IoT-based supply chain digital twin
< Project overview >
This project has developed a reference framework for IoT-facilitated improvement on supply chain information asymmetries and therefore inventory decisions. The reference framework has specified the components and key features needed for the development of a supply chain digital twin (SC-DT). The project brings together two complementary academic teams, at Cambridge and Oxford universities, building on and leveraging relevant previous research streams.
The Cambridge team has expertise on building SC automation architectures using sensor-coupled agent software and SC data analytics. The Oxford team has expertise in SC complexity and agent-based modelling with application to supply chains and financial markets.
The academic teams worked closely together throughout the duration of the project. They examined, through reviews of state-of-the art industrial use cases and research papers, and surveys and interviews with company representatives who are using or developing DTs, where IoT and real-time data in a SC can enhance a SC-DT. The team also analysed the industrial partner’s – Schlumberger’s – existing datasets, to suggest how an IoT-based SC-DT can provide improvements in their operation.
The main project goals were to:
Identify how IoT-based data capture can be used alongside SC Information systems.
Develop a SC digital twin reference framework for increased prediction accuracy and optimised coordination amongst SC partners.
Identify and conceptually assess key software technologies that can facilitate DTs.
Evaluate how SC-DT can yield operational improvement.
The key targeted Pitch-In barriers:
(1a) Resource access barriers: data sharing – IP, legal, security, underpinning business models and where companies derive their competitive advantage.
(2a) Lack of trust in stakeholders and real potential for IoT; lack of knowledge/understanding of how analytical tools can be used for leveraging IoT data.
(6c) Lack of understanding in how IoT will/can generate value in a given application domain.
(8a) Incorporation/streamlining of IoT-based applications/decisions with existing business processes.
The difficulty and disincentives to share data is a well-documented and chronic problem in supply chains, which results in inventory wastage due to companies predicting the demand and supply for products inaccurately. The SC-DT reference framework and the additional deliverables and work aim to address an existing gap in terms of SC-DT methodologies, standards, and specific and generic solutions and standards.
There is a lack of understanding among supply chain practitioners on how IoT data can be used to predict disruptions and optimise the supply chains against them. The conceptual SC-DT demonstrator and reference framework aim to showcase university-based research in this area, with a view of encouraging adoption of IoT and IoT-based analytics by SC practitioners.
What was done?
We have performed an extensive literature review on DTs and SC-DTs, designed, tested and refined a SC-DT survey which can be used online to generate a SC-DTs database of key information, based on the SC-DT framework (including generic and specific methodologies and metrics).
We have communicated with Schlumberger, the industrial collaborator, and with several companies, via interviews and focussed discussions. We have also participated, presented and obtained feedback at Pitch-In manufacturing workshops in December 2019 and December 2020. We have also additionally discussed our project and potential links with members of other Pitch-In projects, by interviews focussed on DTs/IoT/SC-DT.
We have evaluated existing SC-DT software, according to a set of key criteria identified through literature review and discussions with practitioners.
The project has established a network of academic and industrial collaborators with an active and significant interest in DTs and SC-DTs. In addition, we have analysed specific datasets from Schlumberger, to suggest how an IoT-based SC-DT can provide improvements in their operation.
Our work has identified and assessed key related work on DTs, and proposed a novel unifying DT reference framework. A paper presenting this work and results has been submitted to the Journal of Industrial Information Integration and is currently being revised. In addition, we have proposed a novel SC-DT reference framework, that builds on and extends existing SC frameworks.
Deliverables and other tangible outputs
Deliverable one. Review of the digital twin concept and vision, and of Digital Twin Building Blocks – whitepaper and academic paper (conceptual DT demonstrator and publication).
Deliverable two. Online SC-DT survey (survey, interviews, analysis, dataset enabler).
Deliverable three. Supply Chain Digital Twin academic paper (conceptual SC-DT demonstrator and publication).
