Part Concept Selection and Decision Process

[LaurenceSachs]

Hi there,
I trust you all are having a good start to the new year!

As a design engineer, I continuously have to look at new ways of designing components or products.

I obviously have to take cost, weight, material, corrosion, reliability, performance etc. into account. I am familiar with using a Weighted Design Matrix. Where all the options are listed and scored based on their affect on the relevant criteria. The option with the best score is the design that we use.

Now I am finding this unreliable, and I would like to know whether you have any process of making calculated decisions for designs?

Any feedback would be great!

Laurence

 

[rb1957]

why unreliable ? could you give an exmaple ?

the problem (and strength) of the matrix method is that you need to capture all the relevant dimensions and weight them correctly.

another problem is that over time weights will change and dimensions will probably change.

 

[LaurenceSachs]

Hi,

Well yes, part of the problem is that weighing each of the criteria like reliability, corrosion etc are all user inputs and therefore manipulated. Also when scoring each concept, the scores are related to the designers existing knowledge or lack thereof. It is also possible that the designer is biased towards one concept - this renders the entire process pointless.

L

 

[rb1957]

it does, doesn't it !

in the bad old days we'd have one guy making the decisions, and his judgment determined the success/failure of the project. and he knew this (ie his accountability).

today we have committees and processes, and no accountability ...

 

[MiketheEngineer]

A camel was just a horse designed by committee - probably a dozen or so engineers!! Like me

 

[KENAT]

You may be able to find some kind of reference for some factors, for instance for serious FME(C)A analysis there are data bases with reliability figure for standard component types etc. However even then they'll be generalizations.

Unless the people filling the matrix are appropriately experienced/knowledgeable about the various factors you don't stand any chance. Even if knowledgeable it's still subject to manipulation as both of you mention - be it intentional or subconscious.

Remember the table still operates on the GIGO principle.

The Camel is not really a fair analogy, while it may not have the aesthetic sensibilities of a thoroughbred, the camel is incredibly well designed for its task.

Posting guidelines faq731-376

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

(probably not aimed specifically at you)
What is Engineering anyway: faq1088-1484​

 

[MikeHalloran]

>>>A camel was just a horse designed by committee <<<

It's a fun old saw, but it's completely wrong.

A camel is spectacularly well-designed for its native environment.
... which is very different from the horse's native range.
Any similarity between the animals is either coincidental, or a response to some common evolutionary pressure.


Speaking of which, Laurence, you now have an opportunity to evolve your product, whatever it is, by going back through its accumulated history of design changes, and coming to an understanding of how each change was or was not an effective or appropriate response to evolutionary pressure.

Pay particular attention to major changes or sets of changes; often enough, you will find that part of the 'old' way was a better solution than a feature that was adopted mostly because it was 'new', and because there was some problem with the old version. Very rarely does the perpetrator of a major design change take the trouble to measure how, or whether, the change was actually effective.

But now you have, or can derive, data specific to both old and new versions. Maybe you even have better ways to measure the performance parameters that are actually important.

Take the time. Do the math. It usually leads to a whole new level of understanding.




Mike Halloran
Pembroke Pines, FL, USA

 

[rb1957]

i don't think the camel adage is meant to cast dispersions on the camel itself, i'm sure it is a truly noble creature and another expact of nature designing animals to fit niches (if they don't, they die off).

i think the adage is a comment about designing a horse, and i think you have to admit a camel is a pretty dam'd ugly horse !

and, yes, i know this is all tongue in cheek ...

 

[dhengr]

Rb1957 hit the nail on the head. Isn’t democracy and design by committee a wonderful thing, nobody has to be responsible for anything. In the good old days, a good design engineer, manufacturing engineer and/or lead designer did the matrix thing in their heads, whatever their biases and knowledge and experience level, etc. Hopefully, they got together and compared their notes and the boss made a decision based on the sum of that knowledge. The more formal matrix has the potential of being somewhat less biased by one person or element or another, but now the full time job becomes continuously monitoring and adjusting the multiplier factors, based on real good experience of all of the factors, and not being biased as you do this, and it probably should still involve the same committee as above. That’s your job in making the matrix work for you.

