A first experience with Simpack in real-time – the view of an analyst
I still vividly remember the first time I experienced our real-time simulation technology. It was around 2012, and I was working on a customer project involving a vehicle model with a custom Pacejka-like tire implementation to meet specific requirements. After weeks of refining every subsystem and coding the tire equations as a Simpack user routine, I was finally satisfied with the result.
Since the customer used a driving simulator and we had just released a new generation of real-time technology, we agreed to demo the vehicle model on a basic desktop simulator. We set up a simple test rig: a gaming-style steering wheel and pedals with torque feedback, running on a Windows laptop converted to real-time Linux.
Then came the critical moment. I loaded the model, sat down (in an office chair) and turned the wheel. After countless hours of tuning equations and analyzing 2D plots, feeling the car respond to my inputs was a game-changer. As I accelerated, braked and steered, every reaction reflected the physics I had built. My work had come to life—it was no longer just numbers; it was physically engaging with me in real time. I was experiencing my model. The experience struck me so deeply that I instinctively took a picture—not of the setup, but of what it represented: the first time I ever drove my own multibody model in real time. Here it is🙂:
Simpack real-time multibody dynamics
It would be beyond the truth to say that Simpack “invented” real-time multibody dynamics, but factually, the software was the first to implement this technology at an industrial level, and Simpack has led the way by making key innovations.
In a SIMULIA blog post published in April, I discussed what makes Simpack’s technology unique, particularly how it formulates the equations of motion. One key advantage I described is Simpack’s relative coordinate approach, which naturally leads to a numerically efficient and stable model, making it inherently well-suited for real-time simulation applications.
This advantage comes from Simpack’s tendency to produce ordinary differential equations (ODEs) rather than the more complex differential-algebraic equations (DAEs) typically generated by absolute coordinate formulations.
Actually, DAEs arise with classical “absolute coordinates” formulations, and the problem is that solving these equations increases the computational complexity. These DAEs lead to higher solver overhead and reduced numerical robustness, both of which are problematic for real-time applications that require deterministic and fast execution.
On the contrary, Simpack’s relative coordinate formulation defines the motion of bodies with respect to their parent bodies using a tree-like topology. As a result, the equations of motion are predominantly ODEs, which are simpler to solve numerically and exhibit better performance characteristics for explicit or semi-implicit integration methods commonly used in real-time simulation.
Simpack does support the creation of DAE-based models when needed, but its unique formulation also allows for the generation of pure ODE models, something not achievable with many other multibody simulation technologies.
The advantages of ODE-based systems for real-time applications include:
- Lower computational load per integration step
- No need for constraint stabilization
- Improved numerical conditioning
- Simpler solver structures, which are more amenable to fixed-step solvers often used in hardware-in-the-loop (HiL) environments.
Simpack Code Export | early applications
Because Simpack’s solver technology is so well adapted to real-time, running Simpack on a driving simulator was always a possibility. We started promoting this technology in 2001. Back then, our technology was named “Code Export” as shown in this picture from our old website.
Code Export was our first generation of real-time technology and was based on code export generation. With this approach, the multibody model was symbolically pre-processed and compiled into target-specific C code that could be deployed on dedicated real-time platforms. This method ensured high execution speed and determinism but required a separate code generation workflow, which increased integration complexity and reduced flexibility.
Simpack Realtime | our modern approach to real-time multibody simulation
Simpack Code Export was used in the field for about 10 years, giving us deep insights into many challenges associated with real-time applications. One key learning was that “just” generating real-time code was not sufficient to satisfy customers’ needs.
The second generation, introduced around 2013, marked a significant shift. Unlike code export approaches that required proprietary hardware and operating systems, Simpack Realtime was designed to run a standard Simpack installation on standard real-time-enabled Linux platforms. This was made possible by the new Simpack Realtime solver technology, which included “smart parallelization”, combined with advances in real-time Linux kernels, memory locking, and deterministic scheduling.
The idea of running a complete “off the shelf” multibody solver in real time was a completely new idea in the industry.
So what makes Simpack Realtime groundbreaking?
It’s fast, flexible and built for modern engineering workflows. It enables running multibody simulation models directly in real time from within the Simpack environment. There is no need for lengthy code generation or tedious compilation steps—just load your model and hit run. We call this direct real-time.
Our direct real-time approach has consistently delivered significant value to our users. By leveraging the full capabilities of the Simpack solver without limitations, engineers can run their models on their real-time simulator without compromising on model fidelity or performance.
Another key advantage of this approach is that model simplification and solver tuning are no longer necessary.
One powerful example of this flexibility is that Simpack Realtime can be combined with Software in the loop (SIL) by importing Functional Mock-up Units (FMUs) seamlessly using the Simpack FMU Import feature. This means you can integrate third-party components such as power steering or battery management modules and still achieve reliable real-time performance (indeed, you must ensure that the third-party components are real-time compatible).
On top of that, you can easily modify Simpack model parameters between runs—and for certain parameters, even adjust them on the fly while the real-time simulation is running. This enables rapid iteration and interactive testing like never before.
As one of our customers worded it, “We really love your workflow because it is transparent”.
As an example, in the automotive industry, engineers have relied on lookup tables to approximate suspension kinematics in real-time simulations—a method, while effective, introduces some intermediate data processing steps. With Simpack Realtime, those approximations become obsolete. You can now run fully detailed suspension models, including dynamic bushing elements in real-time, eliminating the need for precomputed data and enabling a more accurate, physics-based simulation—directly and without compromise. Over the years, we’ve also ensured that third-party tire model integration is fully compatible with real-time execution across all major tire models.
It has been demonstrated that Simpack Realtime can seamlessly work with models that include flexible bodies. Here is a paper published by one of our customers: Employing Real-Time Multibody Simulation in Driving Dynamics Development by Venrooij et al.
In my experience, flexible body integration significantly enhances the driving experience for driving simulators.
It is worth noting that beyond the automotive sector, Simpack Realtime is widely used across a diverse range of industries, including rail, wind energy, drivetrain systems, construction and even consumer products. As an example, you can watch this webinar discussing Simpack Realtime for wind turbines.
Today, Simpack Realtime enables you to deploy the same model across multiple hardware configurations, seamlessly covering every stage of the real-time development process from early standalone runs on desktop simulators or real-time hardware, all the way to full deployment on high-end motion platforms. If the standard platform is not sufficient, it is even possible for customers to code their own custom communication layer.
And it doesn’t stop there, with integrated real-time animation feature, you can visualize your simulations as they run.
We hope this post has shed light on why Simpack has become a trusted leader in high-performance real-time simulation, empowering engineers to push the limits of what’s possible. But this is only part of the story, innovation is at the core of everything we do and we’re excited about the future.
Stay tuned—we’ll share more exciting developments in future blog posts. Are you ready for real-time multibody dynamics simulation?
Tom Burton is a SIMULIA Industry Process Senior Consultant.