Originally Posted by
MechEngr
This specific automation claim seems a bit suspect. I agree that including automation as part of the initial design process is a good idea and delivers the least expensive adoption. But there are many ways to automate an operation. The usual impediment to automation is wide ranging variability. If that was the case it would not require a design change to focus on lowering the variability. Manufacturing is always welcome, and should be encouraged, to decrease variation to make a more consistent product.
Likewise jigs, fixtures, and, to a lesser extent, order of operations, can also occur without a design change.
Most remaining problems flow from the lower level variability. It's possible to make a design that is more costly, but is also more tolerant of that variation. This adaptation is not is a specific impediment but if the variation is reduced the adaptation should also respond to automation. This adaptation is often things like peeled shims that are customized based on making measurements of the varied gaps. If the underlying parts are consistent the gaps will be consistent and it may be possible to eliminate the much or most of the need to peel shims.
Perhaps there is some specific process that requires human flesh to perform the fastening work?
If I were to place a bet where things are difficult to automate it would be routing wires/cables - but that would not involve the image of a dirty trashed factory.
One big difference is that Airbus is a multinational company that also serves as a show-piece, supported by multiple governments. Showpieces can have money spent on cosmetic dressings.
I saw this in a facility that assembled medical equipment: while there was no need for sterile conditions and no one was wearing hair coverings, the workers were all issued lab coats. Perhaps it was for esprit de corp, but under the veneer it was a rather unproductive and relaxed operation. My favorite was a case of simple magnets to hold a cabinet door closed. Two different versions of the same model had arranged them differently. So I went to the floor and asked the assemblers. Two were working the station and when I pointed the differing requirements, they each claimed each version was right, and then looked at each other because they had been doing all the assemblies different ways, inches from each other, and never noticed the difference. Now - one might say - it doesn't matter, and I suspect that is right. But there had been an engineering drawing change to make them different. So, what problem in two otherwise identical cabinets, caused engineering to find one assembly unacceptable? Turned out - engineering didn't know. But they had lab coats too.
The main factor isn't to make some operation more easily automated, but to eliminate the function entirely. Doing that does require design change and, with it, some form of re-qualification. One company I worked for made fighter planes and at some time they discovered it was less expensive to buy a 500 pound slab of aluminum and mill off 470 pounds of it (or so) rather than assemble the part from 30 or so pounds of aluminum in the form of 50 little bits of sheet metal and a few hundred rivets. Of course if one of those brackets got damaged, one could grind off the rivets and replace the bracket for a few dollars. In the one-piece version a completely custom one-off part would have to be made, the remains of the stub removed, and the repair part riveted into place. Maybe pay now, maybe pay later.
Obviously, changing the length of the fuselage from series to series and model to model must have required re-qualification, so redesign would have been far easier to add in. Not sure if Boeing put that effort in or skipped it. It would be interesting to know if they did. They have made other, more significant structural changes and such changes would have saved a significant amount of money..
While much of the above is true, it's hard to see much relevance to the 737.
Round about the same time the 737NG was on the drawing board, I was involved in the initial development of a regional jet (we got as far as making presentations to a whole bunch of airlines before the project was cancelled). One thing that sticks in my mind is just how much of the design work was centred on manufacturability - the processes, techniques and automation that would go into the building of the aircraft.
I stand by my view that trying to retrofit all those manufacturing innovations to a 30-year-old programme would have been a non-starter, and clearly Boeing decided likewise.