A tragedy took place along a jetty as a fishing vessel was unloading its catch while a heavy ocean surge pulsed into St Helena Bay on the West Coast of South Africa. Disaster ensued as one of the jetty bollards, to which the ship was tethered, broke suddenly and was launched over the deck where it struck the chief engineer. Some months later One Eighty was appointed to determine the root cause for the failure of the bollard.
The first point of call was to examine the physical evidence. The bollard structure had fractured through the base plate as seen in the photo below. A picture of the fracture surface captured on the day of the incident showed no signs of corrosion, even though the rest of the bollard’s surface was completely covered in rust. This suggested that the failure occurred catastrophically and was not due to slow crack growth over time (fatigue). Further analysis of the failure surface with a scanning electron microscope confirmed that the bollard broke through a fast-brittle mechanism i.e. catastrophically.
Compositional analysis established that the bollard was made from grey cast iron. The jetty on which it was stationed was built in the early 1960’s. Although grey cast iron is no longer typically used for bollard construction, this material selection is typical for the era in which the bollard was made. But how strong was the bollard?
Tensile tests showed that the ultimate tensile strength of the bollard material was around 125 MPa. The material also showed very little elongation before failure. These results were consistent with the evidence that the failure occurred in a fast-brittle manner and would be expected from grey cast iron. A technical drawing of the jetty was eventually uncovered which specified 5-ton bollards i.e. the bollard should not break when subject to loads less than 5 tons.
The question was now posed: was the bollard overloaded or was there a defect in its material which caused it to break within its specified 5-ton capacity?
To answer that question, One Eighty investigated a material property called fracture toughness (K1C units MPa√m ). This would indicate of what size defect would have to exist within the casting or on its surface for it to fracture at a stress considerably lower than its ultimate tensile strength. The results showed that a sharp-edged flaw of 4mm or greater would be needed. No evidence of a flaw of this size could be found on the fracture surface. This now proved that the bollard must have failed at a load of 5 tons or greater.
British standard 6349 for marine structures, which was only introduced some decades after the bollard in question was installed, provides a standard mooring arrangement for each vessel class. These arrangements are designed to allow ships to sway gently with the ocean while remaining safely tethered to the dock or jetty. The schematic below shows how the fishing vessel should have been secured, where each ‘line’ consists of a single span of rope.
The clear evidence however, was that on the day of the incident the vessel was not tied up according to the standard mooring arrangement described in BS 6349. Instead, to hold the ship still in the turbulent ocean, they had looped each line multiple times from the jetty bollard to the ship and back again before tying down tightly. The schematic below indicates in red the rope loops employed by the ship’s crew on the day of the incident.
The Finite Element Method was then used to produce two separate models; one of failed bollard with the prescribed mooring arrangement and the other with the actual mooring arrangement used on the day of the incident. The purpose was to compare, for each arrangement, the stresses applied to bollard. The results showed clearly that the arrangement used on the day would have produced loads greater than 5 tons while the prescribed arrangement would have resulted in loads well within the bollard’s limits. The images below show the model developed for the actual mooring arrangement.
In conclusion, the bollard is grey cast iron which is consistent with standard bollard material selection of its time. Even though the surface of the bollard appears rusted, no flaws exist in the casing which would reduce its mechanical properties from the 5-ton specification. On the day of the incident, the shipping vessel was not tied up according to the standard mooring arrangement specified in BS 6349, and as a result, the bollard was overloaded. This caused a catastrophic fast-brittle failure.