Work

Integrating Domain Adaptation and Causal Discovery in Digital Twins for Plastic Injection Molding

Year: 2025

Type: conference-paper

Abstract: The development of digital twin (DT) technology is transforming the manufacturing industry, allowing for applications such as real-time process monitoring, predictive maintenance, and optimization of production systems. However, traditional DT frameworks often maintain a high-level perspective and do not consider the challenges arising from real industrial data. Factors such as changes in input materials and environmental conditions can hinder the functioning of DTs and are frequently overlooked. Using the example of the plastic injection molding industry, this paper highlights the necessity of including technologies such as domain adaptation and causal discovery in the development of DTs. Domain adaptation enables adaptation to changes, while causal discovery provides a deeper understanding of the underlying process dynamics. Their combined adoption allows DTs to achieve improved robustness and flexibility, extending their applicability across diverse manufacturing scenarios.

Source: IEEE International Conference on Pervasive Computing and Communications Workshops and other Affiliated Events (PerCom Workshops)

DOI: 10.1109/percomworkshops65533.2025.00050

Authors: Gian Marco Paldino, Olivier Caelen, Marouene Oueslati, Marc Ansay, Tojo Valisoa A. Johanesa, Gianluca Bontempi

Institutions: Université Libre de Bruxelles, Sirris, University of Mons

Language: English

Topic: Digital Transformation in Industry

Subfield: Industrial and Manufacturing Engineering

Field: Engineering

Domain: Physical Sciences

SDG: Industry, innovation and infrastructure

Open Access status: green

Funder: Service Public de Wallonie