A Bad Infusion - What Would You do?? (Pictures inside!)

[AndrewK]

Resin issue? No its not resin quality issue.
Can it be saved? Yes refer TeddyDiver.
Is it necessary to correct? perhaps not talk to designer as the aft face that goes all the way across and into the hulls may be doing most of the work.

If it was caused by air leaks then the laminate during the infusion would have gone translusent as it got wetout and with time opaque look would have developed, You would have also seen air racing through the transfer mesh.

The other cause of dry laminate is due to mixed in air and dissolved gasses in the resin. The higher the vacuum the bigger this problem becomes, but also once you get your process worked out the higher the vacuum the better the laminate.

Also is the problem worse on the bottom laminate? if so it could be that the feed over the top was too fast for the number of perforations. What was the perforation spacing and hole size.

Should you infuse the more crucial aft face and mast beam? yes but not untill you sort out the process. You may have to start degassing, use slower mesh, more perforations. Do you have smaller non critical parts to make?

[Ilan Voyager]

Originally Posted by CatBuilder

I do know my designer designs his beams with a safety factor between 4 and 6, so 400% to 600% more than needed on the beams. I may be able to use it elsewhere though. Looks like it would make a nice (though heavy!) table.

Would love to hear what some of those infusion experts here think of the quality, too.

My concern with extra patching is that it does nothing to address the real problem: Possible poor adhesion to the core. I think the glass would be fine, since it's a lot of Uni where the problem is. However, if there is some peeling effect and something that is not well stuck starts to spread, my beam will unravel. Or maybe I'm just being paranoid. I just don't have the experience to know.

A structural engineer works with security coefficients ranging from 1.3 to 2.5. That depends of the data known about the involved stresses and the quality of the shipyard. On small boats where most of the engineering is made smelling the wind added of common sense, this NA uses a coef of 4 to 6 because he knows that:
1- by experience the real stresses in multihulls are 1.8 to 2.5 the calculated static stresses. Static stresses do not count the stresses induced by accelarations and decelerations, nor vibrations. On a small boat except for competition there is no need of FEM and other niceties. There is also an empiric knolewdge.
2- by sad experience the NA knows that the quality obtained by the amateur builder range from the best (rare) to the very poor (common) so he has to take a big security coef, so the poorly built boat has a chance to survive a few years.

The real mean security coef will be around 2, do not count on 4 or 6. It would be ridiculous to design a structure with such high coefs, except when you have not idea of the real stresses (that happens...).

When an infusion has gone wrong and that the piece is a structural one. No way: you have to make it again. Betting is forbidden in structural pieces: that can kill. Infusion is for pros whatever say the equipment sellers and promoters of the system. AndrewK is very right. Probably too much vacuum without degassing. Some foams are deadly if the vacuum is too high.

Train with non structural pieces. It seems you had similar problems with the cylinder mold plywood before: you need training.

You can fix a bathroom floor, or the closets doors with some resin but not a beam component. The beams are the the most stressed pieces of the cat along with the the hulls at the beams attachments.

[Pat Ross]

I go with Ilan Voyager and gonzo's thoughts. You need to redo it and if you can use the panel for a cabinet/table that is good.

Pat

[redreuben]

When the breeze comes up, the sea is difficult and the family is on board, what then ?
If you wish to fully enjoy this boat you must be able to trust it !
Remake the piece and use what you can elsewhere.
RR

[idkfa]

While Ilan is right, it is still not the best answer or an engineered one.

The structure could be redesigned to account for this weaken part of the aft box beam.

It may be the most (cost & time) effective to do an additional laminate around the bottom edge. Where it fails first.

You need an engineer to have a look. Perfect your process on some smaller pieces still, come back to it, I doubt it is a lost.

IMHO

[gonzo]

I think you can treat it like a damage repair and fix it.

[CatBuilder]

Thank you all for the thoughtful responses.

The only thing I take a little objection to is that I need training. Sure, we can all learn and improve our processes, but I think I can handle infusing a small panel like this beam. The very first picture in this thread shows the result of my hull infusion. Perfection.