Deliverable four. Review of existing software technologies that enable DT paper (whitepaper; to be used for an additional journal paper, during the dissemination phase).
At the end of the project, we will undertake a range of dissemination activities including conducting a joint workshop with a roundtable discussion on existing SC-DT solutions, on how barriers to adoption can be addressed, and publish a report outlining contribution to Pitch-In objectives and KPIs in the area of IoT for supply chains.
SC-DTs are considered the key enablers of the Fourth Industrial Revolution. This interdisciplinary project has made significant advances towards the theory and practice of SC-DTs, including identifying the state of the art theory and practice, challenges, and open research questions, metrics, application domains, existing software and methodologies, and their features.
The end of project online workshop will be attended by academics and industrial collaborators, leading to a discussion of what a SC-DT means for industrialists and researchers; how different SC-DT viewpoints can be reconciled with multi-echelon supply chain perspectives; what challenges and trade-offs they see in adopting different SC-DT propositions, and what issues need to be addressed to ensure long term success in adopting SC-DT.
The Cambridge and Oxford PIs, together with their team members, will consolidate results, identify research gaps, and use feedback and new industrial collaboration opportunities from the project workshop, to write a joint grant proposal.
In parallel, we will further disseminate the results and lessons from the project and workshop to:
Write a paper for International Journal of Production Economics Special Issue on digital twin and data-driven optimisation for hyperconnected physical internet (submission deadline 30 June 2021).
Publish a second journal paper – finalise and submit Supply Chain Digital Twin journal paper (based on deliverable three; journal to be decided, eg International Journal of Production Research). In addition, the Oxford and Cambridge PIs aim to collaborate as part of a Turing AI World-Leading Researcher Fellowship (proposal submitted in February 2021, Mike Wooldridge (PI), Ani Calinescu (Co-I), Alexandra Brintrup (Collaborator), on the SC-DT-related work packages and deliverables.
The academic teams worked very well together, supporting each other through regular meetings, shared workspaces to ensure smooth and effective feedback, and regular meetings with the industrial collaborator. The team adapted swiftly to the Covid-related changes, and we benefited from generous and competent support, time and feedback from the Cambridge and Oxford Pitch-In managers. We very much appreciated the opportunity to participate in related Pitch-In events and webinars, and to discuss ideas and our work with team members from other Pitch-In mini-projects.
An earlier workshop with participants from industry and academia would have helped to brainstorm ideas and receive earlier feedback. Also, perhaps we could have made better use of the available student internship opportunities. Unfortunately, much time was spent to compensate for the Covid-related disruption and ensure a smooth and effective project progress.
Access to IoT datasets would have helped to further validate the reference framework and software methods (such as machine learning and agent-based modelling), and to enable more case-study specific progress. Also, a broader range of industrial collaborators would have helped to ensure more opportunities for regular meetings and discussion of application-specific features of SC-DT. We hope to be able to achieve this during the follow up project.
Access to deliverables, resources and media content
Sharma, A, Kosasih, E, Zhang, J, Brintrup, A, and Calinescu, A (2020) ‘Digital twins: state of the art theory and practice, challenges, and open research questions’, arXiv preprint arXiv:2011.02833. Available at: https://arxiv.org/pdf/2011.02833.pdf
SC-DT paper/reference framework (please contact the PIs for a copy of the current draft, which is currently being finalised for a journal submission).
DT survey (access to be given by request).
Copies of presentations at workshops/meetings (please contact the PIs for copies of these).
DT software whitepaper (please contact the PIs for a copy).
What has Pitch-In done for you?
Pitch-In represented a great opportunity for interdisciplinary collaboration on the highly relevant SC-DT topic. It has also led to new collaborations with the Pitch-In Cambridge and Oxford managers, to illuminating discussions within and across Pitch-In mini-projects, and with academic and industrial collaborators. We are planning to build on the project results and extend them through a grant proposal.
Anisoara Calinescu – University of Oxford
Alexandra Brintrup – University of Cambridge