Implementation of LIVE Digital Twin Enabled Smart Maintenance Using Smart Structural Sensors

This research paper presents the implementation including detailed design and development of a data-driven predictive maintenance methodology called LIVE which stands for Learn, Identify, Verify, Extend. The four-phase methodology is aimed at developing digital twins for the purpose of predictive maintenance. A case study on a lab-built pipe structure was conducted to evaluate the effectiveness of the LIVE methodology. The modal testing data from the physical and virtual models are used to validate and optimize a network of smart structural sensors (SSS). An experimental modal analysis is conducted with defined healthy and unhealthy cases to determine the vibrational behavior of the pipeline system. The outcome of the case study is a collection of algorithms capable of predicting and locating the induced fault in the system as well as a precursor to the online maintenance synonymous to digital twins. The developed algorithms show a high prediction accuracy of 95%. This study demonstrates the potential of the LIVE Methodology as a promising solution to problems in the predictive maintenance space.