Lap simulation

[JP Sultana]

Is there any software that lets you simulate concepts of a car i.e. track/wheelbase, CoG and the usual inputs, for generic typical FSAE circuit layouts?

[Simon Dingle]

http://fsae.com/eve/forums/a/t...20686151#83720686151 (http://fsae.com/eve/forums/a/tpc/f/125607348/m/11120586151?r=83720686151#83720686151)

http://www.optimumg.com/software/optimumlap/

I've not tried it myself but I expect it fits what you're after.

[Max Trenkle]

I swear, have people gotten too lazy to google?

[onemaniac]

Originally posted by Simon Dingle:
http://fsae.com/eve/forums/a/t...20686151#83720686151 (http://fsae.com/eve/forums/a/tpc/f/125607348/m/11120586151?r=83720686151#83720686151)

http://www.optimumg.com/software/optimumlap/

I've not tried it myself but I expect it fits what you're after.

Optimum Lap is a point mass simulator.

[Max Trenkle]

Some might say... point-mass is all you need. http://fsae.com/groupee_common/emoticons/icon_wink.gif

[js10coastr]

Originally posted by Max Trenkle:
I swear, have people gotten too lazy to google?

http://lmgtfy.com/?q=are+people+too+lazy+to+google

[MCoach]

Originally posted by js10coastr:
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Max Trenkle:
I swear, have people gotten too lazy to google?

http://lmgtfy.com/?q=are+people+too+lazy+to+google </div></BLOCKQUOTE>

http://www.giyf.com/

[JP Sultana]

Geez sorry guys!! I've had OptimumLap for ages, im just trying to double check some stuff, and was asking around for any options. Please, do relax the tone as its my first year and if you do this the newbies you'll scare them away from asking anything.. But thanks anyway for the input!

[Max Trenkle]

Lol if you read around on the forums you'll see why we are so skeptical.

Usually it's something like this"

"hi new Formula from ____ first year dnt know were 2 start wat is best engine and suspenshun plz help thx"

[JP Sultana]

Yeah totally understand you! I've seen them around the forums quite a lot! We have quite a good idea of how we're going round stuff at this point and things looking up! But I wanted the different Lap Sims to see if i could do some primitive data collection exercise with my kart, to compare laps and techniques and such...it was more of a personal question to be honest.

[BZ]

Here's one by Lotus Engineering,

http://www.lotusfiles.com/engi...esoft/LVS_insert.pdf (http://www.lotusfiles.com/engineering/lesoft/LVS_insert.pdf)

I've never tried it or talked to any one who has used it but it might be closer to what you're looking for. I know that they offer one copy of the software for free to an FSAE team, anything more and it's at discounted rates.

[EPMPaul]

If you really want to go bananas, it's possible to code one yourself. There is a fairbit of information in here here (https://dspace.lib.cranfield.ac.uk/handle/1826/4791?mode=simple) or here (https://dspace.lib.cranfield.ac.uk/handle/1826/1091?mode=simple). I wouldn;'t recommend redoing the whole numerical bit because frankly mech engineers suck at programming in general(I'm one so I would know). However, it does give a good indication if you want to get one started in MATLAB.

Personally, I've actually coded one and it's somewhat of a bitch(as it stands it's quasi static with no combined grip). The trick is formulating it so it doesn't shoot MATLAB in the face when you run it. This sounds stupid but it took me a fair while to actually get one of these things running to where the results weren't complete horseshit. Of course, you probably won't get the fancy graphs that way but it does give a certain understanding of what's going on in the code

BTW, if you have access to the TTC forum I posted a thread with the bit of code that finds the max speed in a corner which is pretty much the worst part if you want to get started. Having read the two above papers, there's probably a way to expand that using only one variable for longitudinal force and put that in the optimization. As to how exactly, can't say. the problem with combined is that you need to be able to predict what;s coming at the next corner so you're gonna need to solve for the entire course in one shot/function in the optimization routine. Otherwise there's probably a way using a constraint function that stops you from braking and putting the throttle at the same time but the optimization procedure would probably figure that on its own providing you allow a variable brake distribution(otherwise it might put some throttle to keep you on the limit in braking). Also, you'll need a decent engine map seeing as otherwise you can't really modulate and won't have a variable to play with that enables you to not shoot the tires to crap when you floor it.

Last but far from least, you're gonna have to get a decent combined grip tire model. So hours of fun at processing tire data in perspective.

Currently, don't have the time coz we're actually trying to build a car and I'm no longer on an internship in the middle of frickin nowhere with nothing better to do with my evenings. Might get around to getting one setup at some juncture but definitely not this year.

As for commercial codes that do it, I believe that ADAMS/Car has a way of doing it but getting a stable ADAMS model for full vehicle simulation is probably as tedious as the MATLAB way from my experience. Otherwise, chassisSim does one I believe. Seeing as Danny Knowlan posts here regularly he'd probably help out to get you a license. Carsim also seems to have a sponsorship thing going on and a lapsim so you could hit them up too

The only issue with commercial packages is you're going to need to input a shit ton (yes, it s a unit from the SI system... New for 2013) of data and you might end up spending your life getting the data in the first place. So it's gonna be a tradeoff between how much time you spend to what you actually get out of it. If you're already trained on ADAMS, it'll enable you to build a model to exactly the level of complexity you want. only issue is you need to be trained coz it's a pain to use (MSC doesn't seem to know/understand/comprehend/enjoy the term user friendliness).

