Digital twin research has been started for many years, but there is no
standardized tool or method recognized by the academic community. In
this study, system dynamics was proposed as a tool that can solve
integration problems such as multi-scale, multi-physics, and
multi-disciplinary, which are continuously becoming issues in the
digital twin field. By utilizing this, it was revealed that various
heterogeneous data from various protocols or platforms can be integrated
into one model. In addition, the five-step DT construction process was
explained using actual data. Through the simulation results, it was
confirmed that the potential problem caused by the latent effect can be
found, and the method of optimizing it using the results is described in
a rough way, which can contribute to the usability aspect.
In this study, in order to explain the model building process using
system dynamics, the update of variables is omitted (most statistical
model-based modules are implemented with the lookup function). In actual
DT simulation, real data is accumulated and run-time update of root
causes is possible. Details on this can be found in Sücüllü & Yücel
(2014) and Richardson (2015).
In this study, only the service phase was implemented among the 4 phases
of the total lifecycle, but design, manufacturing, and retire phases can
all be added by using the module addition method described in this text.
Since the system dynamics model for each phase of the total life has
been studied a lot, the DT model can be completed by integrating them.