A new free vehicle dynamics resource - Dan's Vehicle Dynamics Corner

[DougMilliken]

I think Bill had something like this in mind, but he's probably seen worse(?). I took off the identifying information, this is from an actual aligning torque compliance test, with the power steering operating. A degree (+/-) of steer hysteresis on center is pretty awful, more on the left front (dotted line) than the right.

This figure appears near the end of my lecture on Data for Vehicle Modeling, after a number of good looking K&C plots. I put this one up to see who is awake. I've given this talk to a number of FSAE teams over the last ~10 years.

[ChassisSim]

Hello everybody,

Well this discussion has certainly taken on a life of it's own hasn't it!

Some really good food for thought about compliances. I've done a lot of work in two formula that brings this to the fore. These categories are German DTM and highly modified sports sedans. I've learnt a lot from both of them so let me share some lessons learnt because it is very easy to get lost.

The first principle about compliances is double check with a hand calc if you need to worry about it. Once you are starting to deal with outrageous spring rates and large downforce you need a simple model to break down what is doing what. It was really important when I was resolving some questions with DTM and was a great help and saved a lot of heartache.

Also have a look at your logged data and use that to calculate some loads and or perform rig tests. This is really important to nail down steering compliances as well.

Lastly you can learn a lot from a K&C rig. For those of you based in North America it's worth paying Morse Measurements a visit. Here is the link,

http://www.morsemeasurements.com/

All of the above will tell you what you need to worry about.

Also let me also illustrate a key principle about simulation. Start simple and get cute later. Don't do it the other way around you will get lost. I really needed to state this because I could see this discussion spinning off in that direction. The ultimate validation of this is one of my US customers who have won the LMPC category at the Sebring 12 hour two years running,

http://www.chassissim.com/blog/chass...ebring-12-hour

Their model started simple and then got added to. Given I haven't had that much to do with that model their results and the fundamentals behind it speak for themselves.

Good food for thought.

All the Best

Danny Nowlan
Director
ChassisSim Technologies

[BillCobb]

Gee Doug, now I don't have to show and tell. The "Where have you been" remark pertains to a statement made previously that compliances have never been talked about in the industry or literature.

As for tools to accommodate bushing, Cardan joint, and flexible beam elements, We had SNAC (Static Nonlinear Analysis Code) running on mainframes from punched cards in the mid 60's, then from foreground TSO sessions on mainframes soon after.

Knable & Associates sold a version of this called SKAT (then to SSNAP in the late '90's and then on to SuspensionSim in the 2000's. SuspensionSim is available now thru the CarSim folks.

As you can see (first 2 pics are 1 file split into 2 images, control arm bushings were an optional element, as was frame or structural compliance (think twist axles) besides wheel bearing springs. Running with and without settled all bets on their necessity. 'Theoretically", you still need the nonlinear pressure - angle data from the steering valve to get the steer vs. Mz compliance done right. Even for manual steer cars the Cardan joint stiffness(s) can be a real source of error.

So now there is SimMechanics able to perform as much as you want, if you have the license and the savvy. Need a bushing press, though (supplier data) and maybe some fixturing to get the structural elasticity.

[Z]

I, too, fully agree that structural compliances should play a much greater part in VD simulations than is currently the case (ie. in some of the Simulators mentioned here).

But, big question again, HOW IMPORTANT IS ANY ONE "COMPLIANCE"???

My point is that until a thorough ERROR ANALYSIS is done on these things, it is all just a moot argument.

For example, a current fashion in the production car world (well, last few decades) is to give the wheels large longitudinal compliances, mainly for less NVH from low-profile tyres over small sharp bumps. Such large compliances, if done correctly, may NOT have any significant effects on laptimes, at all. In fact, they may improve laptimes? But as noted before, small "deflection steers" can have massively negative effects.

