Title: AI, Innovation and Global Transformation: Interdisciplinary Perspectives on Technology, Business and Society
Editors: Dr. J. Preetha, and Dr. Siddhartha Mehrotra
ISBN: 978-81-69857-64-2
Chapter: 10
DOI: https://doi.org/10.59646/809/10
Authors: Anand Kushwah, Anshu Anand, and Shawan Mondal
Abstract
Digital Twin (DT) technology represents a paradigm shift in the lifecycle management, design verification, and health monitoring of advanced mechanical systems. By creating high-fidelity, bidirectional cyber-physical counterparts of physical assets, DT architectures enable real-time anomaly detection, predictive maintenance, and closed-loop operational control. This chapter presents a comprehensive, Physics-Informed Neural Network (PINN) and reduced-order modeling (ROM) based Digital Twin framework tailored for high-speed multi-axis mechanical assemblies and rotordynamic drivetrains. The proposed architecture synchronizes multi-rate industrial Internet of Things (IIoT) telemetry streams with continuous-time physical degradation laws. An adaptive unscented Kalman filtering scheme is integrated to dynamically estimate non-linear states and update residual fatigue metrics under non-stationary stochastic operational loads. Experimental validation across a multi-stage industrial gearbox and rotor test rig demonstrates state tracking accuracy within 2.14% Normalized Root Mean Square Error (NRMSE), while accelerating dynamic simulation speeds by over relative to standard high-fidelity finite element solvers.
Keywords: Digital Twin, Cyber-Physical Systems, Physics-Informed Neural Networks, Reduced-Order Modeling, Predictive Maintenance, Non-Linear Rotordynamics