Coupled Simulation

Today, the coupling of simulation software from several different physical domains is essential to adequately map and simulate modern technical systems.

This fact especially applies to vehicle systems: modern vehicles are complex mechatronic systems, where electrical and hydraulic components interact with complex control algorithms and the classic mechanics of the vehicle. A vehicle is controlled and operated by a driver and its tires possibly interact with a yielding surface.

Co-Simulation
© Photo ITWM

Co-Simulation

The adequate mathematical modeling of all these components and sub-systems generally requires domain-specific, problem-based modeling and simulation software. Typically, this involves the creation of mathematic equations of various types and requires different numerical solvers. In addition, the models usually run in significantly different timescales – depending on the physics, but also on the modeling approach (micro-/macro level).

Integrated system models simulate all sub-systems and components

In an integrated system model, all sub-systems and components – from various domains, in different software tools, in different mathematical descriptions with different numerical solvers – must be coupled for simulation.
Initially, this is a major technical challenge for the software; the simulation tools must have the capability for data exchange during operation.

Most modern tools support this requirement, for example, if they are supported by the FMI standard, a tool-independent interface standard for co-simulation and model exchange. (see www.fmi-standard.org)

Furthermore, the numerical coupling scheme is of critical importance: This is what defines when and what sample data are exchanged; in a parallel simulation of separate coupling partners, the exchange variables have to be appropriately predicated and extrapolated.

We deal with the development of numerical methods that implement the various coupling schemata to ensure an efficient and stable co-simulation. In this context, Fraunhofer ITWM is developing co-simulation master software and deploying it to client projects where a coupled simulation can be structured and administered. By the way, simulation partners that support the FMI standard can also be integrated.

Distributed co-simulation and co-simulation in real-time context

To achieve an efficient co-simulation, it is frequently essential to work on different computing units. These could be modern, multi-core computing systems with shared memory or separate distributed processors, that communicate with each other via a network connection.

Distributed simulation strategies are particularly important for coupled simulation under real-time conditions. In this context, not only must every simulation partner be real-time capable, the data exchange and extrapolation must be performed efficiently enough or must be adapted to achieve stable, coupled real-time simulation.

Our co-simulation master software is developed with special focus on these real-time applications. The typical application areas are HiL/SiL tests, but also the test bench area as well as interative driving simulators. At our driving simulator RODOS, we have already implemented a distributed, coupled real-time simulation: The coupling partners were an MBS vehicle model, four tire models, the simulator and a human operator.

Gekoppelte Simulation
© Photo ITWM

Coupled real time simulation