Lifting-eyes, known as pad eyes, were welded into the top corners of a module for lifting the module on to an offshore platform. Each pad eye consisted basically of a 50mm thick vertical plate measuring approximately 2m x 1m which contained a hole reinforced by rings welded to each side of the plate. This plate was welded to a 35mm thick base plate 0.5m wide and also to horizontal and vertical stiffeners.
Four pad eyes were inserted into the top corners of the module and welded to the main girders. Welding of the pad eyes, during sub-assembly and on the module, was carried out with basic electrodes baked at 400°C and stored at150°C and the joint regions were preheated to 100°C in accordance with the recommendations of BS 5235.
The steel was to DIN St-52-3M (analysis 0.18%C, 1.49%Mn, 0.43%Si, 0.017%S, 0.027%P). The steel had been aluminium treated during production. Magnetic particle inspection and ultrasonic examination of the pad eyes before they were inserted in the module showed that the welds met the requirements of ASME VIII and the pad eyes were then stress-relieved in the furnace.
Final inspection was carried out after all welding on the module had been completed and some toe cracks were found in the cruciform joints in two of the pad eyes. In addition to this there were indications of lamellar tearing in the base plates of each pad eye under the full penetration T butt welds.
Fortunately the toe cracks were shallow and could be ground out and repaired even though the access for the welder was somewhat restricted. However the only access for the repair of the lamellar tears was from the inside of the module. The base plate in each case had to be air carbon arc gouged and ground in the overhead position to reveal the defective areas which were then checked by dye penetrant examination. Welding repairs were then carried out with the preheat of the joint regions increased to 150°C to compensate for the lower energy input when welding in the overhead position.
Inspection of the repairs was carried out four days after completion of welding and the success of the operation was largely due to the skill of the welders working in the overhead position with restricted access and preheat.
Causes of cracking
Because of the restricted access as well as production deadlines, it was not possible to carry out metallurgical examination to confirm causes of cracking. The toe cracks reported were most likely to have been caused by hydrogen and may have been due to some relaxation of the welding procedure in respect of preheat, energy input or drying of electrodes. It seems likely that for the delay that occurred, in fact seven days, before HAZ cracking occurred some additional stress would have to have been applied to the welded joints.
Pad eyes of similar design have been fabricated previously from BS 4360G 50D steel without any indications of lamellar tearing. The St-52-3M steel used in the present case was known to have been aluminium treated, in which case the principal non-metallic inclusions would be manganese sulphide with a smaller amount of alumina inclusions. The MnS inclusion content of a plate is directly related to the sulphide content and a knowledge of the sulphur level can provide an estimate of the short transverse (ST) ductility of the plate which is one of the factors that governs the risk of lamellar tearing. The other factor is the degree of restraint of the welded joint.
The ST ductility of the steel involved was estimated to be 5-15% in terms of reduction of area. Correlations between %STRA and the risk of lamellar tearing that have been built at TWI indicate that the St-52-3M steel would be likely to suffer from lamellar tearing in highly restrained joints. It is possible that additional stresses could have been applied to the pad eyes some time after welding because of the application of heat to other parts of the structure. This could have been caused by preheating or welding at positions remote from the pad eyes which could have set up reaction stresses which would have added to any residual welding stresses.
Stresses of this nature can cause lamellar tearing or they may cause the faces of very tight pre-existing flaws to open up so that they can be detected by ultrasonic examination. It is also known that such defects can be made more readily detectable by stress relief heat treatment, which can cause some separation of the crack faces. For this particular highly restrained item, it was recommended by TWI that the final NDT inspection should in future be carried out not less than one week after welding or other heating operations on the module have been completed.
To avoid any possibility of lamellar tearing, plates for pad eyes should be ordered with guaranteed adequate through-thickness ductility. In highly restrained fabrications, the recommendations given in BS 5135 to avoid hydrogen cracking should be implemented with rigorous control and supervision.