The problem as I see it today (and this is an older guy’s perspective, maybe more experienced and knowledgeable about some things ) is that we have way to many inexperienced people doing way to many independent little parts of the total, with insufficient guidance or knowledge of the whole machine; then this is all thrown in a blender and out pops the finished part. Actually, the camel is not a bad horse, particularly in the desert where water is in short supply and their larger hoofs deal better with the dry sand and poor footing. That design committee didn’t do too bad a job. But today, we have inexperience people using CAD who really couldn’t do drafting otherwise. It must be right if you can do it in CAD, right? No knowledge of the real facts required. We have inexperience people using modeling software and FEM and FEA who couldn’t take a first shot at a simple beam design without this software. They don’t even know that they are modeling it wrong, after all the program spit out results to 12 signif. places, that must be really really right. Then many companies have nobody with the knowledge, experience and overview ability to guide this whole process.

This is a fundamental problem (flaw) with the direction engineering and design is going these days, and I don’t think there is a big enough column (definable element) in your matrix, or enough multipliers and adjustment factors to account for this deficiency/problem. All of these new tools and methods have the potential of being improvements in the way we do things, and the way we can finesse designs and analyses today, but we are not combining and using them correctly. They will not take the place of real experience and product knowledge, not in my lifetime, and companies are making a big mistake thinking otherwise.

 

[MiketheEngineer]

Yes - it was meant to be an adage confirming what DHENGR so eloquently put into words..

When I run into this type of situation, I run it by a few good engineers and a few good field people - they are the ones that are going use it, install it, rely on it.

If that all goes well - then we consider putting it into play. Obviously ROI now becomes involved and the bean counters have to buy off on it.

 

[dhengr]

I take back the nice things I said about Rb1957 in my earlier post, after the way he has disparaged the elegant ‘slightly humpy, big footed, desert horse.’

 

[LaurenceSachs]

Hi there,

Thanks for the input.

Any more suggestions or spreadsheets that you use?

Attached, See the matrix I may use in the future:

 

[tomwalz]

I am not sure you are getting the part where you have to find folks who know what they are doing.

Data never tells the whole story. It just can’t. You need input from folks who have been there and done it.

Last month I received a request for a quote for a piece of tungsten carbide that was to be 12” x 12” x 0.5”. It was to be used as a skid plate on a flying vehicle. Tungsten carbide is often used as a wear plate so it made sense.

However tungsten carbide is very weak in some ways. With careful and expensive grinding of a much larger part you might be able to grind down to the desired dimensions. However the betting here was about 50:50 on whether it would snap under its own weight if you picked it up by a corner.

It would definitely shatter under any sort of impact.

I did the engineering team a favor and sold them an STB (Standard tooling blank) for $20 and told them to experiment with it. Specifically clamp one end in a vise then hit it.

When you see something like a TRS (Transverse Rupture Strength) of 500,000 PSI, it looks really strong. You have to have somebody who knows what it means.

If I had been more clever I could have tried for an R&D contract to produce the part. One of those low ball deals where you make a fortune by charging for every change. (You need somebody smart enough to keep you out of those.) I hate those. I have been in too many accidentally.


Thomas J. Walz
Carbide Processors, Inc.

http://www.carbideprocessors.com

Good engineering starts with a Grainger Catalog.

 

[HDS]

How many people really use these matrixs?

 

[rb1957]

tom, how much money did you cost your company ?

dhengr, i called it "a truly noble creature" ... geez, what more do you want ? my membership card to the "camel lovers assoc." ?

 

[MiketheEngineer]

I just love camel hooves.

 

[SWComposites]

The only real value in these "Weighted Design Matrix" type analyses is to fool management into thinking you know what you doing. The criteria and factors can be manipulated to get any answer you want. See the above comments about using experience to guide design decisions.

 

[rb1957]

@mike, hooves or toes ?

@SWC, i don't think i need to fool mgmt that i know what i'm doing, i think i have to fool them that they know what they're doing.

 

[KENAT]

I've got to agree that unless it's some kind of high volume/high cost/high 'risk' or similar component, it seems an awfull lot of effort to go to.

Posting guidelines faq731-376

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

(probably not aimed specifically at you)
What is Engineering anyway: faq1088-1484​

 

[MiketheEngineer]

Right - TOES!!!

Just too beautiful!!!