Parallel-Chiller Optimization Using Continuous Barnacles Mating Optimizer Considering Chiller Availability and Cooling Load Variations

Heating, ventilation, and air conditioning systems are known for their high energy consumption, and the operation of parallel chillers is an important way to increase the energy efficiency of these systems. In the field of optimal chiller loading, algorithms are used to find the best loading combinations of chillers for different systems. Existing studies assume that all chillers in a case study are available and included for optimization. Furthermore, due to the iterative nature of stochastic algorithms used for optimal chiller loading, they can present huge variations between results, leading to suboptimal performance. This study aims to solve these problems by modifying an existing optimization algorithm, which is the barnacle mating optimizer. By allowing manual overriding of the on and off statuses of each chiller, these modifications improve practical applications in cases of chiller maintenance or breakdown. Furthermore, optimization precision was increased by introducing continuous cycles in each optimization procedure, followed by filtering and sorting of the best loading distributions. With numerical testing, the proposed continuous barnacles mating optimizer performed on par with its native counterpart at high cooling loads while providing up to 11% energy savings and reliably following optimization constraints at lower loads. Manual chiller switching functions of the modified algorithm also demonstrated good stability despite having fewer chillers during high-load situations. This study encourages improvements other than power consumption in the optimal chiller loading field to improve their feasibility in practical applications while addressing the 11th goal of the Sustainable Development Goals, prompting another step towards more sustainable buildings and cities.