On March 13, I had the pleasure of giving a guest lecture in Professor Bentley Oakes‘s LOG6953FE – Digital Twin Engineering graduate course at Polytechnique Montreal. Professor Oakes has been offering this course in the past two-three years and it is always a pleasure to share some of our topics with the audience at Poly MTL.
Title: I Reap What ISO: Engineering Digital Twins on Standardized Architectures
Abstract: Digital twin technology has experienced burgeoning adoption in a wide array of domains, from automotive systems through smart agriculture to healthcare. Contemporary digital twins come in many flavors: physics-based and simulation-heavy twins, AI-enabled predictive twins, and human-centered twins equipped with rich visualization and decision-support cockpits. This diversity led to highly heterogeneous and often ad-hoc architectures, hindering reuse, interoperability, certification, and long-term evolution. Standardized reference architectures help alleviate these issues by providing high-level design guidance that different vendors can follow to render their digital twin solutions comparable and compatible. Unfortunately, at the time of this lecture, there is only one architectural standard available for digital twins: ISO 23247 (Digital twin framework for manufacturing). Although originally developed for manufacturing systems, the ISO 23247 standard has been extended into a handful of other domains, such as automotive systems, battery systems, space debris detection, and AI simulation. Although these extensions were initially developed mostly out of necessity (on account of the lack of alternatives), growing evidence suggests that the standard translates well to various domains and delivers reasonable benefits.
In this lecture, we argue that digital twin developers should reap the benefits of the ISO 23247 by informing their design decisions from the standard or directly implementing it in digital twin solutions. Through illustrative examples, we will identify the strengths and limitations of the standard and discuss the ways in which the standard aids the engineering of digital twins. Ultimately, we position standardization as an enabler of robust, evolvable, and trustworthy systems.
Bio: Dr Istvan David is an Assistant Professor in the Department of Computing and Software, Faculty of Engineering at McMaster University. His research is centered around modeling&simulation, machine learning & AI, and the application of such foundations in the design of complex heterogeneous systems, such as digital twins and cyber-physical systems. Dr David is an award-winning author, reviewer, and organizer of flagship scientific events in his communities. Occasionally, Dr David likes to pretend that he’s an engineer.
