Towards the Development of an Operational Digital Twin

Abstract

A digital twin is a powerful new concept in computational modelling that aims to produce a one-to-one mapping of a physical structure, operating in a specific context, into the digital domain. The development of a digital twin provides clear benefits in improved predictive performance and in aiding robust decision making for operators and asset managers. One key feature of a digital twin is the ability to improve the predictive performance over time, via improvements of the digital twin. An important secondary function is the ability to inform the user when predictive performance will be poor. If regions of poor performance are identified, the digital twin must offer a course of action for improving its predictive capabilities. In this paper three sources of improvement are investigated; (i) better estimates of the model parameters, (ii) adding/updating a data-based component to model unknown physics, and (iii) the addition of more physics-based modelling into the digital twin. These three courses of actions (along with taking no further action) are investigated through a probabilistic modelling approach, where the confidence of the current digital twin is used to inform when an action is required. In addition to addressing how a digital twin targets improvement in predictive performance, this paper also considers the implications of utilising a digital twin in a control context, particularly when the digital twin identifies poor performance of the underlying modelling assumptions. The framework is applied to a three-storey shear structure, where the objective is to construct a digital twin that predicts the acceleration response at each of the three floors given an unknown (and hence, unmodelled) structural state, caused by a contact nonlinearity between the upper two floors. This is intended to represent a realistic challenge for a digital twin, the case where the physical twin will degrade with age and the digital twin will have to make predictions in the presence of unforeseen physics at the time of the original model development phase.

A Review of Machine Learning Methods Applied to Structural Dynamics and Vibroacoustic

“…Although the surrogate can be a classifier, the target is more often composed of continuous values which predict the physical system response. In front of this, most of the supervised regression methods are appropriate to construct surrogates, like Support Vector Machines [178], Gaussian Process Regressor or Kriging [312,313], Neural Networks [304], Random Forest, Gradient Boosting, etc. As surrogate models rarely involve high dimensions inputs, since the ML inputs are the parameters of the physical problem, the choice of Deep Learning methods is not always as advantageous as in SHM applications.…”

A Survey on AI-Driven Digital Twins in Industry 4.0: Smart Manufacturing and Advanced Robotics

“…In this respect, Kabaldin et al [ 154 ] selected RNN to estimate the statistical model of the dynamic state in cutting. ML models, such as ANN [ 155 ], and probabilistic modeling methods, such as the Gaussian process [ 156 , 157 , 159 ], could likewise be adopted to develop a surrogate model and implemented in a control context [ 157 , 158 , 160 ]. Alternatively, model order reduction techniques can transfer highly detailed and complex simulation models to other domain and life cycle phase, e.g., building efficient finite element model for dynamic structural analysis through reducing the degree of freedom, while maintaining required accuracies and predictability [ 161 , 162 , 163 ].…”

Development of a digital twin operational platform using Python Flask

“…As noted already, in many digital twin systems, there is a requirement to output control signals and other commands to the physical twin. This is also possible in the proposed DTOP framework, and an example is shown in Section A.1 in the Online Appendix (see also Gardner et al, 2020).…”

“…The digital twin concept has been widely studied in recent years, as described in the recent review papers: Fuller et al (2020), Jones et al (2020), Liu et al (2020), Minerva et al (2020), Wagg et al (2020), and Niederer et al (2021). In the context of engineering applications, a digital twin has four main elements: (a) models (both physics- and data-based), (b) data, (c) digital connectivity, and (d) knowledge (both contextual and expert; Gardner et al, 2020). In order to realize a digital twin in practice, an operational platform is required.…”