clustering-based approach, eliminating the conventional classification of consumers into domestic, commercial, and industrial categories. This topic is essential in the context of India’s growing focus on reducing distribution losses and improving energy equity, particularly under the Smart Grid Mission initiated by the government. The existing tariff structure, based on consumer type and time-of-use pricing, fails to account for actual consumption behaviour, resulting in inefficiencies and potential unfairness in billing. This paper addresses the limitations of conventional pricing strategies are addressed by proposing a data-driven method, which groups consumers according to their actual load patterns. The novelty of this study lies in its objective pricing policy, which allows the distribution companies to impose tariffs based on actual consumption patterns and not on pre-determined categories. This approach entails the use of unsupervised machine learning algorithms to group consumer load profiles which are obtained in a real-life Indian distribution feeder. This study incorporates sample time-series load data that was gathered within a specific time frame over different consumer behaviors and synthesized for this study. K Means clustering algorithm is used to determine the different load patterns, and then intra-cluster similarity-based tariff design was done. Each cluster has a fixed base price based on the co-varaince. The proposed cluster-based tariff design shows a significant reduction in revenue (about 50%), highlighting the need for future strategies such as adaptive cluster pricing to balance utility sustainability with consumer affordability.