According to a report from Texas A&M, the loss of the keel and capsizing of the "Cynthia Woods" was due to a design flaw rather than faulty repairs. Of course, they may have a vested interest in pitching it this was, but anyway that's probably at least partially true. I read a short story on it here:http://www.salon.com/wires/ap/us/2009/07/17/D99GB7RG4_us_missing_sailors_report/index.html
Report on Cynthia Woods
- Thread starter John
- Start date
- Status
Of course A&M is a very interested party. The investigator they quote is the assistant general counsel for the university and a probable witness and interested party in a lawsuit against the builder. The report is a response to the Coast Guard investigation (a disinterested if not always wonderfully competent investigative body) who found that the loss was largely due to faulty repairs done by University staff and students.... Anyway, on first blush I generally assume paid for testimony in a lawsuit is grossly biased. So, I don't see any reason to give this report much credence.
--Tom.
Wow. Talk about a vested interest!!!!!
Let me see if I got this right- the cop blames me for running a light, but a hire an investigator (myself) to determine the light was faulty??
I understand the university's position (defending the repairs they made previously), but if the findings weren't as a result of an INDEPENDENT investigation, I think their findings are worthless at most, skewed at the least.
Their findings are completely opposite of the coast guard investigation? The repairs had "absolutely" no bearing on the case???
Why am I not surprised their associate counsel (read:lawyer) not only released their findings, but was also one of the "investigators"?
I'm just a lowly sailor...can't wait to hear Roger Long's take on this.... Roger?
cup
Let me see if I got this right- the cop blames me for running a light, but a hire an investigator (myself) to determine the light was faulty??
I understand the university's position (defending the repairs they made previously), but if the findings weren't as a result of an INDEPENDENT investigation, I think their findings are worthless at most, skewed at the least.
Their findings are completely opposite of the coast guard investigation? The repairs had "absolutely" no bearing on the case???
Why am I not surprised their associate counsel (read:lawyer) not only released their findings, but was also one of the "investigators"?
I'm just a lowly sailor...can't wait to hear Roger Long's take on this.... Roger?
cup
A story about a report is as useful as a story about anything else when no tangible fact are shown. N. Korea detonated a nuclear devise recently. Their claims and the seismic effects produced by the event haven't yet been reconciled. The report on the Cynthia Woods event made by the CG and the report by A&M will have to be reconciled. At this point I would lean towards favoring the CG report.
All the various keel failures for which photographs and other information have been made publicly available appear to be in accordance with applicable standards and accepted industry practice. The standards describe minimum structural characteristics at a few specific points as guidelines for the rational engineering and design that it is expected will be applied to the structure as a whole. Accepted industry practice is often to blindly adhere to thickness and specifications in the standards without conducting any real design and engineering. Justifying some of these engineering absurdities by pointing to a recognized standard that they comply with at a few points is shameful and deceiving.
As I said on one of the Titanic broadcasts, “Any idiot can make something strong enough. Engineering is about making structures light and economical enough to be practical in the real world.” There is a philosophy and ethic about these trade off’s that is as important as the structural calculations. A fiberglass hull that is one inch thick instead of half an inch will obviously have much greater reserves against puncture (assuming the same lay up schedule). However, it’s weight will dramatically effect performance and the boat will probably be too expensive to survive in the market. Acceptance of a possible few additional hull breeches in the fleet is a rational and reasonable engineering decision. However, the performance penalty of making a keel stub, located right next to a lead keel casting, twice as thick, could never be practically measured. A fraction of an inch shaved off the top of the lead casting (an expensive material itself) would completely offset the performance penalty. The additional labor and material expended at this point would also be insignificant to the overall cost although they could be measured. Groundings and haul out miss handling are highly likely and foreseeable and the benefits of this extra structure are almost certain to be realized whereas the floating object that extra hull thickness would justify may never be encountered.
The lawyers’ various experts will be throwing calculations, numbers, and standards back and forth at each other wracking up huge bills but it will be hard to find a shred of common sense or rational design analysis anywhere.