This is what the hull infusion looked like when I did it:



That's the whole hull. 15 meters long. The vacuum line is at the keel on this inverted hull. The other side away from the camera has an infusion going on as well at the same time. I infused it all in one shot and got the results you see in the first picture in this thread. Perfect.

I've also infused both rudders, both dagger boards and another beam. All perfect.

This beam situation was an anomaly, but there is one thing that was different in this infusion as compared to my other ones. I had a vacuum pump failure and had to switch the pump out mid-infusion.

There was a brief time where there was a backup of air into the laminate from the vacuum side toward the resin side and the infusion was going very slowly for about 10 minutes. I noticed the slow rate of flow and decided to switch the pump. Switching pumps brought the flow rate back up to normal.

I had dirty oil in the pump (milky white, actually) and I think it wasn't able to pull the full volume it normally does (6CFM), although it was pulling down to full vacuum.


ANDREW: The holes in the foam were as ordered from Divinycell. This foam is Divinycell H100, perforated. I believe it is 6cm spacing, but I am not recalling perfectly. There was one difference in flow rates between top and bottom, but it was just the opposite. In one part (a part that came out very well), the top lagged behind the bottom. It worked that at times resin came up through the holes before the resin advance line on top made it to the same area. This only happened in one area, but that area came out well. Also, I use a resin stop (stop the flow media/mesh 2" from the vacuum tube) to allow faster and slower areas to catch up to each other. This worked as well as it normally does and allowed that slower top to catch up to the rest of the infusion. For degassing, I normally run the pump at full vacuum for about 30 minutes to an hour before starting the infusion process. Is this enough time?



If I looked at fixing this beam (preferable), what would the repair joint look like?

The beam, essentially, is a bunch of uni going across it the long way. The idea is to have uninterrupted uni strands going from one side of the beam to the other. If I cut into those uni strands (ruining their strength), what type of joint would I use to keep the same strength as the original laminate?

If that doesn't work out (because it's uni), then I will just use this one for some other part of the boat, despite all the extra weight it will add.

[TeddyDiver]

Originally Posted by CatBuilder

The beam, essentially, is a bunch of uni going across it the long way. The idea is to have uninterrupted uni strands going from one side of the beam to the other. If I cut into those uni strands (ruining their strength), what type of joint would I use to keep the same strength as the original laminate?

What comes to my suggestion it's not to cut the strands. Grinding them means the strand layers become tapered in the sides of compromised area. New layup goes over the edges and after curing the "patch" edges are grinded again so that there's overlapping enough to overcome the strenth issue. The result is a bit thicker laminate around the patch..
BR Teddy

[CatBuilder]

Something else also comes to mind after someone PMd me with an idea and reading Teddy's idea.

What about grinding off the bad part of the laminate along the uni strands?

For instance, just grind off the entire length of the beam, but only 12 inches in from the edge, picking one uni strand (and the rest of the beam) to leave whole? Maybe that uni strand and a few of its neighbors get a taper, like Teddy was talking about.

Then, I could just hand laminate a new, full length uni lamination along that 12 inch strip that was ground out, maybe with some triax in there for thickness to match spec.

This beam already will have many layers of biax around the edge of it to connect it to the other parts of the box. It is the core adhesion that is the questionable situation. If I can lay new uni (with the added triax), I'll be in good shape, but have a little bog fairing to do since the hand laminate will be a bit bulkier than the infusion.

Or... I could even infuse the repair strip, one side at a time.

Anyone see any problem with that idea for a fix?

[P Flados]

From your picture, this is a vertical surface on the forward side of the rear beam.

The primary loads I would worry about for this beam would be shear (for vertical bending), and torsional loads (twisting).

The amount of fiber specified was probably more to achieve stiffness than for raw strength.

This surface has a lot of triax as these fibers carry shear loading to support vertical bending loads and the diagonal strands are the fibers that carry the primary stresses for both shear and torsional loads.

