Aluminium boat with no welds?

No need to teach me, I've been riveting aircraft structures together for more than 30 years!

During that time the whole industry shifted away from annealed bucked rivets to self setting fasteners like the Avdels. I doubt you'll find a bucked rivet in any modern aircraft, they are virtually all bonded, friction stir welded or fastened with self setting fasteners.

With that change has come the virtual elimination of "working rivets" as you describe. These modern fasteners stay very tight for many years, making for far less maintenance than back in the days when bucked rivets were the norm.

 

Technically that is correct.

However most marine applications use 5000 or 6000 series. The 5000 series is non-heat treatable. These are unaffected by the heat of welding, thus have no drop in mechanical properties. If the alloy is tempered, ie strain hardened, then there is a slight drop from the unwelded to welded, in the HAZ, of the mechanical properties.

In the 6000 series, which are heat-treatable, there can be as much as a 50% drop, depending upon the alloy and temper chosen. But the point of using a 6000 series is that it is an extrusion; that is to say a shaped bit of ally, shaped to your requirements. Therefore when you design the structure and select or design an extrusion, the higher stress regions are located in the unwelded higher strength regions. Where as the welded joint, the HAZ, is located in the lower stress regions where the lower as-welded properties do not affect the performance of the joint..

No one heat treats post welded 6000 series to obtain the same properties as before in boats. It requires a significant amount of control, and a large oven (as Jermey notes)! Also depending upon the type of temper, as the artificial aging process is also rather complex.

 

Ad Hoc,

According to Pollard:

5086-H116 Before welding: Tensile (psi) 40000, Yield (psi) 28000
5086-H116 After welding: Tensile (psi) 35000, Yield (psi) 19000

From the web;

5059-H321 Before welding: Tensile (psi) 54000, Yield (psi) 39000
5059-H321 After welding: Tensile (psi) 48000, Yield (psi) 23000

When trying to maximize strength and minimize weight the ideal is an unbroken (unwelded) primary structure out of 5059-H321. So each one of the structures are looked at to build in such a way as to minimize the exposure to welded joints in areas needing high strength. So preliminary FEA should be used to strategically locate welds, and design to try and eliminate the number of welds.

----------

I did a design study for an very large aluminum front landing craft door and how it should be designed and built. The strongest design was welding everything together except the top floor plates (only two needed and butted over a frame flange) which were riveted using sealed aluminum rivets.

There were many frames + flanges to which the floor plates were bonded and riveted.

Bonding adhesive was used to secure the plates to the frame flanges (24 hour cure), and then where riveted, starting on the inside of the door working our way outwards, alternating between the two plates.

The bonding also filled the gaps on the slight plate separation. We had a drain plug at the top rear of the door (Large pipe plug), that we used to remove all of the chips from drilling the rivet holes. The door is normally water tight and the plug was used for inspection.

Essentially, except for the center butted plate area, there were two rows of rivets over a bonded frame flange. The bonding prevented the rivets acting in shear, and two rows gave excellent clamping force to the bonding.

So I was impressed by the proper application of rivets and bonding as compared to welding. If done correctly, and the bond is proper, a two row riveted joint can be very strong, water tight, and be effective in fatigue.

Mark

 

Mark Catt

I’ve no idea who or what is Pollard, sorry. I also don’t understand imperial units too. However, I do recognise an error in your interpretations of the figures. The value of “tensile” is the UTS. In the attached, the values of “yield” are shown as location “A” on the stress-strain curve. So from O to A is the “yield”.



The values of “Tensile” are those located at “R”. Well beyond yield. One never designs to the UTS, always the yield.

So, back to your 5059. You quoted H321…I’m not sure why you would select a very highly strain hardened temper, other than its “wow” factor of having a higher yield to the annealed temper.. It is very difficult to fabricate, it would be harder to rivet and it very prone to exfoliation corrosion unless supplied with an ASSET cert (ASTM B928-04, for 5059) and you are careful with the edges.

If you look at the unwelded properties of your 5059, the O temper is 160MPa, the strain hardened H321 temper is 270MPa. So there is the wow factor.



However, when looking at the as-welded strength, no surprises that the O temper properties of 160MPa remain. But the H321 is now also 160MPa, owing to the welding heat annelaing the alloy.



So your selection for riveting can take advantage of the higher properties, but it comes at the expensive of production…and possibly future warranty for corrosion.

