AMACHINELEARNINGAPPROACHFOR PREDICTIVEMAINTENANCEANDRESOURCE ALLOCATIONATTHEUNIVERSITYOFZAMBIA

Abstract

Predictive Maintenance essentially involves predicting breakdown of the equipment to be maintained by detecting early signs of failure in order to make maintenance work more proactive. The application of such maintenance strategy requires the cooperation of several agents and involves knowledge and skills, it requires processing and computing of relevant data signals in the from the equipment in order to extract meaningful knowledge. Preventive Maintenance in a broad topic making it impossible to cover all the subtopics in this research paper. Having this into consideration, this paper focuses on the Predictive Maintenance Model (PMM) for an Office Air Conditioner (OAC) using machine learning techniques such as XGBoost and LSTM networks. This research paper examines the advantages, limitations and advancements in predictive maintenance enabling to minimize downtime, energy consumption, cost savings, and increasing comfort in office workplace. One of the most useful features of the paper is the contribution to the body of knowledge by delivering a UI tool for facilities managers to make decisions quickly using machine learning predictive modelling techniques.In addition, the paper also highlights future work opportunities such as the inclusion of external data sources, development of adaptive models, enhanced IoT integration, and improved hyperparameter optimization methods