My vehicle dynamics web project.

[Kozy]

Hi everyone. I'm not technically 'new' here, but I don't post here very often. Coming back to these pages has made me realise I really should be here more often!

I'd like to share with you a little web project I am working on, called Black Art Dynamics.

Like many people here, I'm a vehicle dynamics enthusiast, although I lack the kind of backgrounds many members here have. I studied Motorsport Engineering to HND level, but then left to take up a job as a design engineer in the rotating machinery industry. Since the HND, I've been learning the subject through self study by mathematical modelling various systems in Excel, and earlier this year I decided to transfer those models into web apps and put together a website to host them as a kind of free resource, also as somewhere to document my progress through the world of vehicle dynamics.

The homepage can be found here: http://www.blackartdynamics.com

So far I've written apps for basic load transfer, brake bias, engine thermodynamics and power train effects on straight line acceleration, and I have a few more in the back catalogue to convert, as well as ideas for future apps

To combine with these, I'm also writing articles explaining them, which incorporate small snippets of the calculations within the posts to create kind of interactive blog posts. Here is the first article I have put together, laying the foundations of tyres and load transfer concepts. http://blackartdynamics.com/Chassis_Art ... ansfer.php

(Please, bear in mind I am attempting to make this information accessible so it is very base level stuff for people here!)

My brake bias tools are currently written for Honda and MX5 models (because I own an MX5 and an ATR), though I do have an 'open input' version in development. The drop downs are designed to encourage people with little or no vehicle dynamics, or engineering based background to come and check that aftermarket upgrades made to their cars were actually going to give desired results.

This was my initial intention, to empower casual tuning enthusiasts to make better calculated decisions. Here is my MX5 version: http://blackartdynamics.com/Brakes/Index.php

This tool uses an algorithm I developed myself to solve for the point of front wheel lockup, calculated as the tractive force - tangential brake force = 0. Obviously this varies with every parameter change but the algorithm solves for it instantly giving results for acceleration and stopping distances in reaction to changes to components and chassis.

There's more content available within the menu bars and lots more to come. Prior to about 4 months ago I had never done any web development so this project has been a very steep learning curve.

As this project progresses I hope to build up the level of technical content, it is still early days but I am enjoying creating it immensely, even if to date, most people I have shown it to glass over a bit!

Anyway, I would absolutely love to hear what members here think of it on any aspects, design, content, technical details etc.

The site is not IE friendly by the way.

[Greg Locock]

Just idly wondering if the goal is communication why you deliberately made your site so that 20-58% of its potential audience can't see it?

Good luck.

[Jersey Tom]

Interesting I suppose.. but ultimately, where's the proof? Click some buttons and make some numbers or charts pop up.. but where's the evidence to show it isn't just BS? Brake temperature charts for example.

And +1 to needing to make the site functional to IE etc.

[Kozy]

Interesting I suppose.. but ultimately, where's the proof? Click some buttons and make some numbers or charts pop up.. but where's the evidence to show it isn't just BS? Brake temperature charts for example.

What proof do you think there should be? It's all school textbook stuff. Energy in, energy out, F=MA, that sort of thing.

There's nothing new, groundbreaking or especially complicated, it's just taking well known models and putting them into templates for the less engineering minded enthusiast to use.

I welcome people telling me where things don't look right, but I'm not entirely sure how I can prove anything?

And +1 to needing to make the site functional to IE etc.

Just idly wondering if the goal is communication why you deliberately made your site so that 20-58% of its potential audience can't see it?

Good luck.

IE accounts for less than 15% of views, and I'm surprised it is even that high to be honest. I do need to work on it, but frankly I've had enough to learn working with the standardized browsers (everything that is not IE basically).

[Tim.Wright]

I'd agree that you need to explain what models you are using and what assumptions you are making otherwise:
a. No one are going to learn anything
b. No one are going to trust the output.

e.g. the brake calculations. What assumptions are you makeing for cooling? What is the ambient temperature? What pad compound is this? Are you accounting for the change of friction coefficient of the pads with temp? Are you including volumetic losses? WTF is the mush factor??

[Kozy]

Mush factor is something I need to remove! A guy running a small US racing outfit asked me to add a couple of things on an earlier iteration and 'mush factor' was one of those things. I agree it doesn't really mean anything, its only a ratio of pedal travel to pedal force, both of which are given and more informative.

