A study of a keel failure and the loss of the Cheeki Rafiki

Feature article written by Eur. Ing. Jeffrey N. Casciani-Wood, HonFIIMS

The recent loss in mid-Atlantic of the a.s.v. CHEEKI RAFIKI and the sad loss of four lives has brought very clearly to public and professional attention the problem of fin or pendulum keel detachment and consequent capsize of many apparently perfectly good boats. This is a very serious problem from a marine surveying point of view and the author has made a detailed study of a very similar loss that took place in February 2007 in the English Channel.

To put the matter into perspective it was first necessary to look into some published historic data. Because of their concern over the number of reported keel failures, the International Sailing Federation (ISAF) developed a Keel Structure Working Party to investigate and report on the problem. As a starter they produced a data base of known failures and investigated 72 of those that had been reported since 1984 and which involved no fewer than 24 deaths by drowning. The problem with the data, however, is that it records only the worst cases in which the keel had been pulled off or otherwise lost from the hull usually resulting in a capsize and an actual total loss. The marine surveyor investigating such a casualty will frequently find that it is very difficult to establish the prime cause of failure and, in particular, to differentiate between a loss caused by a design and/or manufacturing fault and a casualty resulting from human error such as a collision or a grounding. He must be particularly careful to avoid listening and paying credence to usually ill informed bar room or internet gossip.

It may be considered that the number of deaths when compared to the number of people sailing for pleasure is very small BUT IT IS STILL TWENTY FIVE DEATHS TOO MANY!

The Casualty
It must be stated at the outset that this study is not intended to point the finger of blame at any person or group of persons but only to enable marine surveyors to understand the possibilities and to increase his skill in, not necessarily forensic, examinations of small craft. The vessel whose loss is considered in this study was Bermudan sloop rigged, of 10.65 m LOA x 3.32 m breadth and 2.2 m mean draught. The depth of the hull was not given. The boat was of frp and pvc foam sandwich construction and the ballast keel was of the lever bulb type with the arm constructed of prefabricated steel. In late 2007 she had sailed from Plymouth towards Southampton following an out of season repair and maintenance.

In the early hours of the following morning the pendulum keel became detached and the boat capsized, causing the loss of life of one crew member. She was the first of ten boats in a class developed by a Dutch naval architect for use in Recreational Craft Directive (RCD) Category B waters and had, apparently, been designed using the American Bureau of Shipping (ABS) standards and had been built and marketed in Holland. Unknown to the designer, however, the builder had sub-contracted the construction of the hollow keel attachment and stem to another steel fabricator who had no marine experience. The fabricator had changed the design of the keel to make manufacture easier and to reduce costs but without adequately assessing the stresses to which the keel stem would be subjected to in service. Additionally, the owner had contracted a British yacht designer to optimise the yacht for IRM and IRC1 racing and that had involved adding a further 160 kg to the pendulum type keel torpedo ballast bulb. At the end of a successful racing season, the yacht had been delivered to Plymouth for repairs and maintenance. When she was taken out of the water, a considerable amount of detachment of the keel’s epoxy filler and loss of anti-fouling paint was found. It would appear that there had also been some evidence of the likelihood of fine cracking in the steel adjacent to the fillet weld but that had gone undetected before the boat was refloated for the voyage to Southampton.

Figure 1. The stress strain relationship
Figure 1. The stress strain relationship

The wreck was eventually salvaged and taken into Plymouth where investigation showed that the fabricated keel had failed just below the fillet weld connecting the fin to the taper box which was inserted into the hull. Detailed inspection of the wreck showed that defects were also found in the keel taper box welds and that two of the three keel bolts had also failed. Laboratory metallurgical analysis confirmed that the keel had suffered fatigue failure in the fillet weld area, which had been subjected to high bending stresses. The fabricated unit was unable to withstand the bending stresses developed in service and that had led to a condition of fatigue failure and resulting in the keel stem cracking just below the hull and consequent loss of the keel and capsize of the boat. Another yacht of the same type had also suffered a very similar fracture of its keel stem but, in that case the cracking had been noticed before the keel had completely failed. The existence of the second case provided confirmatory evidence about the structural inadequacy of the keel design and manufacture.

Independent analysis of the original design calculations confirmed that the keel had not achieved the required Safety Factor. Further analysis of the keel design, as built, showed that that also failed to achieve the required safety factor and by an even larger margin and that the subsequent addition of the extra bulb weight had exacerbated the situation.

Some Essential Definitions
The tensile strength of a material is the maximum amount of tensile stress to which it can be subjected before failure. The definition of failure can vary according to material type and design methodology. The idea of tensile strength is an important concept in engineering, especially in the fields of materials science and mechanical and structural engineering.

The marine surveyor should be aware of and understand the three typical definitions of tensile strength. The various definitions are shown in Figure 1 which shows a typical stress-strain graph for low-carbon or mild steel.

1. The yield strength is the stress that a material can withstand without permanent deformation.
It occurs at the yield point which, unfortunately, is not a sharply defined point for mild steel but is usually accepted to be the stress which will cause a permanent deformation of 0.2% of the material’s original dimensions.

2. The ultimate strength is the maximum stress a material can withstand.

3. The breaking strength is the coordinate on the stress-strain curve of the point of fracture.

Figure 2. Head box and stem as originally designed
Figure 2. Head box and stem as originally designed

The Keel
The original keel design had the keel stem encastré with the head box which, in turn, was bolted to a cruciform matrix inside the hull. The modification, made by the fabricator, attached the stem to the head box by means of fillet welds. The construction as originally designed is shown in Figure 2.

