Automated vs Manual drilling in aircraft assembly 7

[victorzv]

Hi all,

Automated drilling for fastener installation has been used in airplane assembly for long time. It is known (or believed) it provides better fatigue life. Particularly, I mean wing box stringer-skin joints.

I want to quantify the advantage of the automated drilling process over the manual one in terms of fatigue life improvement. What physical mechanism allows the automated drilling being better in fatigue? Higher interference fit? Unfortunately I wasn't able to find any reference on this subject in the Internet.

Could somebody give me an idea or reference to published data on the subject?

 

[rb1957]

michael niu's text book has a little on this.

 

[victorzv]

Thank you, rb1957!

It was a valuable piece of information. The pink text book does contain a chart with comparison of relative fatigue lives and Kt for automated rivetting and hand bucked rivets. No details though.

A question to all : Is it reasonable to assume that the automated drilling provides:

1. Closer tolerance of the hole diameter;

2. Smaller hole surface roughness;

3. Smaller variation of hole dimensions?

Are there any other factors differing automated and manual drilling from the structural point of view?

Does anyboby know about assigning smaller initial flaw size for the automated drilling?

Thanks.

 

[Stache]

Since I have thousands of hours drilling holes I can say it all depends on the person performing the drilling process. As you are well aware most mechanics/technicians don’t sharpen their drill bits on a regular time schedule and use different drill speeds. Because of this factor many holes are elongated, not straight, or over sized. Using an automated drill machine, which is serviced and maintained to exact standards, takes one of the variables out of the process. If you want a constant standard hole every time use the automated machine.

This isn't to take away from the technicians, but repeatable standards are a must in most cases of production. Technicians should be used for repairs where they can make judgment calls.

Just one man;s opinion.

 

[WKTaylor]

victorzv...

Automated hole drilling/reaming/countersinking** is always associated with automated fastener installation. The combination is what actually produces the high fatigue durability.

Note: automated drilling processes provide manufacturing elements that are "obvious" [when stated] but are NOT intuative, such as:

Part locating/jigging/clamping.
Drilling** "stiffness" for extremely low vibration.
Drilling** control of bit "angularity" [perpendicularity] and absolute location.
Drilling** force measurement [for bit-wear, disimilar material penetration, etc].
Drill** bit cooling and lubrication.
tool changes for types/diameters of fasteners..
Etc...

Regards, Wil Taylor

 

[victorzv]

Thank you gentlemen for your replies.

The picture is clearer for me now.

However, is there a quantitative measure of the automated drilling/fastener installation benefits in terms of fatigue life for joints with, say, Hi-Lok or Taper-Lok fasteners?

The only approach I heard is a smaller initial flaw size for crack growth analysis. Again, precise numbers are of an unknown origin.

Regards,

Victor Zv.

 

[WKTaylor]

VictorZv...

The "Big Picture" with automated fastener drilling, reaming, NDI and installation [with data feed-back for QC] is the extremely high uniformity of the installations which is statistically verifiable. The extreme uniformity of installation provides a very high degree confidence in the use of design allowables that were developed based on the installation precision. These higher mechanical allowables [weight reduction], combined with 24-hr-a-day production capability make automated assembly attractive. this leasve only ~1--2% of fasteners to be installed by hand [with lower allowables].

Look at the last chaper of MIL-HDBK-5 and you will have a lesson on statistics related to raw materials properties and fastening allowables.

What happens with fasteners is that installation variability can be a killer. The "A" level technician can install fasteners with a high degree of precision [consistency]; wheras the "D" level technician hits/misses the quality mark. Unfortunately the variable quality of the "A-thru-D" level technicians HAS to be factored into mechanical allowables, to a degree [Shop QA has to enforce a minimum performance level]. The automated assembly method, on the other-hand, rarely have a bad day... and data traces can reveal incipient quality problems. Now take the automated assembly process and mechanics A-B-C-D and run static and fatigue tests for fastener mechanical allowables [design numbers]. NOT remarkably, the consistency of the automated assembly, is statistically significant with high confidence: X-# shear +/- 1% [typical]. Where-as the Techs may have X-# shear +2%, -10%. Guess what: the automated assembly allowables can go to 99%-# with 95% confidence [or better]; and the hand installed rivets MUST have allowables of 90%-# for a 95% confidence.

Note: talking with the famous Dr Jack Lincoln [Mr USAF DADTA, now deceased], I asked "why the highly beneficial effect of cold working holes for fastener insallation [mod we were working on] could NOT be calculated into the durability calculations "as is", without the dramatic knock-downs he required". He put it bluntly: "failing to meet expectations on a SINGLE fastener installation [out of tens-of-thousands] could put the whole structure at risk for undetected premature fatigue failure at that location. His conservative calculations requiremnts allow for this potential quality defect. If we were building, operating and maintaining acft like "space shuttles", where every little detail was tracked and the absolute quality was mandatory, for the price of absolute weight control, then he would accept the "rosy numbers". Obviously, acft in production and operation and maintenance, are subjected to too many hidden defects to allow this." Now if You consider that we have lost (2) space shuttles to material failures, even his "rosy" assessment "of Space Shuttle quality" was not practical: things happen... defects exist, factor them into the design for safe operation.

