Group Purchase of Aerodynamic Modeling Program
 

I have just joined but I see an obvious need for us to be able to do aerodynamic modeling on our proposed mods. I looked all over this site and others for someone who loaned out or could run some tests for me but found no one. The big automakers have done this modeling to design their generic vehicles and have programs specifically designed available but I doubt they will be open source for us, they tend to be conservative and need people like us to show them that the public does want better vehicles. We need this capability in house because sometimes logic and intuition fail in the aerodynamic world. There is software out there but it is too expensive for an individual to buy. But as a group purchase we could spread the cost so that it is affordable to members. I am not sure just how we could work it. We would be buying one license that would run on one machine. I do not know if we could link as members and run our individual test or if we had to submit our 'run' to the 'Keeper of the Program' and he would send us the results. And a data base of successful mod data would be kept for those new members who wanted to do the same. Once the basic car or truck model was generated then the added mods would be easy to change n adapt. That way we all benefit.

For example, I own a 2006 Ford F150 XLT Supercab w/ short bed. Once I have modeled the aerodynamic bed cover or the best air dam / belley pan then other F150 ecomodders could use my data to improve the design of their own. It seems foolish that we are each spending so much time re-inventing the wheel. I want to make my vehicle more fuel efficient, not spend a massive amount of time researching what might work and meanwhile, not modding my vehicle.

Is anyone else interested? Is there a way to do a survey?

I know that Solidworks is about $4000.00. That seems scary but I also belong to "Cast Boolets" forum as a cast bullet shooter and we only took two weeks to donate that same amount of money from members to buy our own server. There may be programs out ther that are both cheaper and better. How about telling us here if you know of one. Not just CAD but actual flow modeling is what we need. Lets hear from you.

Why spend that much money? There's a lot of open-source software from the academic world. Try looking here: CFD Online - Links - Software

Great site for maybe finding a software package that will be powerful enough to actually realistically predict how a shape mod to a car will react. I am a skeptic because I have been burned so many times. To often freeware has not been able to do what it promised. Unless you become a super geek on it and can tweak it to work. And few of us really want to become a specialist programmer in some limited program. Did you have a program from that site that will do CFD of complex 3d shapes like a car? I mentioned Solidworks because I know it is a mainstream program that will be supported and can do the task. I looked at OpenFOAM but it only runs on Linux and I do not want to learn that too just to run OpenFOAM.

The point is that not everyone can invest the time to learn all this stuff. So if the ones who did/do have the time could help out all of us to better our vehicles by setting up a place where each of us could test our ideas and then have that available for the next guy to build on. The next person could look at the tests and not have to do it again. And either use what was proven to work or they could try something new. Eventually a new ecomodder could come on board, look up airdams for his particular vehicle and seeing what works and what does not, built it one time, efficient and right. That means more cars actually on the street instead of their owners sitting on the couch saying "I'd do it if I could be shown how." and nothing happening.

I personally have spent weeks, most evenings, reading threads here and I still do not have answers that many of you had years ago. Lots of opinions but no real hard data on how to DO the mods.

Making the CFD model for our vehicle is probably beyond most of us who are not attending college or university somewhere where you can pick other minds and get advice easily. So if we had a good CFD program where the basic model (for that person) could be shared we would make more progress. And even if we don't do the CFD here we could still establish a library of car models that members could use.

My guess is that a) most folks would rather try and confirm on their own, but more importantly, b) any given modification will perform quite differently on different vehicles, and we have a hundred different vehicles or more on the go here...

Not that it's a bad idea. Good discussion starter.

windrider919 -

There are people here with access to (relatively) low-cost CFD programs that they use in pursuit of their degrees. However, the time and computer horsepower involved makes a communal venture complicated (not fair to them!). Here is a thread that is tangentially related to your question :

http://ecomodder.com/forum/showthrea...-cfd-1636.html

What would be cool is if we could solicit a patron to provide a Server + CFD software gratis. Members would only be responsible for providing their own car designs.

