Tackling nonlinearities in microgrid codesign through linear programming approaches

Purpose A microgrid system is composed of several components such as producers or storages, destined to age depending on how they are used, through the energy management system (EMS) which can be ruled by optimization. According to the modeling granularity and frequency of updating components data, aging consideration introduces nonlinearities. The characteristics of certain case studies – such as increased energy demand or uncertainties in renewable production – may require long-term simulations. In some cases, specific constraints necessitate the use of a mixed-integer linear programming (MILP) formulation. As a result, computational times can become very long, hindering for search of the solutions. This study aims to build an optimization framework that addreses these various challenges, focusing Toulouse airport study case. Design/methodology/approach This paper investigates a microgrid codesign optimization approach hybridizing linear programming (LP) and MILP for optimizing both sizing and EMS decision variables. This approach, built to find a compromise between computational calculation time and precision on technical-economic considerations, takes account of nonlinearities related to the battery capacity degradation and overruns of subscribed power. Findings Our case study deals with an airport microgrid, simulated over 22 years old, using photovoltaic and wind turbine energy productions to be become an autonomous platform. This method is effective and robust, integrating battery replacement costs, which may increase the net present cost by up to 20%. The obtained results show the feasibility of achieving up to 60% energy autonomy. Originality/value This contribution introduces an original way of managing nonlinearities in LP optimization models, including both quotient-type and product-type (bilinear) relationships between decision variables, converting an optimization variable into an input data known in advance.