Predictive Maintenance of a Ground Vehicle Using Digital Twin Technology

The safety and reliability of ground vehicles is a motivating factor for periodic maintenance which includes fluids, lubrication, cleaning, repairs, and general observation of key subsystems. The scheduling of maintenance activities can occur at different rates such as daily, weekly, or perhaps operating time based on collected historical data and general guidelines. The availability of a digital twin (DT), which offers a virtual representation of the vehicle behavior, enables virtual system simulations for different operating cycles to explore the dynamic behavior. When field operating fleet data can be integrated with the digital twin estimates, then this supplemental information can be combined with the existing maintenance plan to provide a more comprehensive approach. In this paper, a digital twin with a statistical based predictive maintenance strategy is investigated for a wheeled military ground vehicle. The underlying models and mathematics are presented to establish a basis for this engineering tool. A case study is examined in which a DT is utilized in a computer simulation mode as a physical vehicle was unavailable to generate numerical data for signal features and condition indicators. Representative results show a high validation accuracy and reasonable training times can be achieved in support of predictive maintenance classification models. The steady progression towards a virtual engineering design and product support framework for transportation systems relies on the presence of digital twin technology and prognostic-diagnostic methodologies.