Statistically Feasible Joint Chance-Constrained Scheduling of Integrated Distribution Network and District Heating System

Abstract

This paper proposes a statistically feasible joint chance-constrained scheduling framework for integrated power distribution networks (PDN) and district heating systems (DHS). The proposed method constructs data-driven uncertainty sets directly from samples, eliminating the need for prior distribution assumptions. It integrates joint chance-constrained programming (JCCP) with robust optimization (RO) to reformulate the original problem. The resulting model is both tractable and computationally efficient. Additionally, we introduce a novel constraint-specific uncertainty set reconstruction technique. This technique refines the uncertainty set by incorporating optimization-relevant information. It significantly reduces conservatism while ensuring system violation probability requirements. Comparative studies with state-of-the-art uncertainty optimization methods demonstrate the advantages of our approach. The proposed method improves computational efficiency by two orders of magnitude. It also achieves more cost-effective solutions than the best-performing benchmark method.