Abstract

The Digital Twin (DT) technology has emerged as one of the most prominent technologies for different applications including the machine condition monitoring, fault diagnosis, and predictive maintenance over the past decade. However, a major challenge in its widespread adoption is the development of comprehensive and generalized Digital Twin solutions. To address this, the LIVE Digital Twin framework has been introduced as a structural framework to develop and operate Digital Twins. LIVE stands for the main four stages of the framework: Learn, Identify, Verify, and Extend. A crucial aspect of LIVE Digital Twins is the integration of both Low-Fidelity (LF) and High-Fidelity (HF) simulations to manage various stages of Digital Twins’ life span. This paper uses the LIVE Digital Twin philosophy for predictive maintenance of rotary machines and focuses on the creation and application of an integrated dynamic Low-Fidelity simulation required as a main feature of this system. As part of this effort, a Simple Structural Dynamics (SSD) model was developed based on Finite Element Analysis (FEA) and Newmark’s time integration method. The Simple Structural Dynamics model was applied to a case study involving a rotary machine, where fundamental frequencies, mode shapes, and transient responses were analyzed for both healthy and faulty conditions. The results obtained using Simple Structural Dynamics were compared with those generated by a High-Fidelity simulation, demonstrating that Simple Structural Dynamics effectively predicts the system behavior while remaining computationally efficient enough to perform real-time simulations using the sensor data collected. The Simple Structural Dynamics proved to be computationally efficient, and it is highly scalable. Furthermore, the study thoroughly examined the impact of different defects, including cracks, unbalance, and bearing faults.

IPC Classification

Keywords