Key points are not available for this paper at this time.
This paper describes the development and evaluation of a near-optimal control strategy for ice storage systems. The strategy is based upon simple heuristics that were developed from daily and monthly simulations of cooling systems with internal melt, area-constrained ice storage tanks. Dynamic programming was used to obtain the optimal control trajectories which minimized an integrated energy and demand cost function for both the daily and monthly simulations. In addition to leading to simple heuristics, the monthly optimal control results were used as benchmarks to evaluate the performance of both conventional and the new control strategy. For a range of partial-storage systems, load profiles, and utility rate structures, the monthly electrical costs for the rule-based control strategy were, on average, within about 3% of the optimal costs. In contrast, the monthly electrical costs associated with the most common conventional control strategy, chiller-priority control, were as much as 20% greater than optimal, whereas a simple storage-priority strategy yielded costs that were within about 6% of optimal. The rule-based strategy can be easily implemented within a small micro-processor controller and only requires measurements of the system cooling requirement, building electrical usage, and state-of-charge of storage.
Drees et al. (Tue,) studied this question.