Model Predictive Control of Solar PV-Powered Ice-Storage Air-Conditioning System Considering Forecast Uncertainties

This paper proposes a dynamic programming (DP)-based stochastic model predictive control (SMPC) method for the economic operation of solar PV-powered ice-storage air-conditioning (PIAC) systems. The forecast data of PV generation and building cooling load are considered as stochastic variables in this paper. To deal with the uncertainties of the day-ahead forecast data, Latin hypercube sequential sampling, Cholesky decomposition and Simultaneous backward reduction are adopted to provide representative scenarios for SMPC. The value function matrix is employed to solve the receding-horizon optimization problem formulated by DP. With updated short-term forecast information, SMPC is able to reduce the impact of inaccurate forecasts on the operation of PIAC systems. A study of typical operation cases demonstrates the effectiveness of the proposed method, which ensures the satisfaction of cooling supply and yields solutions closer to the global optimality than the traditional MPC method.