Automobile angular momentum

[GEspo]

Will a vehicle rotate around its sprung center of mass while turning? I’m trying to nail down, without CAD, my inertia tensors for components like engine, body, fuel, in rigid body dynamics software. Something is telling me tire loads need to be considered.. thx

(EDIT: considering global and body frames, will a vehicle rotate around its com in the body frame)

 

[BrianPetersen]

It depends. At low cornering loads, it will rotate around the imaginary intersection point of the (steered) front and rear axle centerlines.

What, exactly, are you trying to do? I'm a "car guy" and I have no idea what an "inertia tensor" is.

 

[GEspo]

Yes, I was thinking the same, depends on v. The vehicle(ie com) will rotate around the radius of the turn(global frame) up to a certain v at which time the angular momentum will overcome the friction forces on the tires and create either oversteer or understeer, or total loss of traction(rotation in the body frame). This "overcome" the friction forces is spotty for me and an intuitive guess. I probably can find the answer by digging into the angular momentum eqn..

The inertia tensor is associated with any mass and is part of the angular momentum(L) eqn L=Iw.. I inertia tensor(3x3 matrix) and w angular v...

When setting the tensor we are only concerned about what the mass will do when rotated "freely"?

 

[onatirec]

I'm not a car guy, but I believe a vehicle rotates around its roll center, as compared to center of mass. Physics would lead me to believe that sprung mass alone could not adequately characterize this - as all massive components would see inertial effects, not just sprung mass. I suspect the sprung masses would be more susceptible to introducing off-diagonal terms in an inertia tensor.

(roughly speaking, inertial tensors describe mass distributions which are used to calculate rotational accelerations under applied torques or vice versa; it's just a higher dimensional expression of T=Ia)

 

[GEspo]

Good point about the roll center(axis).. There is yaw and roll to consider, yaw being the rotation of the front of the car relative to the back, and roll being the rotation of the top of the car relative to the bottom etc...(and pitch too).

OR do we have it backward and yaw roll and pitch are effects of angular momentum...
Now we're back at the beginning of will a vehicle rotate around its center of mass (sprung or total mass)


Here's a new formulated guess: the forces acting on the tires are opposing the "freely" rotating center of mass, creating the roll pitch and yaw effects

 

[BrianPetersen]

Are you talking "rotate" in world co-ordinates when viewed from the top (plain ordinary steering), or "rotate" when viewed from the front of the car in vehicle co-ordinates (body roll), or "rotate" in sided view in vehicle co-ordinates (pitch)?

 

[BrianPetersen]

And, you still need to back waaaaayyyy up and tell us the big overall picture of what you are trying to do.

Are you trying to calibrate suspension settings? And if so, for what purpose? Racing? (and if so, what type)?
Are you trying to design engine mounts or some such thing?

 

[CWB1]

Under what traction and control conditions, and in which axis? Front loose, rear loose, both loose? Acceleration, decel? X, Y, Z? In short, no.

 

[GEspo]

I have a model in multi body dynamics software, each rigid body(RB ie mass) has an inertia tensor. I have a way to create torque at the rear wheels(contacts) so the total RB, composed of individual RBs and their connecting RB frames, move at some v. I also have a way to rotate the contacts(wheels) so that the entire RB will turn at some angle dependent on rate of turn. I'm in the process of seting the individual RB inertia tensors.. or I can create one RB at center of mass (of all RB).. I don't think for the overall sprung RB mass it matters if I have one RB or multiple since I_total = I_engine + I_gas etc, for connected masses..

The contact friction I set at .6 which is what I found for tire friction on dry earth(vs street), since this is an offroad vehicle model.. long travel 4x4 ifs and solid axle 4 link rear.. .all custom hydraulics for dampers, normal springs etc..

Creating torque at the wheel produces acceleration of the RB in the model

 

[GEspo]

In the link below there are 2 calculations for Pitch(yaw and roll as well) moment of inertia, one total vehicle pitch moi relative to total vehicle centroid and other sprung mass pitch moi relative to sprung mass centroid... Is this saying to compute the sprung mass centroid and use that for sprung inertia tensor, then for unsprung masses compute the total centroid and use that for the few unsprung inertia tensors?

https://www.researchgate.net/public..._PARAMETER_MEASUREMENTS_Compliance_Parameters

 

[BUGGAR]

A car does not roll about its "roll center". See my other posts.

 

[onatirec]

So, correct me if I'm wrong, the roll center is the effective point at which the forces from the suspension act on the sprung mass - (edit: which rotates around it's center of mass)?

 

[BrianPetersen]

"Roll center" is an imaginary construct from kinematics. You really need to be using force-based analysis.

We also still don't know what the original poster's objective is. "I have a model in multi body dynamics software" - WHY do you have that model? Why did you bother constructing it if you are not trying to solve something ... tell us what problem you are trying to solve by doing this?

