Cost-oriented capacity assessment for renewable power systems with automatic reserve neural networks and information gap decision theory

The integration of large-scale, intermittent, and volatile photovoltaic (PV) systems into power grids necessitates a reliable evaluation of the system’s confidence capacity to effectively formulate scheduling and maintain power balance. This paper proposes a novel cost-oriented capacity assessment approach for renewable power systems, deriving an automatic reserve neural network (ARNN) in conjunction with information gap decision theory (IGDT). Initially, an ARNN algorithm based on clustering and recognition is developed to effectively predict PV output. Subsequently, based on the prediction error from the ARNN, a cost-oriented confidence capacity assessment model grounded in IGDT is formulated. Simulation results on an IEEE-14 bus system demonstrate the efficacy of the proposed method in evaluating the confidence capacity of renewable power systems. In this paper, the proposed approach improves the PV confidence capacity assessment capability by 85.62%, 86.12%, 88.80%, and 60.73% in the normal environment, respectively. In the extreme environment, the PV confidence capacity assessment capability is improved by 87.31%, 86.13%, 86.11%, and 62.31%, respectively.