Simulated Digital Twins for Industrial Applications: Technical Description webpage

Background and goals

Prominent and mature industries widely utilize software simulation techniques –digital twins – to represent their key processes and products. The availability of a reliable simulated digital twin assists the organizations to diagnose their existing problems, to predict emerging scenarios and finally to optimize their costs. However, the traditional method of creating a simulated digital twin is a costly, resource-intensive and laborious task. Further, digital twins require regular maintenance to ensure ongoing relevance; this maintenance again demands ongoing financial cost, trained human resources (expert personnel) and also substantial computational resource. As a result of these costs, the usage of digital twins is often restricted only to the extremely critical functions of the organization; most industries are unable to unlock and leverage the potential of simulated digital twins at a broader scale.

Our emerging research offers a radically different solution: Recent developments in machine learning and computational neural networks offer techniques to automate the creation and the maintenance of digital twins. Based on our recent research, we can build an automated digital twin as self-learning neural network. This neural network identifies the nuances of the underlying domain and simulates it innately. Such an automated digital twin can largely replace the current traditional practices where digital twins are carefully hand-crafted and painstakingly maintained by a dedicated team of expert personnel. Further, the principles and techniques proposed from our research are transferrable across industry segments and across domains. Hence, we can implement, validate and utilize our novel techniques with varied different industry partners.

Objectives and benefits

We propose a proof-of-concept research project to help take our research from laboratory to the industry. Our primary aim is to rigorously validate the technical applicability, cross-domain flexibility and practical usability of our research. Our secondary aim is to provide rough indications on the financial viability of our methodology as compared to existing options.

Results

The primary purpose of this research project is to trigger a research competence in Arcada UAS for industrial simulations. In this perspective, the research grant will act as the seed fund for facilitating the formation of the research group that then explores further research avenues and thereby builds more sources of funds from industry and academia. Hence, the following outcomes are expected from this research:
1. Offer a starting point for discussions with industry partners who can fund future research in related domains.
2. Participate in academic consortia that execute joint research on simulations
3. Leverage larger grants from funding agencies such as Business Finland and ERC

Societal impact

This research project directly aligns with the purpose of the Fonden för teknisk undervisning och forskning (TUF) and the strategic goals of Stiftelsen Arcada by focusing on the application of simulated digital twins in Finnish industries. Digital twins, powered by reliable AI and big data, offer significant potential to enhance industrial efficiency, promote sustainable practices, and drive innovation. Our research will investigate how digital twins can contribute to a circular economy by optimizing material flows and energy consumption, while also exploring their impact on business economics in the digital society. Moreover, the project will foster smart business collaboration and contribute to the development of an innovation ecosystem by bringing together researchers and industry partners. By focusing on practical applications and real-world impact, our project generates valuable insights that will benefit Finnish industries and contribute to the advancement of technical research and education.
The aims of the proposed research project are intrinsically linked to the core purpose and strategic focus areas of Fonden för teknisk undervisning och forskning (TUF) and Stiftelsen Arcada. We have identified several key areas of significant overlap between our proposed research and the overarching goals of both TUF and Stiftelsen Arcada, demonstrating a clear and compelling alignment.
1. Elevating Arcada's Research Profile through Long-Term Applied Technology Research: Our overall research plan is clearly a large, long-term research projects in applied technology. This will raise Arcada's research profile and international exposure.
2. Building International Networks and Securing External Funding: Our research enables us to create international networks around the research projects. We have already demonstrated realistic conditions for other external research funding (Business Finland and direct industrial funding) for a longer time.
3. Direct Contribution to Technical Development and Research in Finland: The primary outcome of our research is the technical development and research in Finland, especially in the field of applied technology. This is the very first stated aim of TUF and of Stiftelsen Arcada. This direct application of research to solve real-world industrial challenges is a core objective of both TUF and Stiftelsen Arcada.
4. Enhancing Technical Training through Hands-On Experience: Our faculty members and students involved in this project will get practical hands-on experience and hence they can better support the technical training. So, the project improves the competence in technical training.
5. Commitment to Open Access and Knowledge Dissemination: In this document, we have explicitly mentioned that we intend to publish our results according to the principles of open access. This document also described how, and to what extent, the project contributes with openly available results, software, and data.
We also discuss the societal relevance, results and impact of the project. We note a clear connection to Arcada's strategic focus areas, which include reliable AI and big data, sustainable health and welfare, business economics in the digital society, transformative culture and media, as well as circular economy and sustainable energy and material solutions. In addition, smart business collaboration, Nordic cooperation, open research and innovation, and innovation ecosystems will continue to be supported.

Abstract

Major industrial organizations extensively utilize computational simulation techniques (digital twins). Traditionally, these digital twins need to be custom developed (hand-crafted) and meticulously maintained at substantial cost and effort. Our research on emerging AI computational techniques (using machine learning and neural networks) potentially offers a low-code/no-code alternative mechanism. Our research hypothesis is that our low-code/no-code methodology can build, develop, deploy and even maintain digital twin simulations at lower costs and within minimal efforts.