OPTIMAL OPERATION OF THREE – PHASE POWER DISTRIBUTION SYSTEM INTEGRATED MULTI-ENERGY

Abstract

The global energy landscape is undergoing a shift from traditional fossil fuel-dependent generation to decentralized generation fueled by renewable energy sources, driven by the imperative to address climate change and navigate environmental uncertainties. This mass integration of intermittent and variable renewable generation has brought significant challenges such as supply uncertainty and variability to the distribution systems. The transition towards sustainable and low-carbon energy systems countering these technical challenges necessitates the integration and optimal coordination of multiple energy vectors across electrical, thermal, and gas networks. This convergence, known as multi-energy systems (MES), offers significant potential for enhancing efficiency, flexibility, and resilience. However, the inherent non-convexities and intricate couplings within and across these subsystems pose formidable computational challenges for their optimal operation. This thesis presents a comprehensive optimization framework tailored specifically for the operation of MES comprising three-phase unbalanced power distribution systems (UPDS), natural gas distribution systems (NGDS), and district heating systems (DHS). The proposed framework employs a multi-period optimization approach to address the broader class of optimization problems of the integrated MES while ensuring reliable system operation within grid constraints. A novel bi-level solution strategy is developed for three-phase unbalanced power distribution system, decomposing the mixed-integer non-linear program (MINLP) into a mixed-integer linear program (MILP) at the first level and a non-linear program (NLP) at the second level. Second-order cone programming (SOCP) and polyhedral relaxations are employed to convexify the non-linear subproblems, facilitating computational tractability. Notably, the SOCP relaxations for UPDS and NGDS are inexact under the original feasible space, leading to feasibility gaps between the relaxed and non-convex problems. Also, the polyhedral relaxations for DHS results in non-zero relaxation errors in their original search space. To address these challenges, an innovative solution recovery procedure is developed for each subsystem, enabling the retrieval of high-quality feasible solutions from their relaxed counterparts. Furthermore, an iterative successive bound tightening algorithm progressively decreases the relaxation error and improves solution quality while maintaining computational i efficiency. The three-phase coupling facilities interconnecting the subsystems, such as combined heat and power units, are thoroughly modeled within the UPDS, capturing the intricate mutual coupling effects and phase imbalances. A decentralized consensus-based alternating direction method of multipliers (ADMM) approach decomposes the multi-energy operational problem into network-level subproblems, facilitating scalability, requiring limited communication, and preserving data privacy across subsystems. The framework is applied to case studies involving modified IEEE distribution test systems, each coordinating with district heating and natural gas networks. The results validate the proposed strategy, achieving optimal solutions with relaxation errors below specified thresholds. Significant reductions in real power losses, phase voltage unbalance rates, and operational costs highlight the techno-economic benefits of the integrated MES operation. The key contributions include a comprehensive framework addressing challenges posed by each energy system, inexact convex relaxations, solution recovery, and the need for decentralized coordination in multi-energy systems. The insights can guide stakeholders in optimally operating interconnected energy systems while ensuring reliability, operational feasibility, and computational efficiency.