CarloSW2

Of course you can turn that argument right around: I don't want to learn Windoze just to run SolidWorks. (Indeed, I will not use Windoze period. I've tried, and don't need the aggravation :-))

CFD is not my field (what little I've done, some years back, related more to atmospheric modeling), but I know enough about it to know that it's going to be complicated, whatever you do. Building 3D models, especially accurate ones of real-world objects, is not easy. Doing the computations takes a good deal of compute power. Any way you go, there's going to be a steep learning curve, and another steep curve in model building. I think open source is the only practical way to go for amateurs, because to make much progress you need a community willing to share knowledge & models.

Getting someone to donate would be great but I think we could do it ourselves.

I read here that the mod that has the most effect on a modern vehicle is improving the aerodynamics. Members are trying all kind of modifications, to their driving habits and to their vehicles. So it seems that most of them will do something aerodynamic. But as stated earlier, some aerodynamic changes make things worse. An example is the truck bed. It seems that leaving the tailgate down would help with less drag when in fact, leaving the tailgate up causes an advantageous vortex behind the cab. And even though there might be 100 different vehicles types among members, if we could test the mod before the person invests time, materials and money we could stop (some) mistakes.

As an example in the thread start, I also belong to a reloading forum that just a couple of weeks ago bought our own server. That forum has a few hundred less members than this one and yet when the call went out for donations to gather $4000.00 the goal was reached and surpassed to $4800.00 in two weeks!

If we did something similar the ecomodders could use it to set up both hardware and software. Those who contributed would get computer time. Someone here who has the time and inclination could run it, perhaps for a small run fee to pay for their time. I am sure that if we were trying to run a bunch of models the person would get tired of donating free time fast. I am trying to be creative here, thinking both inside and outside the box. Personally, I am disabled and run a part time business on the side so I have the time but not the knowledge. I could learn. Someone else might have both and be willing. Thats why I am throwing the idea out to all of you. Because with as many members who could benefit it spreads the cost to something we could easily afford.

Forum has about 9k members and lists +-1500 as active.
If even 1000 members contributed $5.00 each we could be up n runin!

Ex: It would be worth spending some money to me to check my mods before I build them. I would shell out $50.00 or maybe even $100 to test because it could and probably would save me that much wasted epoxy if I had to change something later.

My opinion is that getting $4000 in donations to buy the software would be the easy part. I'll be much tougher or again expensive to have accurate 3d models to be loaded and then lean how to work with the software.

The ROI from what you might find will be extremely low.

I think you'd be better off just following in the steps of people who went down the drag reducing route before. You can do a lot of Cardboard Aided Design, tuft testing and coastdown tests for $4000+.

I have a University copy of Solidworks 2008. Its such an incredibly complex program. I've been trying to figure it out but you really need to spend a lot of time learning how to use it.

Solidworks has a program called FloWorks that its quite possibly the best CFD program i have ever played with. The challenge is making an accurate model to do the wind tunnel testing with.

Has anyone actually looked at the license agreement for Solidworks? I highly doubt you can buy a $4000 license and have 1000 individuals using it...

Alibre is a pretty good package that is MUCH cheaper than Solidworks.

One major upside to Alibre is that it readily saves in ACIS format. Doing some research, Gambit can make meshes from ACIS files. OpenFOAM can convert Gambit meshes for its own use.

I have access to Solidworks, and I'm not terrible at making 3D models with it. I also fired up SolidWorks FlowWorks Express, which is totally inadequate to my EcoModding needs. FlowWorks Express makes attractive (but very low detail) visualizations of flow, but it's incapable of a quantative comparison of one model vs. another.

I think ChuckM has the right idea: CAD software to make a 3D model, export to OpenFOAM for analysis.

Looking at openFOAM it appears to be a strong program. Instead of trying to run MS and Linux on the same drive I think I would use a second, dedicated drive for the Linux / openFOAM.

