Weight correction factor to get Not to exceed weight of component 1

[ethanuman]

Hi,
Greetings for the day!

We came across following problem while designing a system (assembly of parts) for aerospace. Please share your thoughts on this.

What is the value of weight correction factor which aerospace industry generally follows for a system with nominal weights?

More Insight in to the question is “Designed products are given with some tolerance considering manufacturing. When we model same in CAD package we usually go with nominal dimensions.Hence weight estimates for the particular product (or system) comes from given nominal dimensions (CAD tool estimates with density inputs).”

Is there any standard approach or correction factor to consider process variations and calculate Not to exceed weight of particular product (or system)having nominal weight?

Thanks
Hanuman
Engineer
Aero-domain

 

[rb1957]

i doubt there's any specific factor. you're concerned (i think) with "guarranteed" not to exceed.

if you design to nom dim'ns and meet the not to exceed weight, then you're hoping that there'll be some plus/minus going on (some components over nominal, some under) to maintain your design weight. if you design to less than the not to exceed weight, you'll have more cusshion (to absorb over nominal parts).

you Could to a detail tolerance study, and put everything at the top end of their dim'ns/weights to get a maximum weight for the ass'y. but that sounds like a lot of work. you might look into the few heaviest components, or the few with the widest variation; that sounds much easier.

 

[IRstuff]

"Not to exceed weight "

Ever? That's a tall order, unless you're willing to reject parts or you have tons of margin. The basic approach is to do the statistical analysis, starting with how many you're willing to reject. Once you know whether 2-sigma, or 3-sigma, or even more, is the goal, you can then tolerance your subcomponents accordingly.

But, it's doable if you have margin, i.e., if your NTE weight is, say, 10-sigma away, then you should almost never reject parts.

TTFN

FAQ731-376
Chinese prisoner wins Nobel Peace Prize

 

[rb1957]

it's the way things are spec'd these days, always really but these days with way more "partnering" going on i think it's more prevalent. Partnering arrangements usually have penalties for missing the weight target; and the weight target has two interpretations.

as a design goal, based on nominal dim'ns, average weights, etc;

as a production process, with all the variation built into the ass'y.

if you design to your wieght target, then you can expect some ass'ys will exceed this. signifcant number ? significant exceedence ?? i'd look into the variation you can expect with a few of your heaviest components to reduce the chance of exceeding the weight in production.

i would expect that the engineers were thinking the first interpretation, but the procurment folks and the laywers will want the 2nd !

 

[KENAT]

Back in the aircraft stores field I vaguely recall a +-5% figure if no other information was available - but that is purely from an ever fading memory.

That is, if a stores nominal weight was 1000lb, then as a start point in structural analysis of the lugs by which it was mounted to the A/C etc. it would be assumed it could be 1050lb.

Posting guidelines faq731-376

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

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

 

[MintJulep]

Your basic premise is incorrect.

Designed products are given with some tolerance considering manufacturing

Assigning tolerances is a design engineering function, NOT a manufacturing function.

It is manufacturing's job to make sure that everything they build is within tolerance.

Since you, as the design engineer has full knowledge of and access to the limiting dimensions you are able to calculate the limiting weight for each component.

If you really want the "not to exceed" number it's simply the sum of the maximums of each component. As that is unlikely to happen you may want to apply some statistical mumbo-jumbo to arrive at an "unlikely to exceed" weight.

 

[ethanuman]

I Thank All for sharing Your thoughts..

conclusion what i can draw from these posts is to do some statistical analysis on major weight contributors(if it is big assy)and arrive at the variance.

But i have a system(around 110 sub assembies) with almost all subassemblies weigh similar to each other. this is where problem arised... when i started to study for all subassembliesi i felt that it would be better to follow standard procedure if there is any.

What approach can give us required result when we have only one set of products manufactured?

Hanuman
Engineer
Aero Domain

 

[berkshire]

In 1991 General dynamics ran an exercise on the manufacturing side of this problem. They ran a statistical analysis of actual dimensions of parts being produced for the MD 11 fuselages they were making.
The aircraft was coming out overweight at the time.
They quickly discovered that given a plus or minus tolerance machinists would work to the high side, working on the reasoning that if it was out of tolerance they could always rework the part by shaving a bit more off. The machinists knew that if the part was machined under tolerance it would be scrapped, or be subject to extensive MRB attention.
GD Started a program of getting machinists to set up to nominal dimensions, then tracking SPC data on the achieved tolerances of the parts. They very quickly discovered that the CNC machines could easily hold those tolerances and that the manual machines had no problem getting tighter to the nominal dimension if that was demanded.
The overal weight saving was quite considerable.
This was the beginings of the TQM program and this was low hanging fruit.

The good engineer does not need to memorize every formula; he just needs to know where he can find them when he needs them. Old professor