An Optimization Decomposition Framework for Principled Multi-Timescale Market Design and Co-Optimization

Johns Hopkins University, Baltimore MD

Investigators

Abstract

The United States electricity generation mix is undergoing a fundamental transformation. Older thermal generation units, which can be reliably scheduled depending on system needs, are being replaced by newer renewable-based units with substantially different operational constraints. As a result, the core principles under which existing markets operate are rapidly evolving. New Federal Energy Regulatory Commission regulations in response to the increasing variability and uncertainty that renewable generation introduces have spurred Independent System Operators to make fundamental market design changes. However, the existing design process for market products, which include heavy stakeholder involvement as well as some (but not extensive) numerical testing, often has an ad-hoc flavor and uncertain outcomes. Also, the complex interactions that arise, due to the simultaneous participation of resources in several markets operating at different timescales, have caused some of the recently designed new markets and products to lead to counter-intuitive outcomes. This work addresses these challenges by developing analytical tools and thorough simulation setups that can support more effective market designs based on a sound theoretical framework that takes into account the interdependencies of decisions across markets and time scales. This project develops a principled market design methodology that can evaluate the performance of electricity market products for systems with variable energy resources and other emerging technologies. Central to this research is a novel top-down approach that decomposes the operator objectives into desired market products while identifying multi-market interactions, even across time-scales, and communications requirements for co-optimization. This methodology does not aim to redefine from scratch the whole market ecosystem already in place. Instead, the decomposition approach seeks to define suitable subproblems that fit within the existing architecture, which allows comparisons between the new solution and existing market designs, and the identification of missing market components. To show its relevance to current market design challenges, the proposed approach will be first applied in the design of coordinated real-time energy and regulation markets, and subsequently, be augmented to include additional market products and timescales. The benefits of the proposed market designs will be assessed using extensive simulations of detailed market models that will be run on realistic simulation setups. Thus this research includes a mix of innovative design and algorithmic development grounded in deep knowledge of existing and practical market operation structures and business rules, which will be informed by collaborations with the California Independent System Operator, the Pennsylvania, Jersey, Maryland (PJM) market monitor, and the Electric Power Research Institute. In this way, this work intends to provide novel and beneficial research results that are directly applicable to existing and new electricity markets.