A digital twin-based remote assistance system with on-site augmented reality support for machine maintenance
Date
2024-12
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Stellenbosch University
Abstract
Across industries, the need to work and provide assistance from a remote location was exacerbated by travel restrictions enforced during the Covid-19 pandemic. This is especially true for companies that offer after-sales support of their equipment to international clients. As such, there is a growing need for solutions that allow specialists to provide effective remote assistance to On-site Technicians working on equipment, e.g. performing maintenance. To address this need, this thesis presents the development of a proof-of-concept Digital Twin-based Remote Assistance System (RAS) with Augmented Reality (AR) support for On-site Technicians performing maintenance on machines. Through an analysis of the remote assistance context and stakeholder needs, the thesis formulates a set of functional and non-functional requirements. The requirements are used to guide the design of the architecture for the RAS, which is based on an existing service-oriented architecture for a system of Digital Twins (DTs). The RAS architecture is implemented in a case study, which is based on the remote assistance activity for a specific machine of a South African mining equipment supplier. The case study implementation uses various services of the Microsoft Azure platform for the DT and digital service components. In the case study, the On-site Technician is equipped with a Microsoft Hololens AR headset, which hosts an AR user interface developed using Unity. The functionality and performance of the case study implementation of the RAS is verified against the functional and non-functional requirements formulated in the thesis. The verification is based on the results of two experiments. The results of the verification experiments show that the implemented RAS met the functional requirements, achieving near real-time transmission of operational data, enriched communication between the Remote Assistant and On-site Technician and effective use of AR in supporting the On-site Technician. Non-functional requirements, such as vendor independence, system flexibility, and compatibility with network constraints, were also verified. The implemented RAS could be reconfigured to include other DTs and proved adaptable to different hardware and software platforms – highlighting its potential for scalability and integration into various industrial environments. Based on the results of the verification of the case study implementation, the thesis presents a holistic evaluation of the developed RAS against the thesis objective. The evaluation confirms that the RAS improves remote assistance by providing near real-time, accurate guidance during maintenance activities. The RAS allows On-site Technicians to view relevant data, such as equipment status and maintenance instructions, directly within their AR headset, which can potentially reduce errors and downtime. While the proof-of-concept RAS focused on supporting machine maintenance, the flexibility of the RAS – as enabled by the developed architecture and implementation technologies – is expected to support applications in other industrial sectors and remote assistance scenarios.
Description
Thesis (MEng)--Stellenbosch University, 2024.