Do you recall if the CW keel was an all lead structure or a steel strut plus bulb? I kind of though it was the latter... I don't suspect it makes much difference in your cost analysis. But, it would make a big difference in terms of how energy is transferred in a grounding. Also, bagged boats have much better fiber to resin ratios which result in thinner and potentially more brittle laminates and particularly in single skin areas like the bilge they could result in very different properties than traditional lay-up. Scantlings are a mix of rules of thumb and engineering calculations and admonishments to do testing. When boats start to evolve beyond the historical envelope it is likely to be risky to extrapolate from the old rules. It may be that the scantlings are invalid. It could be the build was flawed. But, of course, if there was gross negligence during some repair it may be that the design and build were adequate and we need to think about testing and repair techniques. Tragically, I think all the noise from the partisan expert witnesses is going to make doing a good timely job of discovering what the key lessons are harder than it ought to be.
--Tom.
I just skimmed through it:
http://tamus.edu/offices/communications/cynthiawoods/finalreport.html
and, at first blush, it looks like a pretty decent job. It certainly come a lot close to passing the straight face test than the one produced by the Coast Guard's expert.
I haven't gotten ahold of a copy the ABS rules for yachts to try and figure out why one report says the thickness met them and the other says it was 1/3 of what ABS requires. However, as in my reply to TimCup, that's not actually the essential issue.
Look at the floors in the report photos. How could any rational designer or engineer pass up a chance to transfer the loads directly into those large members which could distribute them out into the hull through their tapered ends? As a designer, those pictures make me feel like most people do when they look at war photos of broken bodies. The waste.
As I skim through the report and the ISO rules that I have (Larsson & Eliasson) I think they may be "confusing" the skin thickness of the panel with the stiffener requirements... Their results are not reasonable. 3.11" thick solid panels are not the norm in ISO boats of this size but the stiffener height needed to get the I's sufficient would, indeed, be about 3 inches.
--Tom.
I just skimmed the Coast Guard's report done by Bill Richards Design and an independant engineering firm and I've got to say I disagree with you. It takes some reading as the actual report starts somewhere around page 200. But, it seems like a very through bit of work and includes good documentation of the very serious damage done to the boat in its hard grounding (the keel separated from the hull and the hull was deformed-- the picture is brutal). Texas A&M not only didn't bother to get a survey, they didn't even get a professional in to supervise the work their kids did... OMG. It may be that the design and or construction was faulty but I'm having a tough time with the T A&M report that just blows all of that off... And, Richards re-worked the ISO numbers and they were in line with both the design and build. Oh yeah, it appears that somebody drilled out the out the limber holes in the stiffeners significantly weakening them after it left the factory a fact missed by the A&M team... And we get to read Richards own words rather than the words of an expert witness re-written by a lawyer who's job it is to cover the butts at A&M. So, for my money (ok, no money) the independent report of from the Coast Guard is way more substantial and believable than the Texas A&M thing.
--Tom.
I'm not sure what you're saying. Do you mean the bolts should have gone though the floors (stiffeners)?
--Tom.
Overall, I pretty much agree with you. I wasn't clear before that I'm speaking very narrowly to the question of the keel attachment design. That's what interests me as a designer. The questions of post grounding inspection and repair are management/operational issues that I'm just not dealing with right now.
The A&M report is a rehashing of Dobroth's work and obviously suffers some in translation but I see a more solid analysis of the keel strength through that haze. I've seen too much of USCG investigations to start with the conclusion that someone is truly independent just because they are working for the government. Richards seemed to be trying very hard to justify what is to my mind a terrible design. Engineering calculations are like statistics. I have no doubt that you can put numbers together showing that keel to be adequate but that's not the full story in engineering. Dobroth appears to be putting some solid analysis behind the obvious.
The weakening of the keel attachment by the grounding is part of the indictment of the design to my way of thinking. I’ve done very little fiberglass designing but it is a cardinal rule in steel structures that loads go to, or are aligned with, framing members and not to panels; even though steel is a much more uniform and predictable material than fiberglass.