As Per licensing, We would have it on one machine, run by one person who would be loading and running other peoples files. So legally there would only be one license. Putting it on the network would make them {Ape > Banana = Big Pile of..... AS} so that would be a big NO.

The vehicle model library would not need to be licensed I would think. Just as each 1st person created it he would make it available to the other owners of brand X.

So you would go to the list and just click on your car to download your particular 3d model.

Geo Metro
Scion
Accord
Camry
Protege
Megane
Jetta
Elantra
F150
Silverado
Previa
etc..................

Example , How many downloads of the Metro would there be? Lots I bet! And as test runs were done the most efficient car would evolve.

I would love to see some collaboration to conduct CFD analysis of ecomodded cars, thus definitively answering questions about how many MPG rear wheel skirts would give, and comparing the relative effectiveness of several proposed grille block designs, for example.

This collaboration does not require money. As you said, if we create a model, it is our property, and we could share it here under a Creative Commons license. OpenFOAM is Free, so no money there. For CAD software, any software that can handle .DXF (drawing exchange format) is fine. If you're rich, buy a copy of a professional CAD suite. If you're a student or an engineer, use your employer's/school's copy. If none of these apply to you, get free CAD, such as this one that was on the first page of Google:

GSI: CAD X11 Features

The problem with that might be the compute time. As I said, it's not really my field so I can't say for sure, but you might be looking at days/weeks of number-crunching to do an analysis of one model on a typical PC.

The amount of computation required depends on how detailed of a simulation you need to run. Higher resolution and smaller time step mean more accuracy, but there's rapidly diminishing marginal returns on that. My best guess is it would take hours, rather than days or minutes.

I'm moving my home, after a bit of an absence. Thank you all for bouncing ideas around for this.

As mentioned months and months ago in another thread on here about OpenFOAM, I was intending to build myself up a linux server at home so I could try and toy with some of the freely available software, like OpenFOAM (it is, by the way, a total pain in the butt to set up!)

Since I'm moving said server half way across Canada, I can't promise either resources, or results anytime soon. But, I am prepared to put in a bit of elbow grease, and try to keep notes on my absent minded explorations once I set up house again, so I can do so.

----

My exposure and dealings with installing, let alone using OpenFOAM make me think of this analogy:

Having coal, sand, iron ore, and a furnace aren't the same as having a steel fender for my car.

That it's a big leap between the two for me, and most people, but I'm sure there's a few people on the planet that could make their own blast furnace and then blacksmith a workable fender, instead of buying one ready to paint.

But, I think I'd be closer to being in that smaller group of people who'd make the fender, IF I were privy to the knowledge and felt comfortable with the steps to do so.

The same goes with OpenFOAM. I've seen end results displayed that show it would do what most of us want. I'm not so sure at this point that I can acquire the knowledge to do so, yet. Everyone that's good at it, is focussing their knowledge on using the program, not building up documentation for newbie end users, at least!

----

What this idea needs is a lot more doing, and a lot less pontificating (a.k.a. immitating "The Thinker" statue), and reporting back on successes and failures, while we fumble about with

• Setting up a server at home, so others can repeat your progress
• How to's on installing some of the free or close-to-free software available for CAD, model manipulation, airflow simulation
• Documenting our exploration into how the different available programs work, so that we figure out which ones are "best" for us to try and use, and/or modify
• Gathering resources (sites, article/book bibliographies, pre-existing HowTo's) so that there's less re-inventing the wheel

Maybe a sub-forum off of the Aerodynamics forum is needed in this case, to help persue this?

Either that, or say, people of like minds here can dig about, identify an existing forum for this, and we invade our little corner of it, to further our "I'm an amateur that wants to model my car in a computerized wind tunnel" agenda.

Right now, all our pokes about are too incoherent, unfocussed, and not collaborative enough so that any one of us can make reasonable progress. We can change that; we just have to put our minds to it. :)