My knowledge of vehicle suspension is of the school-of-hard-knocks variety, "the bike is doing this, which has this bad side effect, why is this happening and what can be done right now, trackside before the next race, to improve it?" - as opposed to the computer-screen variety.

 

[GEspo]

I've constructed the model for my build of an off road race vehicle. Initially I used the software to optimize the double-wishbone, which I've been successful with, and now am interested in finding out about the dynamics of the front and rear ends and the entire system.. I have degrees in math and comp sci so I'm fairly equipped to solve these types of problems..

The big end game here is to optimize the entire vehicle for various performance metrics... as I work with the software I'll discover its capabilities and if it will be possible to achieve.. From my research in the field of long travel suspension I've yet to find anyone using MBD for their designs, however I haven't contacted the big names like Geiser bros... these trucks can go for close to $1mil btw

NOTE: long travel is usually in the ball park of 20"+ in the front and 30" in the rear... top speeds up to 80-90mph in the desert...

 

[GEspo]

After reading a few papers it looks like for the sprung mass the principal axis is inclined. I reached out to the author of the vehicle parameter pages posted above and he recommended setting the sprung mass and unsprung mass tensors locally..

This paper on airplane inertia discusses inclining principal axis and roll axis

https://apps.dtic.mil/dtic/tr/fulltext/u2/029318.pdf

 

[CWB1]

The bit you're missing is that MBD and other complex models need to be thoroughly validated to be useful. In the racing world there isn't need for chassis builders to get overly deep into design bc they're building to an engineered spec governed by a rule book or customer. Individual race teams then take that spec chassis and modify based on analysis and testing for a particular track and set of conditions, and can feed that back to the fabbies for the next build. SOP for engineering development is to benchmark and thoroughly test existing designs then make incremental changes before settling on a final design. Trying to drive design with a model with no correlation to reality or existing testing is the old "junk in ~ junk out" pseudoscience that most despise. My advice is to buy or build a copy of an existing truck, test/learn, and implement that feedback into your model.

 

[GEspo]

That’s a good take on the situation and I appreciate the feedback. When I set out to build a long travel 4x4 ifs I encountered an industry that was scratching the surface on the technology.. MBD is an extremely useful tool in speeding up production based on the fact that you can test multiple designs in a single day.. As far as junk in junk out that’s yet to be tested on my part. How would you address the fact that without MBD it’s not possible to optimize(mathematical optimization) parameters? have you seen the diff eqn of the DW.. yikes

Since it seems you have experience with racing what is your take on using mathematical optimization for such vehicles?

 

[GregLocock]

Strictly speaking when turning a vehicle turns about the centre of the circular path it is currently on, and nothing else, in plan view. That of course is not what you mean. Try and draw out what you are talking about and it may be possible to extract a definition. Any serious vehicle dynamics book starts with a very boring chapter that nobody reads on coordinate systems and reference frames. It is there for a reason.

Inclination of the sprung mass' principal inertia axes, and the relationship to the roll centre axis (a terrible concept), is a rabbit hole that has swallowed many. MBD software does allow you to correctly estimate the roll pitch and jacking gain, for actual suspension geometries and calibrations.


Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[CWB1]

When I set out to build a long travel 4x4 ifs I encountered an industry that was scratching the surface on the technology.. MBD is an extremely useful tool in speeding up production based on the fact that you can test multiple designs in a single day.. As far as junk in junk out that’s yet to be tested on my part. How would you address the fact that without MBD it’s not possible to optimize(mathematical optimization) parameters?

Top pro teams have all had in-house engineering teams for decades running this analysis and leveraging the latest technologies. Lesser pro teams outsource to experienced firms like my employer to save budget. Even collegiate SAE Mini-Baja teams create these models for improving their cars.

Like most engineers I am a big fan of math models/simulation, have written quite a few, and rely on them heavily. The reality tho is that unless based on actual test data, models are simply a guesstimate with no way of quantifying a change as "better." It may seem like the chicken vs egg, but no engineer worth their salt spends much time building a complex model without benchmark data bc we have all seen/dealt with folks promising to revolutionize the industry based on no experience and bad assumptions. I'm genuinely not trying to dissuade your efforts, simply saying that the path to knowledge and success starts and ends with testing. Basic instrumentation is easy with smart phones and wireless networks. Before I sold my autocross car back in '15, I could capture and correlate every driver input to a track map, GoPro footage, wheel/suspension travel and speed, 3-axis G force, and more without much fuss. Thanks to other interests, nowadays I settle for a simple app datalogging my daily drivers' via OBD to forewarn of likely problems and help troubleshoot.

 

[GEspo]

Sounds like I’m on the right track.. hopefully in a year or so I’ll have my first prototype to test.. I’m no engineering team but giving my best shot to make a safe fun experience.. thx for the comments