Techno-Economic Comparison of Centralized and Distributed Photovoltaic Configurations for a University Building Cluster

This study presents a techno-economic comparison of centralized and distributed photovoltaic configurations for a representative university building cluster under a common battery-storage framework. Direct comparisons of such configurations under identical load and meteorological conditions remain limited in the literature. The Electrical and Electronics Faculty building was selected as the reference building due to its relatively high electricity consumption, and its demand data were scaled to a three-building-equivalent level in order to compare the two configurations at a similar system size. Based on the monthly consumption data of the reference building, an 8760 -hour synthetic load profile was generated in MATLAB, and HOMER Pro was used for the comparative simulations. Economic and technical performance was assessed using indicators such as capital expenditure (CAPEX), operating expenditure (OPEX), net present cost (NPC), levelized cost of energy (LCOE), internal rate of return (IRR), payback period, self-consumption, and self-sufficiency. The centralized configuration yielded better economic performance, with an LCOE of 2.49 TRY/kWh compared to 2.64 TRY/kWh for the distributed case, and a shorter payback period of 6.1 years versus 6.3 years. By contrast, the annual grid-interaction indicators of the two configurations remained close, as indicated by the selfconsumption and self-sufficiency results. Tariff sensitivity analysis also showed that the centralized configuration remained economically advantageous across the evaluated tariff scenarios. For the representative campus building cluster considered in this study, the results suggest that the distinction between the two configurations appears more clearly in economic indicators than in annual grid-interaction metrics. The study may serve as a comparative reference for campusscale PV planning under a common demand and meteorological framework.