Development of a Digital Twin-Based Facility Monitoring Framework for Predictive Management of Physical Infrastructure in Polytechnics in South-West Nigeria

Abstract

Physical infrastructure in Nigerian polytechnics is commonly repaired only after failure, with no continuous monitoring of its condition. This study designed, developed and evaluated a digital twin-based facility condition monitoring framework for predictive management of physical infrastructure in polytechnics in South-West Nigeria. A design-and-development approach combined infrastructure assessment with construction and evaluation of the framework. A total of 143 IoT sensor nodes were deployed across five polytechnic facilities to monitor temperature, humidity, energy consumption, structural vibration and occupancy. The framework integrated sensing, virtual-model, data-integration and analytics layers, using long short-term memory (LSTM) networks for anomaly detection and maintenance forecasting. Over a 12-month evaluation, mean energy consumption fell by 26.9%, emergency repair work orders by 74.5% and reactive maintenance work orders by 61.3%. The LSTM model achieved 96% fault prediction accuracy, outperforming Random Forest (95%) and Support Vector Machine (86%). The estimated annual net financial benefit was ₦9.93 million, with a payback period of 2.8 years. Experts rated the framework highly for functionality (3.60), reliability (3.40) and replicability (3.60). Digital twin-based monitoring is technically feasible, economically viable, and practically valuable for predictive infrastructure management in Nigerian polytechnics.