South Africa faces high energy costs, highlighting the urgent need for sustainable and cost-effective energy solutions. This study investigates the design of a cost-effective photovoltaic energy system that maximises savings and revenue for the school through energy trading. In this study, the school trades with 14 neighbouring households and 125 electric motorbikes. This research first applies Latin Hypercube Sampling to explore the solution space and determine which system parameters have a significant impact on supply reliability, investment costs, revenue and savings. Optimal solutions are generated using Non-Dominated Sorting Genetic Algorithm II for a range of system scenarios. Following this, the most promising scenario is selected and applied to 53 schools in the Western Cape. The results show that number of panels strongly correlates with both supply reliability and revenue, thus reducing the break-even years, while battery capacity affects investment costs and, to some extent, break-even years. Among the configurations tested, scenarios where schools traded with both households and electric motorbikes, particularly when both included their own battery systems, achieved the most favourable financial performance for the school, with break-even periods of less than five years under sufficient roof area and improved reliability for the external entities, with an average improvement of 60%. These findings demonstrate that peer-to-peer energy trading between schools and communities can enhance the financial feasibility and sustainability of decentralised solar systems, offering a scalable model for improving energy access and affordability in South Africa and possibly other developing countries.
Molise et al. (Sun,) studied this question.