Is this for me? (Eng student looking for sim)

Hello everyone, i'm a Mechanical Engineering undergrad and am in a Shell Eco Marathon team. As such we design a car, and testing its dynamics somewhere would be fun and help us develop it. I've read that i can design the car on Automation and upload it to BeamNG. Is this a suitable software for this venture? If not, am i allowed to ask for alternatives here? Thanks in advance.

 

Yes this is a great game for you, and i highly suggest you give it a try. And since you are an engineering student i would also recommend working for the dev team and help shape the future of this game.

 

Well that escalated quickly, since sometimes it feels like half of us on the forums are mechanical engineering undergrads =P

As for the OP:

You would be interested in this. A team member approximated a track and car in game and it was close to reality. But there are some things you should consider.

Automation is a game about creating production cars, not experimental vehicles. It sometimes doesn't cooperate with fairly normal stuff. Because it is a game where you create vehicles from preset parts (Spore style) you might need to download a mod from the steam workshop or create one yourself to get what you want, at which point you should probably just make a mod for BeamNG itself since it will be far more accurate If you're willing to put the time into learning that, info is available on the offical wiki. Sometimes, automation cars can have their ABS freak out, and tires and brakes don't always cooperate. So beware of that.

As for the game itself, aerodynamics are an approximation but will be good enough for your needs. Tire thermals are not fully implemented; tires are always at ideal temp IIRC.

As for the driving feel, if you're on a university team looking for a sim I assume you have access to a good sim rig. Just remember that BeamNG has not implemented some "sense of speed mechanics" that other games have (though there are some if you look in the in game options). Carefully setting up FFB and FOV will be very important if you want to get a feel for how the car will drive. People who race in IRL autocross say that their times in game are similar with a ported track and their vehicle approximated in automation.

Speaking of sim rigs, here is a funny video. A little extreeme for your needs, but it's funny.

 

Welcome here
I'm also a mechanical engineering student (TU/e) and this game can help with the dynamics if you make your own vehicle. Since most vehicles in the eco marathon(here at least) are not based on any real vehicle, you have to make your own in BeamNG, not in Automation. This is quite some work but if you want, you can simulate almost any chassis and suspension setup (apart from carbon fibre due to the stiffness vs physics speed).
I only mod this game so I can't really say anything about others but I think that Formula Students teams use Assetto Corsa occasionally to train their drivers, not to simulate the cars. Driving in a sim will be different and will mainly help with racing lines, braking points etc. Which you don't really need in the eco marathon, maybe to practice driving longer distances.
FSAE:https://www.reddit.com/r/FSAE/comments/702wcv/racing_simulations_driver_training/
Other simulations can be with MATLAB/Simulink but those take a lot of time and manpower to develop. (Arrays start at 1 lol)

I think that if you develop a car in BeamNG with an accurate weight distribution and geometry that you can get close to the right driving feel. You can also simulate engine thermals and efficiency so it might be handy.
The aero simulation in this game works with (collision)triangles, they can be hard to set up properly if you want to simulate the aero of your car properly.

Good luck, and if you have more questions, shoot me a PM or ask on the forums, people are helpful

 

Welcome aboard I line up with the rest of the engineers here as a senior mech engineer/project engineer.
Id say this is one of the best platforms for your purpose. My somewhat biased view is that no other sim will fill that purpose quite as good as BeamNG as it simulates physics in a quite unique and very correct way compared. Now, theres always the problem with forum trolls which can soil any thread in minutes so keep that in mind if you decide to go for it, we're not all polite engineer nerds

 

Put it this way, if you Love cars. the 100% For You, Welcome A bored beamn brother

 

As a simple answer:

• BeamNG has a very realistic physics engine that is great for simulating vehicles.
• Automation is not a good way to create vehicles for research purposes. They have incredibly simplified node/beam structures and there is no suitable body to create a FSAE car.

If you wish to create your own vehicle in-game it is no easy task, however it is certainly not impossible. Put simply, you need to create a mesh (.dae format) which is more or less mid-poly, and create a jbeam (made of node/beam structures). Although I should also mention a mesh is not entirely essential, as it can be visualised without one.

Here is an example: (visualised with simple jbeam and coltri debug)

(Credit @Blijo)

Here are some useful wiki pages to get an understanding of how 'jbeams' work:
https://wiki.beamng.com/JBeam_Physics_Theory
https://wiki.beamng.com/JBeam_Introduction
There are more pages on the wiki in relation to vehicle creation if you wish to read further:
https://wiki.beamng.com/Vehicle_Creation_Megapage

Powertrains are also simulated in great detail and pretty much everything except tire thermals (which are simulated at optimum temperature) are simulated as of now.

I hope this helps

 

Thanking everyone for the replies! Now, onto the specifics:

Our new car is a CF monocoque What are the implications for what you said here? The load transfers don't work as they should?

I own AC, but i reckon it is not such an accurate simulator. To be fair the pilot does not have to be amazing but i'm sure i could lend them my g27 and put some hours to teach them how to drive, as we usually pick pilots by size/weight and not exactly skill...

We have a digital file (.sldprt at the moment) for the car's monocoque as we used it in fluid sims to calculate its drag coefficient. Will it help to already have such info or would i have to model something from scratch and try to equalize it with our data? All in all, i plan on having some more time this next weekend and will try to gauge the effort vs reward of designing our car inside the game, as we still have to finish it in real life and our competition is probably due september.

 

Well, most of the time when you create something for this game, you want it to deform well. That means you can't oversimplify things like a chassis. This means you need many light nodes with stiff beams. This gives stress with this physics engine (2000hz). To solve this or to work around it, you can simplify your chassis until you have the correct weight and weight distribution, stiffnes and suspension geometry. Because that are the things that matter for you.
So that's also the only things you need to have:
Suspension mounting points, geometry, weight of everything etc. Just like you need for any other sim.
Load transfer will work but again, simplified.
The minimum or most simple form is a car that consist out of two connected boxes with a suspension attached to each box. When you make a more detailed but potentially softer or heavier physics model, you just add more places where your car can deform

As far as effort/reward goes, I think it's a lot of fun to mod and I have some experience with making physics skeletons. I think it is worth it if you are okay with a steep learning curve.

Good luck and if you want the formula car that @CrimsonCrocodile showed, shoot me a PM, it's not a great setup but it works. It might be useful as a base depending on your car.