Abstract (EN):
The size and the performance expected from a reservoir dam were defined according to a multicriteria decision aid methodology, involving several partners interested in the different uses of water. Simulation, criteria choice and weighting, additive aggregation and a linear program leading to efficient solutions were the main paths of that methodology.This work takes some of the techniques used in planning to establish rules for the day-by-day management of the reservoir. There are main uses for the water - ecological discharges, water supply and land irrigation - each one with daily targets. Penalty functions are set for non accomplishment of those targets according to the same criteria used in planning. Power generation is faced as an outside purpose, depending on the state of the reservoir, on the other uses of water and on the installed power capacity, previously defined according to the evaluated risk of non profitable discharges by the spillways.
Two levels of dynamic programming, both adjusted to a goal programming procedure, are used in succession according to inflows data available. Monthly information is used to establish optimal states for the reservoir in the end of each month, according to evolution of demands and recent history of inflows. A lower level of dynamic programming uses estimated daily and weekly inflows, as well as updated demands, and leads to day-by-day decisions for the water to be delivered to each purpose.On both models, on upper and lower levels, the aggregation of the values of penalty functions is to be minimised. This functions depending on the various criteria, are to be set beforehand, likely during planning decision process involving all partners.
Language:
English
Type (Professor's evaluation):
Scientific
No. of pages:
17