Abstract
The manufacturing industry is embracing a digital transformation from traditional automated manufacturing to smart manufacturing for improving the agility, efficiency, safetyand sustainability. Digital twin (DT)-driven smart manufacturing has shown the potential
to change the today of manufacturing and reshape its future.
Over the past decade, DT technology has attracted tons of attentions, regarding its
definitions, conceptual models, theoretical frameworks, and practical applications. The
creation of a DT typically involves sampling multi-sources data and developing ultra-high
fidelity models supported by sufficient computing resources. However, the lack of strategic
guidelines for data sampling or model development may result in excessive representation
of their physical counterparts, thereby leading to unnecessary costs and efforts. Even
though companies accepted the need to invest a significant amount of cost in developing a
DT due to the benefits it could bring in the early stage, cost, like in any business, remain
the principal barrier when implementing any new technology. This raises a basic research
motivation: how to improve the creation of a DT for its broader application?
This thesis proposes an information and cost modelling approach and formulates it as
a Digital Twin Cost and Information Modelling (DT-CIM) framework. The information
modelling methodology identifies a set of information entities by graphically and mathematically representing the hierarchical workflow. An attribute, termed purpose-based
weight is introduced to quantify the relationship between each information entity and the
specific purpose of a DT. A cost model is developed by identifying a set of cost elements
in a DT implementation, which are then quantified to provide a cost range as reference. A
characterised information-cost space is finally constructed to explore the trade off between
capability and cost, which refer to metrics derived from information and cost models, respectively. This space intuitively demonstrates the cost variance, including bounds as well
as distribution across various combinations of process signatures and individual impact of
each information entity on the implementation cost.
The case study of the material extrusion (MEX) manufacturing process shows observable contributions to improving the creation of DTs. The selection scope is defined by
37 process signatures among 129 information entities in the MEX process information
model. The selection criteria is defined by the cost estimation for the specific purpose
of predicting the dimensional accuracy of the printed part. Rather than straightforward
selection, additional insights derived from the characterised information-cost space can
support a cost-informed selection and finally improve both the certainty and rationality
of the creation of a DT.
| Date of Award | 9 Dec 2025 |
|---|---|
| Original language | English |
| Awarding Institution |
|
| Supervisor | Aydin Nassehi (Supervisor) & Ben J Hicks (Supervisor) |