Distribution reservoir control with demand prediction using deterministic-chaos method

Besides serving the water needs of the community, the distribution reservoir has two particular functions to perform; to adjust reservoir discharges according to time-related variation in demand and to have a reserve to handle emergencies. If priority is given to the former function, then its water level varies; if reserve function is a primary consideration, then the reservoir water level remains stably high. In actual reservoir operation it is important to maintain an adequate balance between the two functions. Demand forecasting seems helpful in meeting this need. Since the storage capacity of distribution reservoirs is no greater than several to 10-odd hours in terms of discharging time, shorter demand forecasts than before should desirably be made in hour units. We developed a method for making hourly forecasts of water demand, using the theory of chaos control and evaluated its usefulness in solving distribution reservoir problems through simulation. The future water level of a reservoir was predicted based on the data obtained by demand forecasting and water pumps were operated to keep the reservoir water level in the goal range. As a result, an adequate balance could be maintained between the two reservoir functions so that the frequency of pump operation could be reduced to improve energy efficiency.