Optimizing Your Data Capture for Digital Twin Creation in Mechatronic Systems – by Flanders Make

Flanders Make

Digital twins are virtual replicas of physical objects or processes, and they can be used to improve the design, optimization, monitoring, and validation of systems both during development and operation. An important aspect of building a digital twin is capturing the data it requires. Here, we will discuss the data capture from a fleet of machinery that can be used to create digital twins thereof.

The physical setup

The approach focuses on capturing measured timeseries data of operational mechatronic systems and the context to use these timeseries when creating a digital twin. It uses a research drivetrain fleet as reference to apply the data capturing for digital twin creation in practice. The research powertrain consists of different components: power electronics for the inverter, a permanent magnet synchronous machine, a gearbox and a load motor, connected with a timing belt, to transfer the required load on the system.

Several signals are being measured: voltages and currents to capture the electrical behaviour, torques and speeds to capture the mechanical behaviour and temperatures to monitor the thermal states of the components and the environment. In the drivetrain fleet, the right sensors are integrated and measure the signals with the desired accuracy. The data acquisition system (DAQ) is able to acquire the data of all signals at the desired sample frequency.

To facilitate the usage of the captured data, two main components are foreseen in the data capturing approach: a data storage and management solution and a user interface.

The data storage solution

The data storage contains different formats of data (curated timeseries and metadata regarding the test and system). It is based on a data model that defines the data content that must be captured and the relationships between the different data types.

In the application, different data sources are combined: the measured timeseries data, captured via sensors mounted on the machine during operation; documentation, stored in separate files (such as user manuals of specific components, specification sheets and calibration information of the sensors); and a metamodel (model of metadata), which contains information about the test campaigns and the specific system configuration that underwent these tests and links the different sources of data together.

The metamodel contains domain knowledge and provides the links to the separate data sources, making these more valuable. It is the core of the data model and is used to improve the data search in the user interface.

The user interface

When the data storage is in place, a user interface can be developed that connects to it. The user interface is able to search and download the data that is required to build the digital twin. In this case, a web-based interface was used that allows for searching and filtering the different data entities (tests, signals, components, etc.) based on the metadata defined in the metamodel. It uses the data model structure to query the right type of data for the user to conveniently explore all available data.

 The user interface also makes it possible to upload data. In order to upload the metadata, it makes use of an Excel file which is then converted to the appropriate SQL commands.

Digital twin creation

By obtaining the required data from the data storage, the digital twin can be created. Note that the download of measured signals by the user interface takes care of the conversion from the time series format to the required format to build the digital twin.

The digital twin consists of a set of mathematical models that represent the behaviour of the physical system. These models are trained, based on the data obtained from the data storage and can be used for simulation, design optimization, improved monitoring or faster validation, both during development and operation of the system. In this case, the digital twin is used to assess the energy consumption of the system with an accuracy of less than 10%.

The digital twin allows for simulating different scenarios, such as changes in the control logic, the addition of new components or changes in the operating conditions of the system.

Digital Twin Academy

This use case was investigated within the Digital Twin Academy project. This project is being carried out within the context of Interreg V-A Euregio Meuse-Rhine, with subsidy from the European Regional Development Fund.

Flanders Make


Others technology sheets

This website uses cookies in order to offer you optimal navigation and to provide us with statistics without knowing your personal data. By continuing to browse this site, you agree to the use of cookies, in accordance with our cookie policy.