Maximizing Future Flexibility in Electric Generation Portfolios

This paper presents a methodology to obtain flexible future capacity expansion plans under diverse types and sources of uncertainty classified as global and local. Planning flexibility is defined as the capability of a long-term planning solution to adapt cost-effectively to any of the conditions of characterizing the identified scenarios. Global (or high-impact) uncertainties allow us to create scenarios that guide the flexibility-based planning model; whereas local uncertainties allow us to create uncertainty sets that model the imperfect knowledge of each global uncertainty (GU). Our methodology, rather than choosing the most flexible plan among a set of candidate solutions, designs a flexible system that is less sensitive to the choice of scenarios. In addition to minimizing the investment and operational cost, the model minimizes its future adaptation cost to the conditions of other identified scenarios via adjustable robust optimization. Results obtained with our methodology in a 5-region US system under a 40-year planning horizon show how a flexible system adapts to future high-impact uncertainties at reasonably low costs with a low number of adaptation actions. A folding horizon process where GUs are guided by Markov chains was performed to assess the degree of flexibility of the system and its cost under multiple operation conditions.