A Digital Twin-driven framework for predictive maintenance in water infrastructure using BIM and Deep Learning

Asset maintenance management ‎‎plays an integral role in the ‎efficient ‎operation of civil ‎infrastructure. ‎Predictive ‎maintenance (PdM) is a ‎‎technique, where multiple steps of ‎‎data preparation and data analysis ‎‎are employed to predict an asset ‎‎fault. Despite the extensive ‎‎application of PdM in ‎‎manufacturing and oil industries, ‎‎PdM implementation in critical civil ‎‎infrastructure such as water ‎‎infrastructure remains a challenge. ‎This study ‎proposes a digital twin-‎based asset ‎maintenance ‎management ‎approach, which ‎leverages deep ‎learning, industry ‎foundation ‎classes, and building ‎information ‎modeling. ‎Findings ‎from this research are ‎expected to ‎facilitate effective PdM ‎‎implementation in water ‎‎infrastructure.‎