The brake temperatures. Energy into a mass of metal with a specific heat capacity gives a temperature rise. At the same time, and during the cooling intervals, energy dissipates according to Newtons Law of Cooling. Ambient temperature taken as constant at 300K (constantly moving airflow keeps this from raising, but this will be improved with a speed sensitive value in future). Friction co-efficient is constant. Not ideal I know, it's on the to do list, nothing is ever completely finished as there's always things like this to be accounted for.

The temperature thing is just a trend really, the bias calculation is the important one. If you're looking for truly accurate figures for temperature, chances are you work in the industry and have expensive software to do this stuff accurately. To most, it's simply a way of visualizing how well matched the brakes are to the power and weight of the car. It will tell you that 325mm vented discs on a 140hp kit car won't get any heat into them. It'll tell you the standard 9.25" discs on a 300bhp MX5 will cook in no time at all. This is the general aim, and not something I believe you can find in any desciption, anywhere else on the net. At least for free anyway.

It IS my intention to explain it all, as seen in the article I wrote. By virtue of being a project, it is not finished yet, so this is all to come.

Also, on the engine thermo, the relevant assumptions are listed.

[Jersey Tom]

What proof do you think there should be? It's all school textbook stuff. Energy in, energy out, F=MA, that sort of thing.

The brake temperatures. Energy into a mass of metal with a specific heat capacity gives a temperature rise. At the same time, and during the cooling intervals, energy dissipates according to Newtons Law of Cooling. Ambient temperature taken as constant at 300K (constantly moving airflow keeps this from raising, but this will be improved with a speed sensitive value in future). Friction co-efficient is constant.

So all of the "textbook stuff" comes with a long list of assumptions. I feel like that could just be communicated more clearly. More so, the implication of the assumptions is what I think is important to understand and state. The Newton's Law of Cooling at fixed ambient 300K for example... how representative of reality is that? You'd think that the local (to the rotor) air temp and any "forced" air flow would be pretty big influences. How can we wrap some error bars around this thing and know just how close or far off the textbook stuff is?

Ultimately to do so, you'd need some real world test data. Lacking that, might be worth stating that the assumptions / approach haven't been validated. Or maybe a sensitivity study? If the air temp were to vary +/- 30 degrees - how much does that swing your temperature plots? How about sensitivity to convection rate?

Not ideal I know, it's on the to do list, nothing is ever completely finished as there's always things like this to be accounted for.

This is actually a good item to stop and think about for a second. A good quote...

Since all models are wrong the scientist cannot obtain a "correct" one by excessive elaboration. On the contrary following William of Occam he should seek an economical description of natural phenomena. Just as the ability to devise simple but evocative models is the signature of the great scientist so overelaboration and overparameterization is often the mark of mediocrity.

Following that, my approach to these sort of things tends to start with asking, "What is the [minimum] amount of fidelity I need to answer this problem. What's a reasonably degree of accuracy?" And following that, "What modeling approaches are out there, what are the error bars on each of them, and what's going to best serve my need?"

IE accounts for less than 15% of views

Perhaps because it is not IE friendly

[Kozy]

So all of the "textbook stuff" comes with a long list of assumptions. I feel like that could just be communicated more clearly.

Noted.

More so, the implication of the assumptions is what I think is important to understand and state. The Newton's Law of Cooling at fixed ambient 300K for example... how representative of reality is that? You'd think that the local (to the rotor) air temp and any "forced" air flow would be pretty big influences. How can we wrap some error bars around this thing and know just how close or far off the textbook stuff is?

Ultimately to do so, you'd need some real world test data. Lacking that, might be worth stating that the assumptions / approach haven't been validated. Or maybe a sensitivity study? If the air temp were to vary +/- 30 degrees - how much does that swing your temperature plots? How about sensitivity to convection rate?

Good points, but see further down.

Not ideal I know, it's on the to do list, nothing is ever completely finished as there's always things like this to be accounted for.

This is actually a good item to stop and think about for a second. A good quote...

Since all models are wrong the scientist cannot obtain a "correct" one by excessive elaboration. On the contrary following William of Occam he should seek an economical description of natural phenomena. Just as the ability to devise simple but evocative models is the signature of the great scientist so overelaboration and overparameterization is often the mark of mediocrity.

Following that, my approach to these sort of things tends to start with asking, "What is the [minimum] amount of fidelity I need to answer this problem. What's a reasonably degree of accuracy?" And following that, "What modeling approaches are out there, what are the error bars on each of them, and what's going to best serve my need?"