The maximum bending moment and shear force would be applied at point A in the sketch and would have been increased in direct proportion to the addition of the extra ballast weight in the torpedo bulb.

The Lessons
What lessons are there in this particular disaster for small craft marine surveyors?

It is frequently assumed that, when a keel is lost, the casualty can be attributed to the failure of the keel securing bolts. That is not necessarily so and, while it does happen, in truth, there are many other ways in which such a disaster can happen. Probably the most common cause of failure happens when the boat is driven hard aground through bad navigation or stress of weather. Such a casualty often results in her starting to break up in the inevitable pounding or where attempts to tow her off have resulted in extensive bottom damage.

It is essential that the marine surveyor adopts good practice when examining an frp pendulum keel boat when she is in a travel hoist or, better, in a steel cradle, and for to take hold of the bottom of the pendulum and try to move it from side to side looking for any traces of movement between the keel and the hull. If there is any such movement then a much more detailed investigation is essential. It is recommended that the marine surveyor when examining pendulum keels to use a dye penetrant all round the stem at the point where it emerges from the hull paying particular attention to any welds in that area. If any damage is found it would be considered good practice to ask for one or more of the keel bolts to be withdrawn for close up examination. A close up visual examination of both sides of the stem or traditional fin keel should be carried out looking for any staining, rust or signs of a gap between the top of the keel and the underside of the hull.

With traditional fin keels fitted with external ballast, a stain at the top of the ballast may be nothing more than a cosmetic defect but if the stains are in way of the keel bolts or their surrounding areas then the marine surveyor should expect that a serious problem is developing. The discolouration of a cast iron keel is probably nothing more than a cosmetic defect but rust on a steel fin keel may mean a serious corrosion problem or even the clear sign of structural failure. Semicircular gelcoat cracks in way of the ends of the keel or stem are clear indicators that the vessel has suffered a grounding. If that is so they will probably be tensile cracks at the leading edge of the keel and compression cracks at the after end. There will also probably be similar cracks around the heel of the rudder horn if the vessel is so fitted.

The marine surveyor instructed to apply his skill to a yacht whether it be for pre-purchase or insurance purposes should, of course, regardless of the boat’s material of construction always lift the cabin sole boards and examine the bilge space underneath. Any signs of water should be investigated. He should have the space dried out and, if the water returns while the vessel is afloat, carry out a more detailed investigation to find the source of ingress. Any dust, particularly white powdery dust on an frp boat should be carefully investigated as it may well be the indicator of a serious structural problem. It is good practice in his survey reports to suggest that his Clients inspect the space under the sole boards of their boats on a regular basis.

The keel bolt nuts should be closely inspected and hammer tested. A rusty nut on its own probably means nothing but the rust should be cleaned off (Coca Cola makes a good rust cleaner), the nut lightly coated with Vaseline and the nut regularly inspected. Again any reoccurrence should be closely investigated. Similarly any discrepancy in the appearance of the nuts should be investigated. A rusty washer probably means that the bolt is leaking or that another serious issue is arising. It should be thoroughly investigated. Any cracks in the hull or liner radiating outward from a keel bolt are serious and must be fully investigated. Loose keel bolts are an obvious defect and, if they can be tightened, it probably means that the ballast keel has started to part company from the hull possibly due to an undeclared grounding.

It is common practice on boats of frp construction to fit an inner liner or matrix which is bonded at various places to the outer hull. In many areas that bonding is inaccessible but it is also often defective having disbonded because of vibration or other reasons. The marine surveyor should, wherever the bonding is accessible, tap test it with a light hammer. A change in the note of the hammer contact will show up such disbonding and, if such is found, he should investigate the reason why. It may be no more than an easily repaired local defect but it may also be a sign of an existing or developing serious defect. It is possible to investigate such defects ultrasonically or by shearography or thermal imaging but such methods are very expensive. Any fractures, transverse or longitudinal, in the structure of the matrix especially if there are fractures or breaks in any nearby joinery work or cracks in the gel coat or surface paint indicate a serious problem and should be investigated thoroughly.

The final thing that the marine surveyor should look for is any signs of repair work. If any are found then the repair including its history should be investigated very carefully indeed.

It is also imperative, should the vessel be involved in grounding, that, as soon as possible, she be taken from the water and a detailed inspection of the bottom of the hull and especially the keel, stern gear and rudder be carried out by a recognised competent person. In such circumstances an inspection of the engine mounts is also advisable.

Needless to say all defects noted should be recorded on dated photographs.

While there are no legal obligations for pleasure craft owners to record inspections of their boats, the author is of the considered opinion that the marine surveyor should recommend to the Owner for their own benefit that they keep such a log of dates and details of all inspections ashore or afloat of the hull, rig and machinery warrants a specially dedicated logbook including copies of all Reports, repair specifications and invoices.

An important lesson for all marine surveyors to learn is that he must understand the characteristic of the materials from which the vessel is built and the conditions in which she will be operating and that not only at the time of the casualty. This necessarily requires that the marine surveyor MUST adopt and carry out a Continuous Professional Development programme.

One final point deals with the question of what should the marine surveyor do if he finds and reports a serious defect but the owner decides not to do anything about it. If the defects, in his opinion, seriously affect the boat’s seaworthiness he should report that fact to the boatyard or to the harbourmaster or both as well as in writing to the owner. If the boat is subsequently lost and somebody dies as a result the owner could be facing a charge of manslaughter and if he has not warned him in writing the marine surveyor could well find himself similarly charged as an accessory.

So BEWARE.

Latest Tweets from the IIMS