Decision tools for electricity transmission service and pricing : a dynamic programming approach

Abstract

For a deregulated electricity industry, we consider a general electricity market structure with both long-term bilateral agreements and short-term spot market such that the system users can hedge the volatility of the real-time market. From a Transmission Service Provider's point of view, optimal transmission resource allocation between these two markets poses a very interesting decision making problem for a defined performance criteria under uncertainties. In this thesis, the decision-making is posed as a stochastic dynamic programming problem, and through simulations the strength of this method is demonstrated. This resource allocation problem is first posed as a centrally coordinated dynamic programming problem, computed by one entity at a system- wide level. This problem is shown to be, under certain assumptions, solvable in a deterministic setup. However, implementation for a large transmission system requires the algorithm to handle stochastic inputs and stochastic cost functions. It is observed that the curse of dimensionality makes this centralized optimization infeasible. Thesis offers certain remedies to the computational issues, but motivates a partially distributed setup and related optimization functions for a better decision making in large networks where the intelligent system users drive the use of network resources. Formulations are introduced to reflect mathematical and policy constraints that are crucial to distributed network operations in power systems.