Decision Support for Short Term Export Sales From a Hydroelectric System

by Donald J. Druce, BC Hydro and Power Authority, Canada,

Document Type: Proceeding Paper

Part of: Computerized Decision Support Systems for Water Managers


A stochastic dynamic programming (SDP) model was recently developed and implemented at British Columbia Hydro and Power Authority (B.C. Hydro) to provide decision support for short term export sales. The model was written around the operation of Williston Lake, the 40,000 million cubic meter storage reservoir that serves as the swing reservoir in what is essentially a pure hydroelectric system. Adjacent to Williston Lake is the 2416 MW G.M. Shrum (GMS) power plant and just downstream is the 700 MW Peace Canyon (PCN) run-of-river project. The SDP model determines the monthly operating policy for the GMS/PCN plants that will meet the (forecasted) residual domestic load and maximize the expected net export revenue. Uncertainty is introduced through the effect of historical weather sequences on the inflows to Williston Lake, on the generation supplied by other hydroelectric projects and on the domestic load. Monthly releases are contingent on the level of storage in Williston Lake. The slope of the model's objective function at any level or time then defines, not only the marginal value of water stored in Williston Lake, but also, the marginal cost of supply for the B.C. Hydro system. That system marginal cost sets the minimum price for any spot sales in the export market. Additional aspects of the subject are discussed.

Subject Headings: Hydro power | Decision support systems | Lakes | Power plants | Computer programming | Reservoirs | Stochastic processes | Dynamic models | British Columbia | Canada

Services: Buy this book/Buy this article


Return to search