Energy scheduling strategy for a photovoltaic/batteries bank/diesel generator power supply system for an off-grid house

A novel algorithm is proposed to control the overall energy produced by a photovoltaic/battery bank/diesel generator renewable energy system. A renewable energy system is used to supply an off-grid connected house in Sfax, Tunisia. The algorithm computes recurrently the battery depth of discharge and diesel consumption of the diesel generator and then forecasts the photovoltaic subsystem output power every two minutes ahead. The estimated power is computed using an auto-regressive moving average model associated with a Kalman filter. During each calculation loop, hybrid system energy scheduling is accomplished considering criteria that guarantee the maximum use of generated renewable energy, minimum diesel generator operational time, and full-load need satisfaction during the whole day. Numerical simulations for two typical sunny and cloudy days in Sfax, Tunisia, are conducted in order to validate the algorithm and to compare the performance of managed system behavior with that of a standard unmanaged system. It is shown that the proposed energy management system could save fuel consumption by as much as 11.27% in the sunny day and 9.17% in the cloudy day.