AI-Based Surrogate Models of Digital Twins for Industrial Processes

Digital twins represent virtual replicas of physical systems, integrating real-time data and advanced analytics to monitor, simulate, and optimize industrial processes. This research delves into the application of AI-based surrogate models to improve the efficiency and accuracy of digital twins for industrial processes. The study employs machine learning techniques to develop computationally efficient models that maintain high accuracy. The integration of advanced sampling techniques and challenges related to data quality and interpretability are highlighted, proposing solutions to improve model robustness and reliability.

I. Introduction

Surrogate models are instrumental in providing virtual rep-resentations that mirror physical objects or processes, yet they serve distinct purposes in the realm of simulations and digital transformation. Surrogate models, also known as proxy or em-ulator models, are simplified mathematical or computational models designed to approximate the behaviour of complex systems or processes. They serve as surrogates for computationally expensive or time-consuming simulations, enabling rapid and efficient analysis, optimization, or decision-making. Surrogate models provide a trade-off between accuracy and computational cost, making them valuable for uncertainty quantification and optimization tasks [1]