I was on the MIL-HDBK-5 amd MIL-STD-1312 and MIL-STD-1515 comittees for several years in the mid-80s. The "magic" allowables tables for all the materials and fasteners were derived based on lots of testing [some fair some "rigged"]... and data collection... which was then statistically reducd to determine what "hard" numbers to actually place into the tables. Its all statictical "black magic" for confidence sake. The "A", B" and "S" basis columns for material mechanical allowables represent statistical confidence levels. S is spec minimum requirements. "A" is 95% confidence level [as best I can recall]. "B" is lower confidence level [~90% as best I can recall]. When I use "B basis allowables", I often DEMAND that material [individual pieces or lot-samples] be tested to verify that they meet these higher allowable values, for MY "confidence", before using the numbers

Nuf' for now.

Regards, Wil Taylor

 

[WhiteRabbit]

Good point, Wil.

Also, some automated drilling devices provide clamp-up during the drilling process. This clamp-up alleviates the deburring process at the faying surfaces of parts in the stackup.

Some of the higher end machines will detect a "harder" material (i.e. Titanium, Steel) in the stackup and will automatically adjust it's spindle and feed rate to accomodate the different material.

However, one problem that was noted is that some machines that install the fasteners use a lubricant for drilling. This lubricant may affect the fay surface seal between the parts. Usually the faying surface seal is applied and has a limited time for the the fasteners to be installed. The lube has been found to affect the sealant.

 

[WKTaylor]

WhiteRabbit...

Sealant "gums-up the works" with automated assy and is affected by drilling lubes.

I understand that virtually all Boeing automated assembly is done DRY. The precision of assy [part fit-up, hole-filling-interference and countersink fit] combined with very corrosion resistant materials and durable finishes on all detail parts and fasteners, provides adequate strength, fatigue-life, fuel-pressure-environmental sealing and resistance to environmental deterioration needed for an anticipated commercial life-span.

Regards, Wil Taylor

 

[victorzv]

Thank you gentlemen,

My goal is to quantify the difference in fatigue life between automated and manual drilling processes for Taper-Lok installation. My standard approach is to try to conduct an analysis first. The idea was to identify principal physical parameters (like surface roughness, amount of interference, sign and magnitude of residual stresses after machining, etc.) attributed to the processes and run a fatigue analysis. Now I see this approach is not realistic, because it is practically impossible to obtain quantitative measures of these parameters.

The only feasible option, I believe, is to follow Will's directions and to solve the problem using a statistical approach without specifying any underlying physical mechanism. It seems I need to perform numerous fatigue tests (material, fasteners and a range of bearing to by pass ratios are known and they are very limited) for the two types of processes. It is obvious that the results will be valid for the particular type of the drilling equipment.
Next, I determine statistical parameters of the distributions. After that I should be able to generate any conclusion about the differences in fatigue live for an arbitrary probability and confidence level.

The difficulty for me here is a definition of the probability for the ratio of fatigue lives. May I just compare fatigue lives for the two processes corresponding to the same probability of failure? Or should I obtain a sample of ratios of fatigue lives, in other words, should I have several values (each of them is a statistical average) of the ratio? In the latter case the extent of the fatigue test should be very large, I think.

Could you advise me in this aspect?

Thank you,

Victor Zv.

 

[WhiteRabbit]

I totally agree with you, Wil. It does mess up the process quite a bit because the sealant will collect on the drill/countersink cutter bit.

There are programs that are using this process with the faying surface sealant. The fasteners are installed dry but the faying surface sealant is present. It's usually sealant that has a room temperature cure of 200+ days. I cannot remember the exact DMS number of the sealant.

They need the lube because of the titanium tearstrips that are in the aluminum skin panel stackups.

Without the lube it will burn the titanium. Most of the lube will end up on the surface of the panel but it does affect the faying surface sealant.

Certain machines can countersink and drill the hole in one feed rate process.

I have seen the Brotje machines and have been quite impressed. There are minor issues with them (i.e. installing the fastener sideways in isolated instances) but in general they are a great asset using only (2) operators.

 

[WKTaylor]

Victor VZ... Now the rest-of-the-story...

Taper-Loks... Love'em and hate'em.

I worked F-15s [2-depot, 8-field], as-well-as KC-135s for +10-Yrs. These jets used a TON of taper-loks in the wing-skin to sub-structure.

AUTOMATED reaming processes [Jigs/fixtures/guides and "quackenbush-style" drill motors] are about the only way to control the process. The Tapered drills/motors and the assy under work MUST be must rigidly fixtured. The reamer must turn smoothly with "0" wobble/run-out and it must precisely advance into the hole and retreat smoothly. Using this sequence/methodology [+more that I haven't mentioned]. The reason for this is obvious in the pre-instalation "blue-pin" check: a machinist dye is applied to the surface of the TL to be installed, which is then inserted into the hole and GENTLY tapped with a gloved hand or soft rubber or leather mallet to a "calculated head-to structure depth"; then the pin is removed. If the Taper-Lok and hole are properly matched, then +70% of the dye will be removed from the pin in a uniform manner [spotty-pattern of dye remaining]... if distinct patterns [such as "rifling" or "patches"] of dye remain, then the holes/pin are NOT matched. This level of hole quality is self evident and is virtually impossible to attain by hand processes. NOT to mention, the the tooling to drill these holes is generally VERY heavy and awkward.

