Control Systems in Future Power Markets: Market Based Demand Response, Transmission Topology Control, and Synergies with Renewable Integration
Committee Members: Advisor: Michael Caramanis, SE/ME; Ioannis Paschalidis, SE/ECE; Pirooz Vakili, SE/ME; Pablo Ruiz, SE Visiting Scholar; Appointed Chair: John Baillieul, SE/ME
Abstract: National directives emphasizing improved power grid sustainability, reliability, and economic performance motivate innovative research directions in demand response and renewable generation integration. In addition, advances in computing capabilities render feasible new techniques for long standing power market problems. This dissertation proposes and evaluates several novel decision support policies, balanced between improving existing power market operation and preparing for forthcoming technologies and capabilities. In particular, this dissertation considers (i) demand-side decision support and (ii) tractable transmission topology control.
The first part of this dissertation uses dynamic programming and duality theory to develop a decision support framework for a load aggregator (LA) managing flexible loads sharing the same short-term capacity constraints – and in particular – plug-in electric vehicles (PEVs). The LA submits inflexible and flexible bids to a liberalized hour-ahead power market, while monitoring localized network constraints. Flexible bids are associated with a reservation price and the market clears these bids as a combination of energy demand and capacity reserve, as energy only, or rejects the bid entirely. By allowing the market operator to dispatch PEVs, this strategy provides an effective means for coordination with centralized renewable generation. In addition, the provision of voltage support in the presence of distributed solar arrays is considered.
The second part of this dissertation develops sensitivity- or gradient-based algorithms aimed at reducing congestion costs by tractably including topology control in the economic dispatch. Currently, the economic dispatch minimizes generation costs subject to transmission constraints, where the status of each line, i.e., open or closed, is fixed. Recent research shows that by optimally dispatching the network topology along with generation resources, significant congestion costs may be avoided. Optimal transmission topology control requires the solution of a mixed integer program which is computationally intractable for real-sized power networks; however, it appears that much of the cost savings may be attained by changing the status of just a few appropriately selected lines. Therefore, this dissertation proposes tractable transmission topology control policies, which employ sensitivity information readily available from the economic dispatch to select candidate lines to change status.