General approaches for determining the savings potential of optimal control for cooling in commercial buildings having both energy and demand charges

This article presents a general approach for determining maximum monthly energy cost savings associated with optimal supervisory control for cooling in commercial buildings in the presence of utility rates that include both demand and time-of-use energy charges. The resulting tool has a month-long time horizon because of the nature of demand changes and is only useful for benchmarking the performance of simpler and shorter-term demand response and optimal control approaches. Attempts to solve this optimization problem using a centralized formulation failed and, therefore, the benchmarking problem was formulated as a dynamic optimization problem within a multi-agent control framework so that the monthly optimization problem is segmented into several sub-problems where each sub-problem involves system optimization for a shorter period of time, for example, a 1-day period. The daily-scale optimization involves determination of trajectories of zone set-point temperatures that minimize an integral cost function with a demand cost constraint determined by a demand agent. In order to further simplify the daily optimization, dynamics associated with the cooling system are neglected and optimal control of the cooling system is assumed to be based on heuristics determined through upfront analysis. This approach leads to significantly reduced computational requirements and more importantly, it provides guaranteed convergence in the multi-agent optimization. Results for a single-zone building case study are presented to illustrate the potential cost savings. In addition, a simpler and more practical short-term optimization approach with a demand-limiting heuristic is proposed and evaluated in comparison to the benchmarking optimization results for this case study and achieves most of the potential savings.