 

Ad Hoc,

Thanks for the graphs.

I could have referenced several sources for the as per welded reduction in yield but I selected Stephen F. Pollard's numbers because many builders of smaller boats are familiar with his book Boatbuilding with Aluminum, which I though was more of the audience for this post.

I provided both Tensile and Yield in the same manner as presented by Pollard. Basically, to get a general visual for the curves.

----------

The overall theme of my post was to illustrate the difference between unwelded and welded strength. And to point out that, in structures, one has latitude as to the use or location of a weld in overall design.

The example of 5059-H321 was to show the advancement in alloys. It is an option for inclusion on a boat, and has considerable strength when not welded. So the weakness to the O temper is only in the location of the weld.

----------

The example of the door I presented was a classic comparison of weld versus riveting. The welded only approach would have had plates spaced apart centered on the frame flanges, and the structure was to be water tight, so each frame spaced floor plate would have to have been continuously welded to the frame flanges. The riveted approach had only two plates for the entire ramp, and was riveted and bonded to the frame flanges.

I agree welding would probably be faster than the rivet + bonding approach, but the rivet approach, for this particular door type design, provided good strong results in attaching the floor to an already welded base structure, which is a common challenge for aluminum landing craft ramp design.

That is, the riveted floor with its unwelded strength could tie together the remaining welded door structure.

So we have options for how to connect structures and when and where to weld.

For the ramp floor the original plans called for 5086-H32 plate, of which there were different tread plate options available.

6061-T6 was ruled out by the team, and this left the question of should we use 5059-H321. But it was not a good match for the floor requirements, not many tread plate options, and as you pointed out, costly.

Mark

 

MC

But those “tensile” properties are misleading. They are NOT design values. So, I’m not sure why you posted them?

I’m not sure what you mean by this…advancement? It is just a rolled alloy to improve the mechanical properties. This is nothing new and has been done for decades.

If you’re just riveting why rule out 6061…surely your SOR remains the same?

So welding versus riveting..ok, nothing new there. But the SOR could yield an all together totally difference arrangement too, not riveted. A designer has many balls to juggle. The mechanical properties of the material is just one of them, it is also not the first one on the long list, just one of them.

Each “product” has a different objective. How you arrive at the objective, is up to the designer. So long as you achieve it, is all that matters to the client.

 

Ad Hoc,

In reference to your questions:

To me 5086 is old and AA5059 (offered in 1999 -2000) is new. However, its adoption is the advancement I am talking about. It is seen as a viable option and is offered by more manufacturers, at least here in the US.

We could not back up the 6061 with real world data for floor corrosion resistance as compared to our proven track record with 5086-H32.

----------

I guess not being in aerospace, the actual performance of a bonded two row riveted lap joint was very impressive to me for maintaining base metal strength and circumventing weakening due to welding.

Based on my ramp design experience, I am going to explore more riveting+bonding options in replacing welds for marine.

Mark

 

AA5059 is a different alloy, btw. It was developed to be better/different to 5083.

However, either I’m being very thick, or are you just saying, here is a new alloy? Just like 5383 is a new alloy variant of 5083? If so, I’m still struggling to understand the point being made as both are available and both have their pluses and minus for selection, not all related to strength too?

6061, or rather the much better 6082 has acceptable corrosion properties, if used correctly. 6061 has 3 times the amount of copper content than 6082, and is no longer recommended by Classification societies, for this reason.

A successful design, no matter what it is, is a success by reviewing all the inputs and treating them all on merit against the desired outcome. Production and careful placement of dissimilar metals etc is all part of this “design” process.

(And on that note, FYI, 5059 has 2.5 times as much copper content that 5086, 0.25 to 0.1% ratio.)

 

Very interesting discussion guys. It's nice to find a confirmation that bonded&riveted plating is finding its way in this industry too. Now I'd be interested to extend it from technical to economical... In your experience, how do the costs of various solutions compare to each other? Welded (friction stir excluded as out of reach for small builders), vs. various riveting systems vs. bonded plus riveted?
Cheers!

 

For small builders I believe that the combination of epoxy bonding (with suitable surface treatment - Alodine is a fairly easy process to use in a small workshop) and self-setting rivets of a suitable type (monel or stainless), wet riveted, i.e, with the rivets dipped in chromate or an alternative anti-corrosion/sealing agent as they are applied. This is a basic technique that has been very well proven in seaplanes and other maritime aircraft for a few decades so should work well for an alloy boat hull.

