The Value of Considering Autocorrelation Between Inflows in the Stochastic Planning of Water Resource Systems

Reliability programming formulations offer a family ofexplicit stochastic models for planning the operationof complex water resources systems. These models usecumulative probability distributions of the sum ofinflows to characterize their variability in theplanning period. Applicability of these models for avariety of problems has been limited, mainly due tothe assumption of independence between inflows indifferent time periods that leads to the derivation ofconservative operating policies. This paper presentsthree new approaches to overcome this limitation. Theperformance of the proposed approaches is demonstratedthrough comparison of the operating policies derivedfrom these approaches and the Independent Approach. Operational planning of the Manitoba Hydro energygeneration system, a predominantly hydro-based utilitycompany in Manitoba (Canada), is used as the case study. Copyright Kluwer Academic Publishers 1999