Control on Central Main Canal in Arizona, dealing with supply and demand mismatches

Water is supplied to the Central Main Canal (CMC) from the Central Arizona Project (CAP) Canal, a very large supply channel that services municipalities and other irrigation districts, and from groundwater wells. CAP supplies can be prescheduled only a day in advance, while pump supplies are often dictated by power considerations. Currently, if supply and demand flows or daily volumes don't match, operators spread the mismatches through the whole canal and users get too little or too much flow. Under theses conditions, not all lateral canals can be automated to provide constant delivery rates because the Central Main would have to accumulate all the flow mismatches. To address this constraint, the district is considering adding a storage reservoir which is very expensive. In order to use all the available storage in the Central Main to reduce the capacity of the reservoir, the control system can operate the in- and outflow of the reservoir in the most efficient way. Ground water wells are used to as another source for water supply, and energy is consumed to pump the water. When the energy consumption passes a certain threshold, the cost will be 4 times the normal price. Therefore energy becomes a parameter to weigh the effect of the controller. Model Predictive Control can be a good solution to optimize the problem.

Research question:
How much usable storage is available in the Central Main? How much additional storage is needed to provide reasonable service? and how to adapt controller design to this limitation?

Approach:
A model of the Central Main Canal is set up in Sobek and the Model Predictive Controller (MPC) is applied in Matlab to use all the available storage in the Central Main. The realistic supply and demand flows are tested in the controlled model. From the simulation results, the location with the largest storage problems is selected as the potential location for the reservoir. The dimension of the reservoir can be estimated based on the extent of the storage problem.

Committee members:
Prof. dr. ir. Nick van de Giesen - TUDelft - Water resources management
dr. ir. Peter-Jules van Overloop - TUDeflt - Water resources management
dr. ir. Bert Clemmens - USDA-ARS-U.S. Arid-Land Agricultural Research Center
ir. Bob Strand - USDA-ARS-U.S. Arid-Land Agricultural Research Center