Incorporating daily flood control objectives into a monthly stochastic dynamic programing model for a hydroelectric complex

Abstract

A monthly stochastic dynamic programing model was recently developed and implemented at British Columbia (B.C.) Hydro to provide decision support for short‐term energy exports and, if necessary, for flood control on the Peace River in northern British Columbia. The model establishes the marginal cost of supplying energy from the B.C. Hydro system, as well as a monthly operating policy for the G.M. Shrum and Peace Canyon hydroelectric plants and the Williston Lake storage reservoir. A simulation model capable of following the operating policy then determines the probability of refilling Williston Lake and possible spill rates and volumes. Reservoir inflows are input to both models in daily and monthly formats. The results indicate that flood control can be accommodated without sacrificing significant export revenue.