Optimization of Real‐Time Hydrothermal System Operation

Abstract

A procedure for the economic optimization of hydrothermal power systems operation has been developed. The procedure is suitable for on‐line system utilization on a daily basis. To alleviate the inherent high dimensionality of the optimization problem, a time decomposition is used. The weekly and daily cyclic nature of load demand makes it convenient to specify operational periods in terms of weeks, days, and hours. A set of three models, i.e., monthly (four weeks), daily, and hourly, is developed. The three individual models are coupled to perform an overall optimization. A feedback mechanism is developed through the periodic monitoring of system states and the updating of the higher‐level models with a periodicity corresponding to the normal periods of the particular model. The procedure considers the seasonal variations of inflow to the system reservoirs, the variation of transmission losses with hydro and thermal plant loadings, and thermal unit commitment and dispatch. Application is made to a moderately sized hydro‐thermal system in the People's Republic of China. It was shown that the developed procedure is feasible and practical for on‐line daily operations. Significant reductions in the operating costs for the Chinese system have been obtained.