I appreciate where you are coming from. The thing is, I've taken that approach, these models are what I have as a result of it. Yes I could take many aspects much, much further, but what I have done so far gives useful enough results. I am not an F1 engineer, the people the site is aimed at are not F1 engineers, or even any kind of engineer for that matter.

However, to my knowledge there is not a single brake bias model available for free on the internet that solves a threshold braking problem correctly, THAT is the real feature, the temperature calculation is purely secondary, it's just a trinket feature. The need to know the brake temperature at all is minimal, so I feel that the method I have used presents a good enough model for the intended use and audience. There is no need to model boundary flow conditions, convection sensitivity, localised temperature fluctuations or anything like that because put simply, there is no need for that level of accuracy.

IE accounts for less than 15% of views

Perhaps because it is not IE friendly

Maybe, but point stands only 15% of people that have viewed the site have done so via IE. Whether or not the site's compatibility is the cause of that I do not know, but I actually think that's a fair representation of the distribution of browser usage at present. IE is just a PITA for web developers, especially new ones like myself, so I have no interest in catering for it at this point in time. I'd rather focus on getting it right for the other 85% of viewers.

[KeithYoung]

I agree on the assumptions. In my articles I tend to make sure the readers know what assumptions are being made.

My traffic is also mostly non IE. I'd personally focus on giving them a link to Firefox/Chrome etc and then working on listing assumptions and documenting things. After you fix those issues look at the others.

My $0.02

[rjsa]

Maybe, but point stands only 15% of people that have viewed the site have done so via IE. Whether or not the site's compatibility is the cause of that I do not know, but I actually think that's a fair representation of the distribution of browser usage at present. IE is just a PITA for web developers, especially new ones like myself, so I have no interest in catering for it at this point in time. I'd rather focus on getting it right for the other 85% of viewers.

That's a self fullfiling profecy. If I land there once and see a thorn white page I turn around and never look back.

You could be not THAT carefull about how it renders on IE, but completelly unreadable as it is just sends a very negative message on how you do things.

In this case completelly disregarding what you think is not cool. Will you do that with theories you don't agree with?

[Kozy]

Will you do that with theories you don't agree with?

If something appears to be wrong, and another theory appears to be correct, then I am highly likely to disregard the incorrect one. I did that just the other week, with two papers on combustion rates, citing two different model equations linking to RPM, and also one I devised myself linking it to piston speed.

The first one gave a parabolic curve of burn duration vs RPM, the second gave a linearly increasing one and my one also gave a linearly increasing one that should have given a more consistent result across vastly different engines.

The first one suggested that above a certain RPM, the burn duration (in crank degrees) decreased at an increasing rate. This is not representative of reality. Eventually the piston will outrun the burn.

The second theory supported this and gave realistic results within a certain window, though it did not work well for very high revving engines.

My one was just an experiment, which worked OK for high revving engines, until you compared two engines of different strokes at the same RPM, then it all went to pot.

I selected the best one (the second one) and disregarded the others.

Are you suggesting I shouldn't have disregarded them?

My theory on browsers is that for the moment, my time is better spent getting it all up and running for the 85% not on IE, then worry about sorting out IE rendering later. I'm not disregarding it completely, just for the time being. I do need to add the message box stating this though. IE is renowned for being difficult as it's just a lame product (MS allegedly spent more on advertising it than they did developing it), not that many people still use it for the exact reason that stuff frequently doesn't work properly on it.

[AidanC]

I enjoyed reading your chassis article. As an engineer with interest but little knowledge in chassis dynamics I found it explained some basic ideas quite well.

I do agree with other above that outlining the equations/assumptions would make the calculations much more usable. For example looking at the 'bar wall thickness' input the default value is 0 for the mx5. This doesn't make sense and short of running a calculation myself I am unsure what value should be here for my upgraded solid swaybars.

[Kozy]

Hmm, you're right, there should be a tool tip there saying to use 0 for solid bars. I realise I haven't actually got ANY tooltips on those pages, so will need to sort those out ASAP. Thanks for the heads up!

[Cold Fussion]

For me the website loads fine in IE (but good grief IE is a slow POS compared to Firefox), so the people having problems are probably on some old version of IE, which is a diss-service to ones self and the rest of the web community.

[Kozy]

That's good to hear it works on the latest version, I've got IE8 on my laptop which won't load it at all!

Added this new article today on NASP engine performance limits, an introducing a Performance Index system I derived.

http://blackartdynamics.com/EngineLimits/Index.php