The Uni near the top and bottom are all about supporting vertical bending loads. I would expect the top & bottom of the beam to have a higher percent uni fiber as this maximizes vertical bending strength.

I would not worry "core adhesion" for a beam like this. The loads are transmitted directly from one fiberglass surface to the next. The transitions at the corners however are very important and need to be good quality.

Any corrective actions should be based on an assessment of reduced strength / stiffness.

If you think you are at 90% of full, you probably exceed what the designer took credit for. No corrective actions are needed.

If you are 75% - 90%, you are probably OK, but your confidence level would be iffy at best. A single addition layer of diagonal only fibers (+45 /-45) with a thickness of around 25% of the original total triax would be the easiest "fix" assuming that your concern is the original triax. With this surface focused on torsion / shear, the extra needs to be full height, but really only needs to include about 1/2 beam height to either side of the area of concern if it is not full width.

If the original uni near the top an bottom is a concern, I would consider adding a narrow width extra strip of thin uni that wraps around the corners at the top and bottom.

If you think you are at less than 75%, you start considering removing some existing glass. Since the loading of this surface is mostly shear / torsional, a repair could be made to only include removal of suspect areas with either a full overlap or a tapered overlap extending around 1 beam height over existing fiber. Adding extra lengthwise uni wrapped around the corners top & bottom should probably be either full length, or with a good amount of overlap on top of good fiberglass.

Any new glass can be infusion, but you may want to put your resin under vacuum before the infusion to avoid any possible off gassing and you probably do need to make sure your vacuum pumps are in good shape before you start to avoid interruptions.

[cavalier mk2]

I haven't played with infusion but have you thought of taking a cue from your hulls and using gravity to assist? You might consider making a test panel with the layup table in a inclined/vertical position. With the vacuum there should be no problem with lamination pressure and the epoxy might wick/flow more easily. I think the repair is possible but would it be more work than a new laminate. The materials loss potential of this process keeps me friends with wood but when it works the results are beautiful.

[TeddyDiver]

Originally Posted by CatBuilder

Anyone see any problem with that idea for a fix?

Not any major.. what comes to mind is to grind only along the lines. The section btw should shear away like a brasilian waxing
Warming helps if it sticks really well, but then it's also well bonded so there's no point to remove it at all.. IMO

[TeddyDiver]

Just to add.. In the picture it looks like it's the upper part of the beam. That means it's mostly compression loads and they are shared loads with the cabin roof (assumed it's quite similar cored uni construction). So I would first check in a worst looking place if the bonding is reasonable or not before jumping into more conclusions.. It might be, and most likely is what someone allready mentioned, some foamy epoxy so it might not compromise the bond..

[P Flados]

My previous post was before I saw the pic in post 24.

Your statement "This beam already will have many layers of biax around the edge of it to connect it to the other parts of the box." is also important.

First I would at least try to assess what you have. Lay it flat. Go to a good area and see how much pressure it takes for a small but quantifiable deflection. Go to the middle of the bad area and repeat.

If you can barely tell the difference, I would just consider increasing the thickness and/or width of the "biax around the edge of it to connect it to the other parts of the box".

Note that where the connecting biax overlaps good regions there is little or no concern in the first place.

If the area is noticeably softer, you have to consider what kind of loading is on all of the diagonal fibers that cross the "red zone", especially toward the middle of the beam. At the very middle, some of the iffy area looks to be above your red zone.

If I were grinding, I would consider a more triangular area with a 4 to 1 taper (or better) along the edge of good fiber. Lay in triax to restore original thickness across the middle and then grind down any excess thickness over the taper. Go with extra biax width and/or wide biax plus some original width uni when joining surfaces to make the box. For extra confidence, you can consider at least one layer of "connecting biax" to extend completely over the patched region onto good glass.

[rberrey]

Cat another qoute form Mr. Horstmans book.Structurally speaking if you double the thickness , you obtain approximately 4 times the bending strength and approximately 8 times less deflection. Maybe you can add another layer of foam as well as glass, just a little thicker and heaver beam. Rick