The primary advantages are the freedom from any weld distortion, the ability to make the frame and hull self-jigging (pre-drill the holes for fasteners in exactly the right place so that the structure adopts the correct shape as it is assembled - a common technique used on kit built alloy aircraft) and best of all the rapidity with which the frame and hull can be assembled this way. Sure, the epoxy needs time to cure, but the rivets will hold the structure together whilst it does, allowing quite a lot of assembly to be undertaken in a single build session.

The main disadvantage with this technique is the need for a reasonably well-controlled build environment (warm and dry) and the need to prepare the bond surfaces with some fairly obnoxious chemicals in order to get a really good bond. It's also slightly more expensive than welding (Hysol adhesive and stainless/monel fasteners aren't cheap), but doesn't require aluminium welding skills or the investment in welding equipment. This latter point may be significant, as the skills required to ensure high quality joints and fastening with a bonded and self-setting rivet approach are generally low - pretty much anyone can learn enough to make a really good joint in an hour or so of training.

Jeremy

 

Nice summary Jeremy.

I would also add that for more “commercial” vessels, it gets harder. Since there are minimum rule requirements, regardless of the alloy and method being used. So where rule says min allowed 4mm, being riveted wont change that, so any gain, is not realised. Also on larger vessels, it gets more difficult to rivet 10mm or 20mm plate!!

As I noted before, we used on 2 deckhouses, with success. But these were relatively small, less that 10 tonne each, and not major structure, as they were deckhouses, not superstructures.

I thinking rivet/bonding has its advantages in some applications, but those advantages begin to diminishes with the larger structures, especially if commercial build.

 

I agree, as soon as you get to boat sizes where you're dealing with section thicknesses of 4mm to 5mm and above then riveting and bonding starts to get too difficult and welding has to be the better method. The advantages of the bonded and riveted method is really only on structures where the material sections are thinner, which restricts it to use on smaller boats. If the material is really thin (say a skin that's less than 16g/1.5mm thick) then large area bonding becomes more viable on its own as a fastening method, although this then requires big flanges on frames and some means of temporarily holding the skins in place whilst the adhesive cures (vacuum bagging works well). I once helped build some 12ft long seaplane floats with 20g 6061-T6 skins, bonded and riveted to 20g 6061-T6 pressed frames and they were remarkably tough when finished. I have no doubt that the same technique would work OK for small, high speed boats.

Jeremy

 

Jeremy, do you have experience with these Avex (by Avdel) rivets in marine environment:
http://www.speedcomfly.com/sito-ecommerce/file_info/Avex_1604_mm.pdf ?
It is made of 5052 alloy, which should make them very suitable for marine use. I'm still trying to figure out their cost.

 

I've pulled a few thousand of those very rivets on home-built aircraft, but not on a boat. They should work well, but they have limited shear strength, so the joint design needs to ensure that you can get the strength you need by using multiple rivets. This is exactly how riveted aircraft are built, the difference with hollow rivets like this is that you need to use more of them than a solid bucked rivet to get the same shear strength.

Chris Heintz has done a lot of work on the use of Avex rivets in kit aircraft construction and developed a neat way of getting a better surface finish on skins. This article (and maybe the first part that deals with bucked rivets) may be of interest, as Chris discusses the relative advantages and disadvantages of the two types of rivets

http://www.zenithair.com/kit-data/ht-87-1.html

My first kit aircraft used around 16,000 Avex rivets, virtually all of them were 1/8" (3.15mm) diameter, set with a pneumatic rivet puller (very quick and easy to use and pretty cheap to buy). I only recall having one bad rivet set in the whole build, which was purely a function of me forgetting to turn the shop compressor on and failing to notice that the pressure in the tank had dropped too low to properly work the rivet puller.

Jeremy

 

"It does nothing for my distrust of airliners though...."

Fear not , today much of the structure is simply glued together , not riveted.

A big advantage of driven rivets is they can be "fattened up" a bit to go into old holes.

The rivet must be a tight fit and many old structures will have holes that are bigger than spec, drilling one size up is a cure .

But simply fitting the rivet to the hole diameter is strong (carrier aircraft use the fix) is cheaper /faster and the parts are usually at hand.