A Modeling Tool for Fuel Price Risk Management in Power Generation Portfolio Planning

Abstract

Power plants are significant capital investments whose returns are largely determined by such uncertain factors as future fuel prices and environmental regulations. Traditionally regulated utilities in the United States initiate power plant investment decisions through a process called Integrated Resource Planning (IRP). The aim of this centralized planning approach is to comprehensively assess both supply-side and demand-side options for satisfying load in the utility’s service territory over the next 20-30 years, and to establish a road map of actions that will serve load at least cost while adhering to reliability standards. Given the current level of uncertainty associated with many types of power plant investments, industry experts have identified a need for innovative planning practices within the IRP framework that can highlight resource tradeoffs, consider a variety of outcome metrics, and test a broader range of resource portfolios. This document describes the development and evaluation of a decision tool that is capable of capturing the tradeoff between upfront capital costs and fuel price risk in power plant investment decisions. An adaptation of mean-variance portfolio theory (MVP), a financial risk management framework, serves as the foundation of the model. As an extension of previous applications of MVP to power sector planning, this project explores the potential for implementing a parametric approach to fuel price inputs as well as a multi-period decision structure. The decision tool is evaluated under five input scenarios that represent different possible future fuel price trajectories and greenhouse gas emissions concerns. Results suggest that the parametric fuel price modeling methodology demonstrated here may be a valuable approach to future IRP applications of MVP. The multi-period structure is problematic in that outputs cannot be compiled across periods. Additionally, use of a widely accessible but computationally limited software platform precludes sufficient representation of economies of scale in power plant construction. Although these drawbacks may limit MVP applications to the exploratory analysis phase of IRP, with further development, the decision tool developed here could be useful in addressing some of the limitations of current IRP practices with respect to decision-making under uncertainty.