by Gerardo Santillán Martínez (Aalto University), Tommi Karhela (Aalto University), Reino Ruusu (Semantum Ltd), and Juha Kortelainen (VTT)

Simulation-based digital twins (SBDTs) of process plants can be used for a number of important industrial applications. They have various advantages compared to digital twins based on data-driven models. However, wider industrial adoption of SBDTs is hindered by laborious development of their underlying simulation model as well as by the lack of integration methods with the operational process. The Engineering Rulez research project has tackled these issues by developing a novel automatic model generation method as well as a simulation architecture based on OPC UA, a well-established industrial interoperability standard. The proposed SBDT automatic generation method aims to enable a wider industrial adoption of digital twins based on first principle models.

Ever-growing competitiveness in process industry pushes companies to increasingly rely on Industrial IoT solutions for improving operation performance and for increasing cost efficiency of process plants. Digital twins (DTs) of production plants are an example of IoT applications, which are becoming highly popular in process industry in sectors such as chemical, power generation, mineral processing, pulp & paper, and oil & gas. Since they are able to capture the structure and dynamics of the targeted plant, digital twins are a powerful tool that can be used for optimisation and for decision support of operational process plants.

Commercial DTs, commonly based on data-driven models, are developed purely from the available measured data of the targeted industrial plant. These systems rely on black-box models built to capture relations between the inputs and outputs of the plant. Consequently, they are fast to develop and they can be applied to obtain production forecasts or to detect certain production anomalies. However, since they are only based on measured plant information, their results cannot be used to analyse plant operation states that are not included in the collected data. Additionally, they require expert interpretation and are thus difficult to scale up. Moreover, applications based on data-driven DTs rely entirely on the automation and monitoring systems data to provide information about the current plant state.

Simulation-based DTs (SBDTs) are based on on-line simulation of first-principles models (FPMs). FPMs rely on engineering, physics or chemical knowledge to represent the behaviour of the plant. In SBDTs, a simulation model runs together with the plant while online and off-line estimation techniques synchronise the simulation state with the state of the targeted device or process. Information of the current state of the plant can be obtained from this simulation configuration. The underlying simulation model can be used to obtain high-fidelity predictions, including production forecasts of operating regions from which no measurement data are available. Furthermore, SBDTs can be used for developing operator training simulation systems, for production optimisation, or for troubleshooting and failure diagnoses.

Figure 1. Simulation-based Digital Twin (SBDT) structure and its applications.
Figure 1. Simulation-based Digital Twin (SBDT) structure and its applications.

SBDTs are a holistic tool for plant operation support of modern industrial plants. As such, developing the FPMs of SBDTs is a time-consuming and complex task. Although these issues can be partially solved by re-using existing models, developing FPMs remains laborious and expensive. Moreover, the lack of systematic approaches for SBDT generation, which address complex integration of the process with simulation systems and methods, limit wider industrial adoption of SBDTs.

The Engineering Rulez research project has aimed to develop an automatic generation method of SBDTs for industrial process plants, which addresses the presented shortcomings in order to increase industrial adoption of SBDTs. In the proposed approach, laborious FPM development is tackled by applying automatic model generation (AMG) methods.

Existing AMG methods utilise data from engineering sources, such as piping and instrumentation diagrams (P&ID), equipment technical data sheets and control application programs. However, it is not possible to generate high-fidelity dynamic thermal-hydraulic FPMs without information of the process pipeline network. In particular, key parameters for such FPMs are the head loss coefficients, which represent head losses due to elbows or branches in the pipelines. These parameters can be obtained only from information about the physical piping structure and are thus available only after a 3D pipe routing has been accomplished.

For this reason, our approach uses information available from 3D computer-aided design (CAD) models of the plant in combination with other engineering data for rapid development of high-fidelity thermal-hydraulic simulation models [1]. In the proposed AMG method, data included in the 3D plant model is used for calculating piping sections lengths, elevations as well as head loss coefficients of the pipeline network, and to automatically generate a thermal-hydraulic model. As a result, the fidelity of the simulation model is increased compared to the one obtained following existing methods.

Figure 2. Automatic SBDT generation from 3D plant models.Figure 2. Automatic SBDT generation from 3D plant models.
After the first-principles simulation model is automatically generated, a newly developed lifecycle-wide online simulation architecture [2] is utilised to automate the generation of the SBDT. This architecture is used to automate the process of connecting the FPMs to the physical plant; to optimise the simulation model for its behaviour to closely mimic the real process; and to dynamically adjust the simulation results in order to permanently synchronise the simulated and real plant states. Furthermore, the developed architecture utilises the industrial interoperability standard, OPC UA, to avoid the need for point-to-point integration of various simulation instances and methods used over the course of the SBDT lifecycle.
DTs are the cornerstone of the industrial digital transformation. The implementation framework proposed by the Engineering Rulez research project aims to enable a more efficient path for the implementation of SBDTs.

[1] G. Santillán Martínez, et al.: “Automatic Generation of a High-Fidelity Dynamic Thermal-hydraulic Process Simulation Model from a 3D Plant Model,” IEEE Access, pp. 1–1, 2018.
[2] G. Santillán Martínez, T. Karhela, R. Ruusu, S. Sierla, and V. Vyatkin: “An Integrated Implementation Methodology of a Lifecycle-wide Tracking Simulation Architecture,” IEEE Access, vol. 6, pp. 15391–15407, 2018.

Please contact:
Gerardo Santillán Martínez, Tommi Karhela, Aalto University, Finland
This email address is being protected from spambots. You need JavaScript enabled to view it., This email address is being protected from spambots. You need JavaScript enabled to view it.

Reino Ruusu
Semantum Ltd., Finland
This email address is being protected from spambots. You need JavaScript enabled to view it.

Juha Kortelainen
VTT Technical Research Centre of Finland Ltd
This email address is being protected from spambots. You need JavaScript enabled to view it.

Next issue: July 2023
Special theme:
"Eplainable AI"
Call for the next issue
Image ERCIM News 115 epub
This issue in ePub format

Get the latest issue to your desktop
RSS Feed
Cookies user preferences
We use cookies to ensure you to get the best experience on our website. If you decline the use of cookies, this website may not function as expected.
Accept all
Decline all
Read more
Tools used to analyze the data to measure the effectiveness of a website and to understand how it works.
Google Analytics
Set of techniques which have for object the commercial strategy and in particular the market study.
DoubleClick/Google Marketing