‘Can you make me a machine that produces 250 thousand pouches of dog food per week?’ Where do you even start when a customer gives you such a challenge? You will probably begin with some rough sketches on paper. But to make it more tangible and concrete, you’ll soon want a simulation tool to draw up your ideas quickly and validate them. This is a challenge that we can tackle by using five levels of simulation.
The highest conceptual level is known as discrete event simulation (DES). In a RollerCoaster Tycoon-type of setting, you quickly construct your first design draft in 3D, preferably with ready-to-use building blocks. This will immediately show you whether or not the devised machine fits within the physical limitations of the factory floor.
But a DES tool can also take you much further. Discrete event simulations revolve around the timing of all the actions and operations in your process. For instance, it will take 2 seconds for an empty pouch to reach the filling station. It has to wait 4 seconds to be filled up, then another 3 seconds to make it to the sealing module. The simulation shows if the flow of the process is valid, and it uncovers potential bottlenecks. In this example, the process may be straightforward and can easily be evaluated mentally, but you can imagine that when you design the logistics for a big warehouse or an intricate production line, it quickly becomes far too complex to keep track of all process steps.
What a discrete event simulation lacks are translations and (virtual) movements. In the visualization of your process, the empty pouch will just pop up at the filling station, 2 seconds after you press ‘start’. With our Prespective DES tooling, you can easily introduce motions with the addition of splines. By restricting movements to a spline, the possible paths are fixed, thereby optimizing the calculations it takes to generate a high-resolution simulation.
Note that, although we use double precision instead of float variables, a DES-with-splines solution does not simulate all the physics in deep detail. It will, however, give you invaluable insights into the performance of your design. The beauty of these two simulation levels is that you have the power to speed up time. Because you are still working on a conceptual level, the math isn’t too complicated yet. So when you show your ideas to the customer, you can first run the visualization in real-time to get a feeling of the process. Then, you can speed up – a thousandfold if necessary – and show a week’s worth of output in mere seconds. You can even simulate the production capacity of a full year with log files and optimize for throughput.
Deepen with a physics engine
After you have checked and agreed on your design on a conceptual level, you probably want to move into the details and actually validate if the system will work the way you designed it. Maybe you need to investigate a possible bottleneck or want to elaborate on some key actions of the process. That means it’s time to get the CAD model from the engineering department, or make it yourself, and set up the exact behavior of the system. Now, it’s time to dive into the kinematics. In Prespective, you can include the control software within the simulation. With every addition to your model, you’re improving the accuracy of your virtual prototype until it becomes a full-fledged Digital Twin. This third level of simulation is based on physics engines. The Prespective platform is built on the powerful Unity 3D game engine.
The closely linked, but one step deeper, fourth level of simulation comes with the inclusion of FACT (Field for Accurate Collision Tracking). This module is part of our own physics engine PASS (Prespective Accurate Simulation System). It allows you to designate critical regions in your design, where collisions will be checked in high resolution, with a 100% accuracy.
The fifth and final level of simulation can only be done with specialized tooling such as Matlab, Comsol or Wolfram. These calculations are too specific and too complex for standard real-time physics engines. However, Prespective has an interface to include the results of these advanced simulations into your model, after which they will run in real-time with the rest of the simulation.
Note that with the addition of physics engines, you lose the ability to speed up your simulation. These mathematics do not run faster than real-time. The real gain is that the physics, collisions and kinematics are much more accurate – enabling you to validate your system.
This toolset is designed with continuous improvement strategies in mind (i.e. Six Sigma). The first two levels of simulation (DES and Spline solutions) enable you to start at the very beginning, testing out the first ideas, working up towards the first results, in terms of throughput and production planning.
Improving on this brings you into the third level of simulation – with a physics engine included and running on the actual control software. You can actually test and validate if the system works and behaves as it should. The fourth level of simulation brings in an extra superpower. Where standard physics engines run out of their maximum capabilities and are unable to continue producing reliable results (i.e. very small objects interacting with large objects), our FACT system can take over. FACT ensures 100% accurate collision detection on any scale, and does it at real-time. For specific needs, the fifth level of simulation (i.e. Matlab, Comsol, Wolfram results) can further detail the virtual prototype.
In contrast with competitors, we have managed to capture the complete simulation track, from DES to specialized tools, all into one platform: Prespective. You don’t have to translate your DES model for the physics engine you want to use. They automatically understand each other and you can easily start at the highest conceptual level and expand all parts until you have a fully operational Digital Twin.
About the author: Guido van Gageldonk is co-founder & CTO of Unit040, a visualisation and simulation software company from the Netherlands. He founded the company in 2006 while still being a student at the Technical University in Eindhoven. He is a renowned tech watcher in the Netherlands and is well known for his ground breaking ideas regarding Digital Twin Technology.