Energy management optimization of a smart wind power plant comparing heuristic and linear programming methods

This paper aims at optimizing the energy management of a smart power plant composed of wind turbines coupled with a Lithium Ion storage device in order to fulfill a power production commitment to the utility grid. The application of this case study is typically related to islanded electric grids. Our work particularly investigates and compares two classes of energy management strategies for design purpose: a first capable of providing the global optimum of the power flow planning from a Linear Programming (LP) approach thanks to a priori knowledge of future events in the environment; a second, based on a classical control heuristic without any a priori knowledge on the future, applicable in real time. Beyond the future objectives in terms of system design (techno-economical sizing optimization), the comparison of both approaches also aims at improving the predefined heuristic from the analysis of the ideal reference provided by the global LP optimizer. In this scope, a linear power flow model of the power plant is developed in compliance with a LP solver (Cplex). A particular attention is paid to the techno-economic optimization including storage cost evaluation, commitment failure penalties and exploitation gains. Simulations and optimizations are carried out over one year in order to take variability and seasonal features of the wind potential into account.