Development of an Optimisation Model of Air-Cooled Chiller Maintenance Management to Improve Availability Performance

Air-cooled chillers remain important to climate control in many industries. However, ensuring high availability of such systems has remained a big challenge to their operations, especially in developing countries like Tanzania. The maintenance practices currently adopted are mainly reactive, translating to a frequent occurrence of failures that result in shortened equipment life and energy wastage. In this paper, an optimum data-driven model is developed and validated for managing the maintenance of air-cooled chillers to advance the performance of their availability. Quantitative and qualitative data were collected from seven chillers at Julius Nyerere International Airport using a mixed-methods methodology approach at Dar es Salaam, including operations records, performance indicators, and interviews with experts. Multiple regression analysis highlighted preventive maintenance frequency (β = 0.312), compressor performance (β = 0.241), control system reliability (β = 0.225), and energy performance (β = 0.202) as the strongest positive determinants of availability. Contrary to findings in other studies performed under more extreme climates, environmental factors such as ambient temperature and dust had low significance. This can be explained by the stable coastal climate of Dar es Salaam and the relatively controlled infrastructure at JNIA. The model was then validated with an average prediction accuracy of about 98%. This was indicative of its efficiency in predicting the availability of chillers.