So, IMO, a good Simulator should first clearly state which compliances, kinematics, dynamics, etc., it is modelling. And then, most importantly, it should very clearly state the potential errors, or UNCERTAINTIES!!! (as per Prof. Lewy), that result from its choice of things to model, or not to model.
~~~o0o~~~

Perhaps the most effective way of driving this point home to the students is via Bill's suggestion:

... This [K&C rig] was early 60's technology and could easily be duplicated by a creative FSAE team and stationed on a trailer at a track to make some money with.

If not done by a student Team, then this is certainly something that the Organisers should think about.

I envisage something very simple mounted on a box-trailer, with the FSAE car tested up at table-top level so the students can easily see everything getting twisted out of whack. Simple ratchet-straps to apply loads, and fish-scales (0-300 kg) and string-lines/tape-measures to measure loads and deflections. This would give a high level of "believability", and guarantee of no fudging in magic boxes.

Output of this K&C rig could then be fed into a VD Simulator and compared with the "undeflected" car. Open-loop, of course, with no "global fudge factors" applied to either simulation.

The Team with the floppiest car should win some sort of trophy. Perhaps this?



Z

[Tim.Wright]

The problem with modelling compliances on a race car is that they come primarily from the gross deflection of otherwise rigid parts NOT the joints.

Road car compliance can be fairly accurately modelled using rigid bodies and compliant bushings because most of the compliance you see at the wheel (in terms of toe, camber etc) comes from the bushings.

Racecars replace these bushings with spherical bearings which (in my opinion) are MUCH stiffer than the suspension and upright members (because they are basically small blocks of solid steel - and stiffness is inversely proportional to size). Now your total compliance at the wheel is MAINLY coming from the larger, thin section parts, i.e. the links+upright+wheel. Now, only a small portion of the wheel compliance comes from the joints themselves.

Therefore, attempting to model them by introducing point elasticities at the suspension pickup points is not correct. It's not even conceptually correct. I've tried and seen other attempts to model rigid body compliances by introducing point compliances in an MBD model of a race suspension and its an abortion. It simply doesn't work.

So the only ways to characterise this properly, if you are still in the design phase, is to create a full assembly FEA of the suspension (this is a massive job). Thereafter you could introduce flexible bodies into your Adams model or characterise the curves for input to a lookup table.

Unfortunately its a bit of an impossible problem for a university FSAE team. Honestly, who in the academic world would have the tools, the expertise and the experience to be running analyses like this?? Where do you draw the line between cost (in time and effort) and benefit (sim accuracy)?

[dynatune]

Again I am fully agreeing with Tim .. but since there is still a need for some well educated engineering guesses we need some tools.

And the question from "Z" How important is any one compliance ? That is indeed the key question and - forgive me to make some PR at this point - but

since I am utterly convinced that compliance should be considered in any suspension or vehicle dynamics analysis tool, I have to refer to some work that

has been done on creating tools that can be helpful.

I agree that a "K&C" rig would be a very nice thing to have but since it is not yet available to many teams and many teams do not have the knowledge to

run MBS / FEM tools accurately. Ergo, has to divert to some "simple" steps to get some indication. As has been announced in another thread here our

suspension design tool can handle - in an elegant way - suspension deformation (can be read about here http://www.dynatune-xl.com/modeling-sdm.html). One

can for instance simply turn on or off "compliance" on every link and start understanding what is actually happening in your suspension. The tool comes

with link loads in order to indicate where your chassis should not be made of spaghetti. Certainly the tool is not as fancy as MBS software or a

fullblown FEM will certainly be more accurate, but as a pocket calculator it does it's job very neatly. Quite a few SAE teams have started using it.

Coming back to the effect of compliance the handling characteristics: This is also included in all our R&H tools, or better said is actually the backbone

of the vehicle dynamics tools. It is possible to see what the effects of 5 selected "most important" compliance parameters are on handling (linear and

non-linear range). Besides that, their effect on maximum achievable lateral acceleration is instantly calculated allowing to get a direct feeling for

their impact on max. performance (always for a given tire of course). One can quickly turn on/off each compliance parameter and look at the consequences

and indicate to you where to pay attention.

I have seen that one can easily loose 15% or more of the cars maximum g capability by having the "wrong" compliance on the "wrong" place on the car and

that is why I keep on saying that the effects of compliance should be included in any kind of vehicle dynamics simulation software. In someway

somehow....

Excuses for the text formatting ... that happens lately somehow by itself after publishing ....

Cheers
dynatune, www.dynatune-xl.com

[BillCobb]

Famous last words: "Say, that's not supposed to be moving. I did a kinematic analysis of the steering ...."

[DougMilliken]

Hi Bill -- where did you attach (ground, reference) the plate and tubes that the magnetic base indicators sit on? With everything bending around I never was completely happy with any of the places that I've tried.

[ChassisSim]

Hey Guys,

I just posted this latest video tutorial on the ChassisSim website,

http://www.chassissim.com/blog/chass...assissim-v3-26

It's about using the Aero track replay feature in ChassisSim v3.26 to generate an aeromap. It's a feature we have just released that has proven very powerful in testing.

Strictly speaking I know it's highly unlikely you'll use it on a FSAE car but it nonetheless to have in your back pocket if you ever work in a race team.

Enjoy

Danny Nowlan
Director
ChassisSim Technologies

[ChassisSim]

Hey Guys,

Sorry for the lack of content. I've been on the road visiting customers and other engagements in Europe.

Just a quick community notice. A big congrats to ChassisSim customer ORECA for their ORECA 05 chassis winning the LMP2 category of the 2015 LeMans 24 hour. More details can be found here,

http://www.chassissim.com/blog/chass...15-lemans-24hr

A just reward for all their hard work.

All the Best

Danny Nowlan
Director
ChassisSim Technologies