Yes, the bolts should have gone through the floors or have been directly tied to them in some way. Think of how differently the bottom panel would have behaved , even with the inter-laminate sheering and other weakening caused by the grounding, if the keel loads had been fully transferred to the floor members. These extend well out into the hull and end in nice tapers. It would have been necessary to tear these members, with their vastly greater bearing area, clear out of the bottom for the keel to come off and a catastrophic breech of watertight integrity to have been created.
Part of good designing is anticipation of failure modes and response to overstress. If a hull I designed was going to fail due to the extreme abuse of grounding and towing, I’d like to see stress cracks or even complete separation out at the ends of the floors. Detection would be easier and I would know that the full strength of the basic shell had been exceeded. The keel attachment should be a backbone support; not a weak link.
Such a design in this case could have had virtually the exact same weight and vertical center of gravity. The cost difference would have been insignificant in the overall budget for a boat of this type. It might still have required major repair after the abuse this vessel evidently received but the weakening would have been more easily detectable. The damage would have been more likely to be manifested by increased flexing and working than by the sudden and catastrophic failure later on.
AFIK, Richards is an outside expert who's day job is designing production GRP boats. I don't know if he's independent but at least he's taking a pay check from a disinterested party. I didn't get the feeling he was saying anything nice about the design except that it was plausible to argue that it was both designed and built reasonably closely to ISO standards...
Yes, he does. And this is where the fog of "expert" witnesses comes in. He's certainly made some pretty pictures on his computer and makes some reasonable sounding arguments and run some numbers and he may well be right on with them. However, he's being paid to provide evidence which makes it very hard to take his report at face value. I'm especially worried that he totally discounts a grounding that might well have resulted in an insurance company totaling the boat by saying "hey look at the woven roving still stuck to the floors. The repair is still in place so it must be good." Since that is utter BS, I have more than a little trouble believing the rest of his analysis. And, unlike Richards he doesn't show us his work.
Okay. IME, GPR boats are never constructed that way. The floors aren't designed to take the loads from the bolts. But, I can see your argument that they might better be. I note that the hull did fail at the ends of the stiffeners and had a massive (much larger than the hole) delam up the side of the hull. The problem is not so local as Dobroth insists.
Yes, that seems right. The thing is the signs of failure were probably there. They didn't bother to have anybody with any knowledge of GPR structures look at the thing after a grounding that split the hull from the keel. I suspect if they'd just tapped the hull with a coin they would have found the voids. They are huge. And the result was a probably a very flexi structure and that caused the ultimate failure.
Having said all of that. It's a marginal design in lots of ways. I hate it. And, it looks to have been built to a low end standard. The keel is just about the ugliest foil I've ever seen on a race boat. The backing plates were too small (and needlessly). And it is possible that the delam was a result of poor initial construction -- I've seen Kevlar fail to bond well and the failure might be between the kevlar layer and the glass. But none of that excuses Texas A&M for not getting the POS surveyed after they put it on the reef. Particularly in the context of carrying their students that's just criminal.
--Tom.
Totally agree with Tom.
.... and besides a Cantilever Keel root is a MORE serious/complex design than that carrying a simple factor of safety of 2.0 (ABS std.) for simple 'shear' concerns, usually requiring a FS of 4 for 'offshore' boats if you carefully analyse and 'backcalculate' those designers who are reknown for 'bluewater' designs ... and that is notwithstanding fatigue concerns. Without access to the calculations of the finite element analysis done by the 'plaintiffs' engineer and without actual measurement (testing of actual 'coupon') of the material for 'actual' strength values, etc., it appears that the 'engineering report' is tailored to be 'typical lawyer speak' ... always doubtful in good objective forensic analysis/reports. Without such reporting of 'actual' (tested) values any court argument can easily be challenged and discounted. Fiberglass is not a material with consistent strength values as there are many factors such as glass to resin ratios, consistency of layup/voids, areas of untowards stress concentration (stress risers) which also must be considered (this argument would favor the plaintiff in this case). I guess what Im alluding to is that for a 'fair' comparison or evaluation one should actually test to failure under exacting test condition a mirror image or exact duplicate of a keel section to be able to give 'objective' values or at least strength test small coupons of the actual material; otherwise, such a 'report' becomes simple 'lawyer talk' or 'cookbook' engineering analysis.