NOTE: The problem with Taper-Loks is as mentioned above [previous post]... a single BAD fastener installation can result in a poor fatigue life at the X-unknown location... ruining all the [very] hard-work.

NOTE: there were several technical reports published in the 1960s and 1970s describing the "best fatigue-enhancing [interference-fit] fasteners processes" available at the time [Hi-loks, Hi-Tigue, Taper-Loks in interference, high-strenth rivets, etc]... which You need to check-out. Unfortunately searching STINET [

http://stinet.dtic.mil/

] with "keywords" can be a complex problem hit-miss problem.

CAUTION SCC prone materials [7XXX-T6xxx] should NEVER have high interference-fit fasteners installed, due to the high potential for cracking or IGC.

The "world of today" has other "better" options, such as:

Cold-worked holes with interference-fit straight fasteners. Note: depending on the tooling, these installatios can be CX-to-size, ready for fastener installation in the "precision as-cold-worked hole":... or cold-worked to and under-size and reamed to final fit [especially important in multi-material stach-ups]. Hi-Loks, Hi-Tigue, Bolts, ets can be installed in these holes.

There are also sleeve bolts: tapered-bolts [similar to Taper-Loks] are matched with a softer sleeve [Aluminum CRES, etc] that has a tapered-ID [to match the bolt] and a "striaght-OD". The Sleeve bolts install in a reamed conventional hole with slight/precision clearance... then swell to interference when the taper-bolt is forced into the sleeve.

However, you may be in-luck: I think I still have contacts with Taper-Lok installation experts, who know what data is available and precisely how to-do-a high quality installation.


Regards, Wil Taylor

 

[victorzv]

Thank you Will,

Your detailed description has made clear which factors contribute to superiority of the automated drilling. (Actually, you doubled my knowledge of the process). However, I am not a process engineer. My task is just to evaluate how much of fatigue life we lose when performing manual drilling instead of automated one before the Taper-Lok installation.

You mentioned the "blue-pin" check after the automated drilling. Is it conducted for each of thousands fasteners during the aircraft manufacture by the OEM? During repairs we perform the check and the criterion according to the SRM is the same - certain print pattern and 70%+ coverage. Does that mean that the fastener fit is the same as for the automated drilling? I hope that we, at our repair facility, have an advantage over the OEM because we usually deal with relatively small number of fasteners at a time, not thousands, so that we can inspect every hole. Can this alleviate difference between the automated and manual processes in terms of fatigue?

Otherwise, we come to a conclusion that the SRM does not provide the level of safety equal to that of the OEM structure. Actually this is the issue our organization is encountering with an airworthiness body. The OEM (and USAF) SRM for the aircraft specifies manual drilling. I think, in this case there should be two DADTA and two sets of inspection intervals for the locations where automated process is used. However the OEM provides only one DADTA – that with initial flaw size for the automated process. The SRM has been in use for many dozens of years and, it seems, nobody raised a question. Until now.
Could you share your opinion on this issue? Or may be you have heard about similar cases…

As for the newer options, we are probably bound by airworthiness issues not to introduce significant design changes in the structure to be repaired. By the way, our (not completely our though) recent fatigue testing showed that Taper-Loks are still much better in fatigue resistance than a combination of the hole cold expansion and Hi-Loks installed with light interference.

Regards,

Victor Zv.

 

[WKTaylor]

VictorVZ~

Sorry, my computer was down for a planned move... getting back to action just now.

In automated production, things tend to be less rigorous than repair... I think the automated drilling, combined with the pin protrusion test is used exclusively... which I disagree with... but that's another story completely.

Have You considered Sleeve Bolts for replacement of TLs? Essentially a clyndrical OD sleeve has a tapered ID to match essentially a TL pin [they are mated together when recieved]. Refer to MIL-B-85667 & MIL-B-85667/1 thru /13 for details. I believe PB Fasteners [

http://www.pbfasteners.com/

] designed these based on their TL experience. Many bleeve and bolt-pin material combinations are available. Function essentially like a TL... but the sleeve expands as the bolt is driven-in.... and the OD of the sleeve interfers with the hole. The simplicity is that the hole just has-to-be drilled/reamed straight... a MUCH easier job than tapering to exact dimensions.

Regards, Wil Taylor

 

[victorzv]

Thank you Will,

The Sleeve Bolts are interesting.

However, if we switch from TL to the Sleeve Bolts, how can we substantiate the aircraft service life? I would say this is a significant design change. We have to prove that the service life is at least not reduced compared to the original one with the TL.

I see the only way of the substantiation - an extensive fatigue test program. Which cannot protect us from customer's arguments like:

a) you did not test all combinations of part thicknesses;
b) you did not test all range of bearing-to-bypass ratios;
c) your test spectrum is accelerated because you applied less than 500,000 cycles per specimen;

etc., etc...