Digital Twins Speed Up Drive Development

From Simulation to Reality: Successful Virtual Release

Digital twins, modern simulation methods, virtual homologation: development and validation processes in the automotive industry are changing from the ground up in the course of digitalization. This year's VDI Congress DRITEV 2025, which will take place on July 9 and 10, 2025 in Baden-Baden, is dedicated to these topics in detail. René Honcak, ZF Friedrichshafen, will report in his presentation on how virtual prototypes and digital twins can reduce development costs and thus enable decisive competitive advantages. We spoke to him about the associated potential and the specific benefits of these technologies for OEMs and suppliers.
 

Mr. Honcak, one focus of your presentation is on the combination of different technologies. What does your model look like and what role do Model-Based Systems Engineering (MBSE), TwinOps and digital twins play?

René Honcak: Our methodology combines MBSE, TwinOps, IOT and digital twins in a holistic approach that makes optimal use of virtual prototypes. MBSE forms our basis for systematically assigning customer requirements to technical solutions, designs and test specifications for both the product and the digital twin. This methodology helps us to make complexity manageable, especially in the development of multi-physical digital twins, which are both digital and virtual product images. MBSE also promotes efficient variant management as well as the traceability and reusability of models and parameters. By linking requirements management and task management, we plan our resources and simulation time much more precisely.

TwinOps is the transfer of the DevOps methodology from software development to continuous development, integration and testing as well as the development of a virtual product representation. This creates our cloud-based environment to automate simulation workflows, continuously integrate simulation models and test for changes in an agile manner. TwinOps enables us to work collaboratively, with experts from different disciplines accessing a shared, standardized environment. We also use TwinOps to calibrate and continuously improve machine learning and simulation models using a close-loop approach. This ensures that our models always meet the latest requirements. Digital twins ultimately form the virtual image of our real systems. They enable rapid analyses and help us to efficiently evaluate design and functional decisions without having to constantly rely on physical prototypes.
 

Which combination of technologies has proven to be suitable for your purposes?

René Honcak: The use of MBSE and execution in a TwinOps environment for the agile and scalable integration of simulation models and parameterization of models gives us a decisive advantage both in development and in discussions with our customers. This creates a strong basis of trust. In addition, thanks to the digital twin technology and the connection of our test benches as IoT devices, we are able to verify and validate models in a traceable manner. MBSE also offers us the possibility of risk management with detailed transparency regarding model uncertainties or the impact of design or functional changes on the simulation quality.
 

You emphasize the importance of trust in virtual models. How do you ensure data quality and reproducibility?

René Honcak: We ensure data quality with comprehensive validation processes. This means that we carry out sensitivity and uncertainty analyses for simulation results to evaluate the simulation forecast quality. We use statistical tests to evaluate both measurement data and simulation data. This is supplemented by sensitivity analyses in order to understand the factors influencing simulation results and to continuously optimize the models. Based on our MBSE methodology, we use standardized test cases with corresponding metrics to evaluate simulation models. Our Digital Twin Ecosystem also has automated pipelines that systematically cleanse and validate data. In this way, we create a reliable basis for our models and simulations.
 

How important is interdisciplinary collaboration in this approach?

René Honcak: Cross-disciplinary collaboration, for example between hardware, software and simulation teams, is essential for us to successfully implement complex development projects. We rely on systems engineering and model-based systems engineering (MBSE) as a collaborative platform to efficiently exchange information and create a common database for all participants. 
 

What advantage does your modeling V-model offer compared to classic V-models?

René Honcak: The Digital Twin V-model developed by us plays a central role in the virtualization of our product in the sense of a digital twin, as it addresses the high degree of complexity and variety of our simulation models and application variables. Our V-model enables efficient and error-free management and the merging of different simulation models and results by defining requirements for simulation models and the test cases for calibration, verification and validation. Using SysML architectures, it describes simulation workflows and models on a meta-level, which ensures detailed traceability and precise planning of activities and provides the ideal starting point for detailed model development and simulation execution.
 

How do you manage risks in the course of virtual release?

René Honcak: We control risk management with regard to the accuracy and reliability of the virtual release using a structured, multi-stage approach. First, we define clear requirements for the simulation models and the underlying measurement data used for calibration, verification and validation. These requirements are documented using SysML architectures in order to make simulation workflows and models comprehensible at a meta-level. To ensure model quality, standardized tests, sensitivity analyses and uncertainty analyses are carried out to evaluate the accuracy of the simulation results and detect potential deviations at an early stage.
 

Which specific processes are currently benefiting from your technology in particular?

René Honcak: Currently, especially e-drive development and associated processes are benefiting greatly from virtual approval using digital twins. This has enabled us to significantly reduce physical validation efforts, which makes both product development and the application more efficient and is accepted by our customers. In addition, the overview of dependencies of simulation results provides a clear basis for avoiding unnecessary iterations and optimizing the development process. 

Another success is the virtualization of thermal aging tests for power electronics, which has increased efficiency by up to 80%. We also benefit from the TwinOps infrastructure in the area of virtual sensor technology, such as thermal detection, which enables a CI/CD/CT system for virtual sensors through the continuous provision of simulation and measurement data, which can then be integrated into the vehicle software with ECU capability.
 

What are the advantages for ZF and your customers?

René Honcak: Virtual release using digital twins offers significant benefits for both ZF and our customers in terms of product development time, cost efficiency and competitiveness. By reducing physical prototypes and moving testing to the virtual environment, development cycles can be significantly shortened – for example from 24 to 18 months or less – enabling faster time-to-market and meeting our customers' requirements. At the same time, virtualization significantly reduces costs, as fewer physical prototypes are required and complex test setups are no longer necessary. This leads to a more efficient use of resources and greater cost-effectiveness. 

In addition, virtual release improves product quality, as digital twins enable more precise validation and optimization in early development phases. This reduces the risk of errors and increases product reliability, which has a positive impact on customer satisfaction. The ability to reuse simulation results and models through generic planning minimizes duplication of work and reduces development costs in the long term. Overall, this approach strengthens competitiveness by bringing innovative solutions to market faster, more cost-effectively and with higher quality.
 

What are the current limits and what progress do you expect to see in the coming years?

René Honcak: The current challenge lies in particular in the high technological volatility in the area of powertrains. Over the next few years, we expect major leaps in development through the increased integration of AI methods into our development process. In particular, the use of AI for faster analysis of requirements with regard to simulation capability will enable us to further shorten development times and significantly increase efficiency once again.