Chiller Plant Operation Optimization With Minimum Up/Down Time Constraints

Abstract

A chiller plant includes chillers, cooling towers, and pumps. Chiller minimum up/down times (MUDTs) are relatively long among a plant but short as compared with the time interval. MUDT constraints are generally ignored in optimization but handled heuristically. With the numbers of active units and nonlinear heat exchange, chiller plant optimization is a mixed-integer nonlinear problem. Complexity of the problem will increase if MUDT constraints are considered. In this letter, potential energy savings and complexity increase caused by considering such constraints are explored. To obtain near-optimal solutions efficiently, a method is developed based on a recent decomposition and coordination approach. Chiller subproblems are solved in two steps, first without MUDT constraints to establish stage-wise costs. Then, possible state transitions are developed based on MUDT constraints from the states with feasible costs before dynamic programming (DP) is used. For practical problems, MUDT constraints are rarely violated. A second method is developed with DP replaced by local search to reduce computational effort. Numerical testing shows that our second method is faster and without much performance degradation as compared with the first one. However, energy savings are small with complexity increased significantly by using both methods as compared with those where MUDTs are handled after optimization. Therefore, there is no need to consider MUDT constraints in optimization.