An optimization scheme for chiller selection in cooling plants

Providing effective cooling for buildings and industrial facilities at minimum cost is one of the main challenges in the HVAC industry. Considerable effort has been put into the optimization of existing cooling plants, but the chiller selection procedure has been relegated to a second place. This paper introduces two alternative formulations to add the chiller selection into the overall optimization problem: i) a mathematical programming approach with a single cost-based objective function and ii) a multi-objective optimization of capital cost and energy consumption. It was analyzed the case of a cooling plant with a known load profile and 13 air-cooled screw chiller models ranging from 140 to 500 TR available for purchase. The minimum objective value in the mathematical programming approach is obtained by selecting three 500 TR chillers. On the other hand, the multi-objective optimization approach produced a set of nine nondominated solutions (including the three 500 TR chiller selection). Under the second approach it is not necessary to translate the energy consumption into monetary terms. The results of both alternatives are considerable different from the straightforward approach of selecting chillers with total nominal capacity closer to the peak load (900 TR). This reveals the importance of a formal selection procedure.