Cash flow prediction optimization using dynamic programming for a residential photovoltaic system with storage battery

This work presents a method and an algorithm for the management of the storage battery for a residential PV system. The method is intended to minimize the electricity bill with two different strategies: the first one based on the current state of charge of the battery, the latter using weather forecasts for the minimum cash flow prediction with a predefined initial and final battery charge. The battery can be charged by the PV panels only. The investigation is highly motivated for small and medium residential homes using electrochemical storage batteries as they are forecasted as the actual better compromise. The prediction strategy utilizes daily weather forecasts, user's consumption profiles and a simple Dynamic Programming algorithm. A finite state model for the state of charge of the battery allows efficient processing. Experimental data for PV production in several meaningful days of four months and statistical data for the home consumption are used in the simulations. The simulations, with predictions of 24 h and time steps of 10' each, present the performances and their time diagrams for different control conditions: PV without and with storage battery, without and with weather forecasts. The results confirm the meaningful economical benefit of the proposed prediction method and the unavoidable influence of the sky conditions on its efficacy. The low computational complexity of the algorithm suggests its application in real time PV storage battery controllers.