Abstract
Efficient plant operation can be achieved by properly loading and sequencing available chillers to charge and discharge thermal energy storage (TES) reservoirs at optimal rates and times. TES charging sequences are often determined by heuristic rules that typically aim to reduce utility costs under the time-of-use rates. However, such rules of thumb may result in significantly suboptimal performance on some days. Rigorous optimization, on the other hand, is computationally expensive and can be unreliable if not carefully implemented. A novel receding-horizon control (RHC) algorithm is developed to reliably compute near-optimal control for charging and discharging the stratified sensible cool storage reservoir of a chiller plant. The algorithm provides a constant coefficient-of-performance (or cost-per-ton-hour) 24 h dispatch plan under which chillers operate at higher capacity during more favorable weather conditions. The algorithm uses a one-dimensional search requiring at most N2/2 evaluations per day of chiller performance where N is the number of planning horizon time-steps and chiller performance is modeled as a function of capacity fraction and ambient wet- or dry-bulb temperature. Analysis of four hot climates, ranging from dry to humid, indicates 2.4–2.6% energy savings under a flat electricity rate relative to the same plant operating without TES. Annual cost savings from 6% to 9% were found for electricity billed under a simple (10 a.m.–10 p.m.) time-of-use rate with no demand or ratchet components.
