The cost of events in temporally-resolved energy system models: a stochastic programming approach

Abstract To deal with random exogenous events, energy systems require the ability to flexibly adjust their operational schedule. Here we explore how such events can be incorporated into temporally resolved energy system models as forced deviations from planned operation at random times. This leads to a formulation of the investment and dispatch problem as a stochastic program. We apply our approach to a conceptual model of an industry park with renewable generation, subject to randomly timed, homogeneously distributed, events requiring it to lower or increase power consumption. The cost of events occurring can exceed the cost of maintaining the capacity to adjust to them. We find that incorporating a probabilistic sample of such events during system planning can reduce the cost of adjusting to them during operation by up to 50%. Furthermore, investment decisions seem to have a minor impact on the cost of adjusting to the events - the main benefits lie in an optimized operation of the industry park. Thus, in this setting, existing assets can be leveraged to provide a cost effective response.