UC Irvine

In this thesis, the problem of electricity demand shifting for the cooling needs of a large institution using a thermal energy storage (TES) tank is considered. The system is formed by electric chillers, cooling towers and a TES tank that can store energy for the cooling demand of most days, but not for the hottest ones. The goal is to supply all the cooling needed while minimizing the cost. This is done by shifting the cooling demand to night and early morning hours, when electricity is cheaper and due to lower temperatures, the chillers work more efficiently. This is all done with the help of the TES tank, that acts as a buffer storing chilled water. After a series of assumptions and simplifications, the cost function becomes convex and thus a minimum solution exists. However, from previous work only the chillers were considered, omitting the negative effect that other components of the system, such as cooling towers, had on the overall cost of operation. Using data from the operation of the power plant under real conditions, a method to model the whole system is presented in this thesis. In addition, the algorithm relied on the knowledge of an accurate prediction of the cooling demand, which obviously is not known in advance. A method to predict it starting from a forecasting of the temperature is presented. Finally, the algorithm can be easily modified to allow the imposition constraints that limit the maximum power use of chillers, during specific periods, in response to the overall needs of the micro-grid.