Im dumbfounded that ABS specs. (as listed in the report) for FRG sections only recommends a FS of 2. .... in normal design (metals, etc.) any dynamic structure with 'critical stress' targets would at least have a FS of 3 (or more) for 'severe' dynamic service, especially in a extremely complex area such as a keel/wing root. FS=2 typically equates to an 'inshore' (less than 'coastal') boat - IMHO. My prior engineering design was in mega load high mobile steel: cranes, high 'lethal' mobile, stuff ... so I dont know if that would translate to such composite structure using such variable material as FRG. Doesnt look good for the builder - IMHO.
.... and besides a Cantilever Keel root is a MORE serious/complex design than that carrying a simple factor of safety of 2.0 (ABS std.) for simple 'shear' concerns, usually requiring a FS of 4 for 'offshore' boats if you carefully analyse and 'backcalculate' those designers who are reknown for 'bluewater' designs ... and that is notwithstanding fatigue concerns. Without access to the calculations of the finite element analysis done by the 'plaintiffs' engineer and without actual measurement (testing of actual 'coupon') of the material for 'actual' strength values, etc., it appears that the 'engineering report' is tailored to be 'typical lawyer speak' ... always doubtful in good objective forensic analysis/reports. Without such reporting of 'actual' (tested) values any court argument can easily be challenged and discounted. Fiberglass is not a material with consistent strength values as there are many factors such as glass to resin ratios, consistency of layup/voids, areas of untowards stress concentration (stress risers) which also must be considered (this argument would favor the plaintiff in this case). I guess what Im alluding to is that for a 'fair' comparison or evaluation one should actually test to failure under exacting test condition a mirror image or exact duplicate of a keel section to be able to give 'objective' values or at least strength test small coupons of the actual material; otherwise, such a 'report' becomes simple 'lawyer talk' or 'cookbook' engineering analysis.
Im dumbfounded that ABS specs. (as listed in the report) for FRG sections only recommends a FS of 2. .... in normal design (metals, etc.) any dynamic structure with 'critical stress' targets would at least have a FS of 3 (or more) for 'severe' dynamic service, especially in a extremely complex area such as a keel/wing root. FS=2 typically equates to an 'inshore' (less than 'coastal') boat - IMHO. My prior engineering design was in mega load high mobile steel: cranes, high 'lethal' mobile, stuff ... so I dont know if that would translate to such composite structure using such variable material as FRG. Doesnt look good for the builder - IMHO.
Most boats have keel stubs and that is a different engineering situation due to the local stiffening created by the "hat" section. I wouldn't necessarily put bolts through the floors there either. I sure would put in a lot more glass and backing plates than the other keel failure we were recently discussing though.
I haven't paid much, or any, attention over the years to fin keel attachments on canoe body shape boats. However, if it is as common as you indicate to bolt small footprint, concentrated loads, to the hull panels with no tie in to the internal stiffening structure, then designers and builders are doing something that neither I nor any steel boat designer I've ever considered competent would dream of doing.
I won't argue with any of your points about A&M's management and operation or how the reports address them. As I said, I'm simply not interested at the moment in that aspect of the case.
We agree that the design was horrible. I think the design played a significant role in the degree of damage resulting from the grounding and the trecherous nature of the failure mode it set the boat up for later. For Richards not to have pointed this out makes me suspect his impartiality. I might feel differently if I traced all his calculations through but I would be shocked if I could justify that design or overlook its role in the subsequent failure.
As Rich points out, the ABS yacht rules are a pretty damn small fig leaf to hide behind when there are bodies in the water. Their steel boat rules are quite good and very conservative but something sounds very out of whack with their yacht rules. A factor of safety of 2.0 would only make sense to me if the loads were assumed to include those of grounding and hauling off.
