The author writes this before acclimatizing to his native weather being just recently back from Italy, where the workshop took place (see the magnificent villa del Grumello on the right).
But why anyone would volunteer to attend a PhD-level event in a field, where they didn't even study?
The simple answer is curiosity; but there is more to it: mechanics is "adjacent" to robotics. Roboticists write codes to control things built by mechanical engineers. Also, these complex things aren't built by trial and error; the software is involved.

A view of the venue's courtyard 5 mins before the morning lecture.

The simulation software helps to reason about a new CAD model — if it's worthy of physical embodiment. When this new thing is supposed undergo accelerations, then the software used must be of the dynamic simulation variety. When someone's dynamic simulation is true to life, it is possible to simulate model's reactions to programmatically commanded controls too. This way the loop with robotics gets closed: the robot can be tested end-to-end now within the simulation, before it is ever built.

Ways to simulate dynamics

I've learned that there are two main ones: multi-body and FAE (finite element analysis). Further, these two are brought together for the third way (co-simulation) sometimes. The latter is used for the most involved problems, e.g.: the fluid dynamics of the atmospheric airflow are simulated jointly with the (partially flexible) structures of an electricity generating windmill.

Without making any futile attempts of explaining them, I will present their demos instead.

Multi-Body Dynamics Simulation
Co-Simulation

Finite Element Analysis

One the first screen, only the rigid bodies (an unconstrained ball and a fixed beam) are being modeled, plus the forces involved: gravity and contact. The next screen shows that beam must not be rigid. In fact, in many practical cases, the elasticity of materials cannot be ignored. Here a case is explored, where short and strong external force is applied to a flexible beam, which bends, springs back and vibrates in response until it has dissipated all the energy transferred into it. Lastly, the separate simulation of two cannot replace co-simulation, when both types of interactions happen simultaneously and generate feedback to one another.

It may be seen that the ball ends up in completely different configuration on the third screen due to the beam's bendiness even though the initial conditions for the ball and the rest of the environment is fully identical to the one from the first screen.

My examples are a bit simpler than the wind turbine enmeshed in fluid, but these should nonetheless get the point across. If readers are interested to inspect further, these renders were made with a single Python script that's only a couple of hundreds lines long. The script is available via the link.

In writing this, we also skipped over the simplest simulation possible — the kinematic one; it would differ from those above by way the ball's trajectory is realized: instead of experiencing the gravitational pull, the ball would just go in a straight line accordingly the commanded actuation, and would never interact with the beam.

So, if an outsider can get cracking with writing co-simulation codes after just a one week-long acquaintance with the topic, is there any sense in going to this particular MechE workshop in the future? I believe there is.

Summer School Take-Aways

I will not share authors' materials, while the school's program is available on their official page. The structure of it was so that attendees were presented with three types of classes:

• Pure theory. This year we've learned a lot about peculiarities of kinematic approaches, handling SO3 and modeling beams. The last part was useful in making my toy simulations!
• Applied lectures. These presented real mechanical engineering problems and their solutions from fields such as rail transportation, wind energy, hydraulics.
• Hands-on sessions. Attendees got their student copies of the industry-grade simulation software (ADAMS). The most meticulous of us even put them to work simulating motor vibrations on complex electricity powered machine.

Then, there were talks, walks, chats, exchanges of ideas and aspirations, espressos, the Como Lake, and the roar of engines during the lecture hours (from the beautiful birds below).