Sign In
Forgot Password?
Sign In | | Create Account

Lower the Surprise Ratio

Nazita Saye

Nazita Saye

Posted Jul 15, 2009
4 Comments

From time to time I lurk around CFD related forums – just to see what people are talking about. A couple of weeks ago I found a posting on the subtleties of meshing and its relation to accuracy. The discussion rapidly became quite heated – there were some folks who believed that unless simulation gives you 100% accuracy then it is useless while others didn’t believe that accuracy is the only factor in the equation. In my mind’s eye I could see the whole conversation taking the shape of Budweiser’s campaign of “Less Filling, Great Taste” from a few years ago. One of the engineers on the forum finally said: “designers are interested in lowering the surprise ratio … not in getting an answer to another order of magnitude in accuracy. They are just trying to get it in the ballpark of success.”

Reading this conversation made me realize that from time to time we all need to remind ourselves that the correct answer to the question of which is more important: accuracy or getting a usable answer, is both!

I touched on the subject of meshing in my previous post. A lot of people consider meshing a “black art” especially when using traditional CFD tools. You need to tinker with it and coax it until it rewards you with a fine enough mesh so you’d get 100% accuracy. But there’s a price to pay for that kind of accuracy — time. I remember hearing about an engineer at an aerospace company who would spend 3 months on creating a “good enough” mesh.

In the fast-paced world of design, there’s very little room for any activity that distracts the design engineer from what they need to do: design. So it stands to reason that any activity that is spent on other things is not a good spend of time. Transferring data back and forth … waste of time. Deciding when to use which mesher … waste of time. Refining meshes for days… is a waste of pure design time. During the design process, engineers are mostly concerned with doing sanity checks … figure out if they’re going in the right direction. This iterative process doesn’t require 100% accuracy – a 5 to 10% deviation is OK.  Once a few good candidates are chosen then the design engineer can go for broke. Now FloTHERM, FloVENT and FloEFD users are quite lucky because they get to take advantage of both speed and accuracy simultaneously (please look for one of these products under the Mechanical Analysis Multimedia tab to see  what I mean).

Now I betcha you’re asking yourself, what’s this got to do with ROI (the main theme behind this blog)? It all goes towards making sure that the investment that you’ve made or will make in simulation software, pays for itself by enabling you to do what you do best: design in a timely manner.

Until next time,

Nazita

Mesh, FloVENT, CFD, Design Engineer, Accuracy, FloEFD

More Blog Posts

About Nazita Saye

Nazita SayeI have been involved with the CFD user community in one shape or another since 1999 -- when the NIKA team first introduced FloWorks to the engineering community. Over the years I've seen the market evolve and I still marvel at the wide range of products that are being designed with our tools. As the Manager of External Communications for the Mechanical Analysis Division at Mentor, it is my privilege to bring some of our customer stories to you. Visit CFD doesn’t mean Color For Directors

More Posts by Nazita Saye

Comments 4

Post a Comment
I thought the principle was that you progressively refine the mesh until the answers stop changing significantly. Then you stop refining. Pretty sure there's a Flotherm technical paper on this somewhere...

Chris Hill
10:23 AM Jul 16, 2009

Hi Chris, you’re absolutely right… I should have been clearer about what I meant. The aerospace engineer took 3 months to create a mesh that would give decent results because as you said you would need to refine your mesh and check your results and wait for the results to stabilize before claiming 100% accuracy. What I was trying to say relates to that of quick moving design … when you’re trying to get an idea about the feasibility of a possible design. Once you decide that a design variant is going in the right direction, then you would strive for a fully refined mesh.

Nazita Saye
10:53 AM Jul 16, 2009

Hi Nazita, Excellent article, I’ve read a few similar discussions over the years at CFD forum and there always seems resistance between pure analysts and engineers using CFD software. I use FlowEFD(Floworks in Solidworks) on a weekly bases for analysis of the thermal performance of cameras for mobile phones. What I need is just a rough ball park figure initially; even within 30% accuracy is ok for a first look just to get a feeling about the design. We well then refine the design based on this quick analysis and run a more detailed analyse once we have a final design. I think this is where FloEFD is invaluable as doing these quick iterations at the start of the project lets us refine the design before things are frozen. If we took weeks to refine the mesh then we would miss our design window and would have little chance to make meaningful changes to the design.

Stuart Brodie
9:35 PM Aug 12, 2009

Thanks Stuart -- for your compliment and for taking the time to write in. Your type of application, verification during the design phase, is the reason why FloEFD was developed in the first place. So I'm very pleased to hear that your organization is taking advantage of the technology and its resulting benefits. Now that's not to say that we don't have a few analysts who have embraced the technology too. Usually the analyst group in large organizations get involved in the purchasing decision process for the design team. At one company, the analyst who sat through the demo was asked what he thought of the product and whether it would be a good tool for the design team. The analyst said forget about the design team ... I want this myself :-)

Nazita Saye
9:10 AM Aug 13, 2009

Add Your Comment

Please complete the following information to comment or sign in.

(Your email will not be published)

Archives

Tags

 
Online Chat