Urban multi-energy network optimization: An enhanced model using a two-stage bound-tightening approach

The ever-growing interdependence among power distribution, natural gas networks and transportation infrastructure entails efficient and coordinated optimization techniques. This paper proposes an operational optimization model of urban multi-energy networks that encompasses the synergies among the above components, energy converters and electric vehicles. The proposed model defines and encapsulates three energy flows in a convex optimization problem. In the distribution network, the conic relaxation-based branch flow equation is employed to characterize the alternating current power flow. The static characteristics of gas flow are described via the Weymouth equation with convex envelop relaxations. In the transportation network, the traffic flow patterns associated with the interplay among routing and charging behaviors are studied. Since these behaviors are amenable to the Wardrop principle such that the travel cost cannot be reduced by unilaterally altering travel choice, a mixed user equilibrium model is established to describe the related traffic flows. Although the relaxation quality has been improved by deriving the tighter bounds of the variables, to date, bound-tightening approaches have not yet been effectively applied to multi-energy network problems. To strengthen convex relaxations, we enhance the proposed model with a two-stage approach. The first stage improves the variable bounds through sequential optimality-based bound contraction. The second stage iteratively and successively solves the model with a dynamic bound-tightening algorithm. Based on case studies, the interdependence among energy networks is discussed. In addition, the numerical experiments corroborate the solution quality and computational efficiency benefits of the proposed approach.