Demystifying Stochastic Analysis in Resource Planning: An Introduction for Regulators, Stakeholders, and Engaged Observers of Electric System Resource Planning Processes

Navigating uncertainty is fundamental to building a successful long-term electric system resource plan. Whether it is short-term hourly fluctuation in system conditions; medium-term variation in seasonal states; or longer-term changes in factors such as government policy, technology cost, and load growth, the principal objective of the electric system planner is to identify 10- to 25-year system resource portfolio plans that can perform cost-effectively and reliably. Practices such as scenario analysis and sensitivity analysis are widely used in resource planning to navigate uncertainty and manage risk. Both approaches can provide important information about the range of potentially beneficial decisions but may miss identifying adverse outcomes or further opportunities due to the limited futures explored.

In contrast, stochastic analysis often explores a relatively large set of potential futures and explicitly considers risk. It is important to note that, in practice, scenario, sensitivity, and stochastic analysis complement one another across the different phases of a resource planning assessment. Scenario and sensitivity analyses can help identify candidate resource plans as they are well-suited to efficiently exploring the range of potential plans for a discrete set of uncertain futures. The outcome is thus a tractable set of candidate plans to consider. Sensitivity analysis and stochastic analysis may then be used to evaluate these plans, with the latter affording a means to characterize risk over more possible futures. Overall, the combination of scenario, sensitivity, and stochastic methods balances computational efficiency of the planning assessment with providing the most insight into risk and uncertainty.

This brief focuses on demystifying (1) the use of stochastic analysis in evaluating candidate resource plans; and (2) the interpretation of results from a stochastic resource planning analysis, including how the planner’s risk tolerance intersects with the potential selection of a “preferred” resource plan. To solidify the concepts presented, a simple, illustrative example is outlined and used throughout the discussion. Comparisons to traditional or “deterministic” resource planning analyses are made.

Authors Nidhi R. Santen

View on EPRI.com

Keywords