And finally, once you set tire weight distribution and what i call slip angle modifiers, you can spend days/months year mentally masturbating to set a roll center/camber gains and all that stuff and might end up with the same thing as what you would end up with eye balling it from what papers are already out there(I highly recommend this (http://www.mie.utoronto.ca/undergrad/thesis-catalog/files/201.pdf) ) and a steady state script.

As a conclusion to my excessively long and anecdotal rant, lap time simulation is a very time consuming endeavor any way you cut it and may get in the way of actually getting a well built/reliable car made(again believe me I've done it) and tested(working on starting that these days). So look at it from a resource management point of view before spending a lot of time on it. On the other hand, it does give some interesting insight on what people consider to be important in FSAE. (Ever looked at a graph of Dynamic Index(k^2/ab) vs wheelbase with some compensation for the added wheelbase).... Makes that 60 in wheel base choice ever so interesting).

Now back to my CFD lab

[nowhere fast]

One way to create a lap sim that includes combined longitudinal and lateral accelerations is to first describe the racing line mathematically in terms of inverse corner radius versus distance travelled. The points where the derivative of this is zero, and the inverse corner radius is non-zero, will give the points of the track where the corner is at its tightest, and the car will switch from braking to accelerating (usually).

You can then calculate the corner speed at this point by assuming there will be only pure lateral acceleration. Then you can simulate backwards and forwards away from this point, using the current speed and instantaneous corner radius to determine the lateral acceleration and a friction ellipse/circle to determine how much longitudinal acceleration is available.

The points where the speed simulated forwards away from one corner matches the speed simulated backwards from the next will give you the braking points.

This may not be the most accurate way to do it, but I think it is the easiest step up from a pure longitudinal or lateral acceleration only simulation.

[Joe_Little]

Originally posted by EPMPaul:
MSC doesn't seem to know/understand/comprehend/enjoy the term user friendliness.


Ha, true enough! There are some materials that I and others have contributed to, but its still not enough to make it nearly as easy for you to get started with Adams as it is for other softwares. Its takes a lot of work, more than most teams are willing/able to put forth, to really get up to speed enough for it to be useful. On the other hand, the software and training is free (not just the online stuff, the actual onsite courses for industry are free too if you ask and are on a competition team) and you will make $$$,$$$ when you graduate and take a vehicle dynamics analysis job at a major OEM...

Short of Adams though, CarSim seems like the most popular 'reduced-DOF' vehicle solver in industry, though OptimumG seems like they are giving the most love to FSAE, and VI-Grade makes one called CarRealTime that has some model-sync capabilities with Adams/Car. These are all still dynamics solvers (thought non full blown multi-body dynamics packages) and will still be industry relevant skills, though significantly less so than Adams (speaking generally of course).

The 'lap solvers' are different altogether. There are a few out there, and they're relative simplistic (quasi-statics, traditional equations, and so forth), so simply learning to use any of them is probably not going to help you get you a job, but I'd guess they can provide *some* directional design guidance when used within the proper scope. I have neither first or 2nd hand knowledge which are the best ones, but the roll-your-own approach is fairly doable and a good learning experience that is industry relevant (demonstration of ability to apply concepts to processes, scripting, data processing, all that).

After all, good learning experience is what you're here for, right? I know everyone here doesn't want to be an analysis, but if you do, I'd highly recommend prioritizing your skill development over the low investment/low value software approaches. In a few years you are going to graduate and need the skills to keep doing this stuff.

[Tim.Wright]

Can add a few more;

IPG Carmaker
Tesis Dyna4
Chassissim

The last one ist the most gear towards motorsport and lapsim. The next best one for lapsimulation would be Carmaker. I haven't used Dyna4 but I used it predecessor veDYNA but I found the driver model was not up to scratch for lap simulations.

I think all of them are either free or discounted for FSAE.

Others that I know of but haven't used are;
Ansible Aerolap
Bosch Lapsim
Milliken Research VDMS (though I've never actually heard from anyone who use it. Perhaps Doug Milliken can update us on that?)

Tim

[murpia]

I have used Bosch Lapsim 2009 (free version) with great success. I did not try cars and especially tracks as small as FSAE though, so it may not like those ranges of numbers. (e.g. if it has a distance precision of 1m then that works well for Le Mans, less so for an FSAE track).

A few notes:

<UL TYPE=SQUARE> <LI> You might have trouble finding a download for it, maybe Google cache or the Wayback machine can find it?
<LI> If you have MATLAB you can poke around in the data files it uses to view the internal parameters.
<LI> To import Ascii (csv) track data to create a track map is possible, but if it fails it gives no clues to the reason. I think the documentation PDF lists the format, but be careful about using Excel to generate a csv file because it adds extra commas which screw up the import. [/list]

Persevere and you will have success!

Regards, Ian