Full speed on the production line

A recent joint research project between Volkswagen and Siemens

The digital twin is making it easier to create simulation models for complex systems. A recent joint research project between Volkswagen and Siemens shows how this approach improves efficiency in production lines.

A safety door, a rotary indexing table, and a robot are three components in a production facility at Volkswagen where the tailgate for a passenger vehicle is assembled. The robot picks up the tailgate from the rotary indexing table, which first tips into the correct position, and then conveys it to the next station. The closed safety door protects people in the factory from the table and robot movements but must be opened when a new tailgate is placed on the table. The three plant components are controlled and coordinated by software that ensures that the robot and the rotary indexing table will not move if the safety door is open, for example. A software error could cause serious accidents and major material losses. Comprehensive tests – especially those performed at an early stage on a digital twin of the plant – help avoid accidents like this.

Virtual commissioning is standard, but time-consuming

On the VW production lines it’s nothing new: “Before we commission the real production plant, or even individual new components, we always perform what’s known as virtual commissioning,” explains Volkswagen project manager Torben Meyer. “That means we test the functions first using a virtual model of the system, which ensures that the real commissioning, which involves stopping the plant at great cost, proceeds as fast as possible.” Until now, the heads of production have had to budget around six weeks to virtually commission a production line. About two-thirds of that is needed to create the virtual commissioning model in the first place – and that’s a long time. Siemens and Volkswagen therefore set about reducing the time required, as part of the completed EU grant project Maya.  (more about the potential digitalization offers industrial enterprises)

Complex system with a lot of documentation

“A production line like the one at VW is a complex facility consisting of many networked and embedded systems,” says simulation expert Veronika Brandstetter, of the Siemens central research and development unit Corporate Technology (CT). “The individual components may be in-house developments or bought-in products. In most cases, paperwork and development documentation is already available, as well as models that come in different degrees of detail and use a range of notation systems. The people in charge of virtual commissioning face the error-prone and laborious challenge of gathering all this multifarious material from many different co-workers in multiple development departments, and then using it to construct the simulation model of the plant, which is a core element of the digital twin.” One particularly labor-intensive and error-prone stage is reverse engineering, in which the behavior model of existing components must be recreated by working backwards from the original. “The process of creating a virtual commissioning model like this can be made much faster and more efficient if the people in charge are able to use existing and well maintained sources,” Brandstetter observes. 

Vision of a modular system – a technical challenge

The ideal vision would be to make simulation models and other associated information available in the form of a library where the models of the individual components are managed, enabling these elements to be drawn together quickly, like building blocks, to create a complete, virtual system.  “We already have library elements for many system components, especially those we developed in-house,” says Brandstetter. “But this is still an exception for bought-in products.  The goal is to establish the digital testing twin as a standard element in the package.” The challenge still lies in working out how to create a comprehensive virtual system from these digital components, which come from many different sources, with different degrees of detail, and serve different environments. These digital components usually can’t be directly combined with each other.

From multiple individual models to co-simulation

This is where the Siemens experts come in: “It’s especially difficult when the simulation models are based on different, domain-specific tools. For our research project, we concentrated on achieving a solution for this problem,” Brandstetter explains. “For the tailgate assembly, for example, we had to accommodate three separate environments (Process Simulate, Amesim, and PLCSIM Advanced), which simulate or emulate 3D kinematic behavior, electrical behavior, and the control software.

Using the Maya Simulation Framework, which we developed as part of this project, we can now combine models from different simulators. The Maya Simulation Coordinator orchestrates the workflows of the individual simulators to create a single comprehensive system. Comparable approaches were already in place in the area of automobile development, but for the development of manufacturing plant, like our pilot project for VW, it’s a new process.”

Pilot project at Volkswagen

At VW, the control software for tailgate assembly was piloted using co-simulation, as in the example described at the beginning. Workable simulation models already existed for 3D kinematic behavior and electrical behavior of the three components mentioned – the safety door, rotary indexing table, and manufacturing robot – but for differing simulation environments and with features specific to particular disciplines. In the Maya Simulation Framework, they were combined with the control software to form a virtual tailgate assembly system, which they were then able to verify.

Lower costs for virtual commissioning

The users and pilot testers at Volkswagen were impressed by this example. Almost 70 percent of them are convinced that co-simulation will greatly reduce the costs of virtual commissioning. Volkswagen’s Torben Meyer is also happy: “We think that co-simulation offers major potential for reducing the time needed for commissioning in our production lines.”

2019-05-15

Aenne Barnard

Subscribe to our Newsletter

Stay up to date at all times: everything you need to know about electrification, automation, and digitalization.