Yep, the calculations and the pretty Finite Element Analysis (FEA) can give us a really great read on stresses but what it can't show is that some designs can't be correctly fabricated or worse, can't be inspected and verified after fabrication...or also very bad, can't be repaired after damage. That is the part that seems to be lacking in some of the analysis on CW.. As I said before in this thread, the design, as drawn, probably satisfies the applicable standards .. but the damaged section was compromised by either the initial fabrication or the repairs.. There are ways to figure that out, but I doubt that sophisticated destructive testing required on the actual sections will be done. I like Ross's analogy.. if ya hang a weight supporting shelf by the narrow end only, ya better know what you are doing with the anchors!
I still say it was a faulty repair and not a bad design. Either the bolts were loose enough to cause a wobble effect or the keel didn't bolt up smoothly against the fiberglass after the repaired grounding(S). A&M is just trying to save their ____.
I am not by any stretch an engineer. My knowledge has been gained by reading and experience. Many structrual problems can be anaylised by considering the levers involved. Every kid learning to pull nails knows that the fulcrum needs to be close to the nail for easy pulling. When I bought my house forty years ago the first owner had hung pictures on ten penny nails driven into the studs. A two foot crow bar would pull them but it crushed the gypsum board. To distribute the load from the crow bar I padded the wall with a piece of 3/4 inch plywood. When I look at a fin keel, I picture in my head a long lever trying to pull the bolts through the hull. If the attachment is inadequate the bolts will pull out. If the bolting surface is up to the job then it will try to take a chunk of the bottom of the boat with it. If the bottom of the boat is strong enough then the keel stays in place. It really is a matter of how far the fulcrum is placed from the bolt.
Of course the design was the major element in how the boat responded to the grounding. Everything from the shape of the canoe body, the built displacement, the fairing of the floors into the topsides and, well, everything were factors. It's a very complex problem made more so by the fact that the initial impact on the keel was off-center on the "T" bulb and there were multiple secondary shocks as they pulled her off. I really don't want to defend the design and I don't have any loyalty to Richards either. But, composite failures are complex and they can occur within the laminate. I don't think there is any way to design around that. If we sail in plastic boats we, as users, have an obligation to be aware of the danger of inter-laminate failures. Take home lesson: if you run aground hard enough so that your hull has a major deformation and so that you can see daylight between the hull and the keel, have an expert check the entire hull. Richards does point out that the design is for racing and does note that the result was a boat more susceptible to damage.
ISO has a section where you are supposed to work out the additional tensile strength of the forward bolts and the additional sectional modulus for the aft floors needed for grounding. Richards includes the worksheet in his evidence but just eyeballing it from the photos the aft floors don't look to be tall enough or to have sufficient flange area... I suspect that was a design or build error that resulted in excess damage when they grounded.
I understand that you're looking at lessons we can learn looking forward. And, I agree that the design played a big part in the grounding accident damage and the tragic result. However, I still feel that Richards is right to say that absent the grounding and negligent inspection and repair work there is no reason to believe the keel would have failed when it did. I don't feel all that sad that the lawsuit will probably break the builder, but, having skimmed the reports, I think the real lesson is that FRP failures are complex and when there is reason to believe the laminate might be compromised it is critical to life and safety that they be carefully inspected by an expert. IMO, it is T A&M's failure to do that which lead directly to a the death. Their failure to do the right thing at the time and their attempt to cover it up are far more serious crimes, IMO, than Cape Fear's.
Looking forward, it seems to me that as racers demand higher aspect keels that serious thought should be given to attaching them in ways that can absorb more energy before the hull fails and perhaps you're right that more though should be given to making failures evident. Cars have crumple zones and alignment jigs. Perhaps we can borrow some of those ideas...
--Tom.
It appears to me that a fin keel is a wide beam fixed at one end and free to flex, with a concentrated load at the end. From there it looks to be a a simple strength of materials problem with a chance of impact loading. Is this an over simplification?
It strikes me that the simple fact is that any boat with this sort of long, narrow keel has less margin of error since there is so much less surface area the point of attachment to the hull. Sure, plenty of boats with this design never lose their keel, but still it seems to me that the design is inherently weaker. A grounding and/or faulty repair job, or a certain type of (weaker) layup that would not cause another design of keel to detach is more liable to with this design.
- Status