Power generation mix evolution based on rolling horizon optimal approach: A system dynamics analysis

Abstract System dynamics is a well-established methodology to analyze the behavior of complex systems through computer simulation. Different from the traditional system dynamics, in this paper, the peak shaving and frequency control reserve constraints are incorporated into power generation mix planning to ensure system security and efficiency. However, it is an intractable task to consider this multi-market equilibrium and multi-period coupling planning problem in system dynamics. To ameliorate this inherent drawback, a novel planning model based on the rolling horizon optimal approach is proposed. First, by adding the constraints to the objective function, the Hamilton function is constructed to calculate the optimality conditions, and then the primal planning problem can be converted to a capital recovery problem. Considering the stochastic characteristics in the future market, a rolling horizon approach is adopted to update the planning information and optimize the power mix continuously. Next, an approximate gradient based on Pontryagin's minimum principle is developed to simplify the optimal iteration processes. To demonstrate its feasibility, the sensitivity analysis suggests that the reduction of renewable energy cost and the wide allocation of flexible resources are two major factors to achieve a high-level penetration of renewable energy.