Investigating distribution systems impacts with clustered technology penetration and customer load patterns

Abstract

Electric vehicles (EVs), photovoltaics, heat pumps and energy storage are changing the demands placed on electricity systems and can pose significant challenges for system operators and distribution companies. Furthermore, clustering of behaviours and technologies throughout different areas of distribution systems can produce broad variation in load curves and impacts on the network. This paper investigates local clustering impacts in a utility service area as a case study to develop methods and gain insights which can be applied to other datasets. Through clustering the variation in technology penetration rates across distribution transformers is revealed, a level of granular variability which has not been well-quantified in past literature. A second clustering framework is the applied to transformer load profiles to identify a small but diversely representative set of novel archetypical local loads. These profiles provide a summary of the dataset variability, showing how simple modeling can begin to illustrate the impacts of future technology penetration across different regions of the system. The results of the case study demonstrate that home EV charging will significantly increase peak residential transformer loading (up to 19% with 25% EV penetration), potentially drastically decreasing their useful life. Results also produced insights into possible mitigation strategies. By taking advantage of alternate charging opportunities (like workplace) the load can be spread across transformers, reducing growth in local residential and aggregate peaks by 2–8%. Energy storage is found to be more effective on residential transformers than business ones, promoting deferral of capacity investment, while simultaneously matching local and regional grid requirements for demand smoothing. In contrast, photovoltaics are found most effective at lowering new and baseline peak demands when on commercial and industrial transformers, particularly for small businesses where moderate penetration scenarios for EVs and PVs showed peak demand actually declining by 1–9%. The data analysis and clustering techniques developed through this case study can provide valuable insight into large datasets for policy development and potentially revelatory illustration of the varying effects of new technology within evolving networks.