M Pereira, J B Speck and G R Razmjoo
J R Still
Formerly of Amerada Hess now with Lochhead Still Associates Ltd
Presented at: 20th International Conference on Offshore Mechanics and Arctic Engineering OMAE 2001, Rio de Janiero, Brazil, 3-8 June 2001
Many of the structural problems that face the operators of Floating Production System Offloading (FPSOs) can be avoided during design and construction. A programme of mitigation was developed to limit the in-service damage in a FPSO hull. During the early stages of engineering, consideration was given to the selection of hull materials, matching weld consumables and welding process. Fracture toughness and crack arrest tests were undertaken to characterise the fracture resistance of the hull welds. The data were used in a fracture mechanics-based investigation of fatigue crack growth and fracture susceptibility. A programme was also developed for the purpose of investigating the optimal method of improving the fatigue life of the hull weld details. The data generated during the investigation were used to develop a fatigue and fracture control programme for the FPSO hull.
Construction of the hull structure was in accordance with class rules. However, class non-destructive testing (NDT) specifies inspection of a limited number of fillet and butt welds within the hull structure. In order to ensure that the quality of welds not requiring class inspection was satisfactory, an inspection programme was established by the operator where additional NDT and visual inspection were applied during construction and erection of the hull structure. It is essential that the inspection programme is not diluted due to the construction schedule constraints which would limit the detection of weld irregularities and potential critical defects.
The offshore industry has been constantly looking for more economical solutions that would allow the exploration of marginal fields while cutting the time of production of first oil, optimise installation costs and maintenance of offshore production fields. The use of FPSOs in various offshore regions throughout the world have been steadily increasing and likewise the requirements for longer periods of service.
The requirement to use a FPSO without dry-docking for longer periods than those stipulated by the classification societies for the shipbuilding industry imposes additional constraints on the quality expected for the hull structure and in the in-service maintenance programme. It also requires that the materials used in the hull structure have particular fracture toughness and fatigue endurance when compared with merchant oil tankers. Additionally, FPSOs are stationed on location for long periods, which in some instances can involve 15 to 20 years depending on the life of the field. Consequently, an FPSO hull structure experiences continuous exposure to the elements and is subjected to loading conditions more severe than those experienced by conventional oil tankers.
FPSOs are offshore structures with the hull shape of a cargo oil tanker and experience severe loading conditions due to waves and loading/unloading at sea compared to semi-submersible structures. Although FPSOs are designed to weathervane into the direction of the prevailing wind and tides, they have to withstand severe weather conditions that an oil tanker, in many cases, would avoid by changing direction. The use of FPSO vessels in deep water is subjected to additional loads in the moon pool area of the hull structure due to the mooring system and flexible risers. This area of the hull structure requires additional reinforcement to withstand the force of the 100 year wave. Conversely, due to the more severe wave spectrum, fatigue cracks may develop in the hull more rapidly than in other similar structures, hence mitigation measures have to be taken to avoid, or detect early, problem areas where more frequent in-service inspection and repair need to be undertaken.
Fatigue and fracture control programmes (FFCP) based on fitness for purpose methods (or engineering critical analysis - ECA) have been applied in several industry sectors and they are well known in the oil and gas industry. ECA has been extensively used as an engineering analysis methodology to assess the significance of defects in several engineering systems, e.g., offshore jacket structures, pipelines, pressure vessels, storage tanks, etc. Although the methodology is in constant development and has produced several well known standards, e.g., BS7910  in the United Kingdom, API579  in the USA, WES2805  in Japan, it has not been extensively applied in the shipbuilding industry and in particular to FPSOs.
The reliability of the FFCP is dependent on the quality of the data for fracture and fatigue assessment of the considered component and on the quality of the stress data available for the analysis. The former would require accurate data on the fracture toughness of the material used in the hull, which implies that fracture toughness testing, e.g. crack tip opening displacement (CTOD), be performed on the selected hull material. It also would require a reduction of the uncertainties associated with S-N data that are in current use for the design of typical FPSO details, as pointed out by Lotsberg  . The latter would require an accurate estimation of the loading conditions of the FPSO, which would be representative of the operating and environmental conditions of the system during its designed life. Therefore, a more accurate estimation of the applied loads in critical structural components, both under static and dynamic loading, would have to be undertaken. For this purpose, the use of sophisticated engineering analysis tools to estimate the various loading conditions on the critical components, accounting for the statistical nature of these loads (i.e. short term and long term wave loading conditions), would be appropriate  . Because FPSOs are stationed, the wave spectrum for design is predominantly that of the region in which it is going to operate, and the wave heading expected during operation. The vessel speed in waves could be neglected while the number of loading conditions and wave heading could be simplified to the more critical loads applied in the structure (e.g. predominant wave headings, wave height, etc.). This apparently simplifies the number of loading cases that would be necessary to predict the load spectrum on the critical components, although it is still very complex in nature.
Nevertheless, FFCP can be applied in the earlier stages of construction of purpose built FPSOs, or new build tanker intercepted to become a FPSO, in order to mitigate the risk of catastrophic failure due to fatigue cracking in the hull and to ensure that a more reliable in-service inspection programme (ISIP) is applied to critical components during the design life of the system. The use of a FFCP for an existing tanker to be converted to a FPSO can be established provided sufficient data can be retrieved from the vessel construction records or additional testing is conducted where data are lacking. It is also important that additional inspection on critical components of the structure is undertaken during the modification stage. In order to apply such methodology a robust testing and analytical programme have to be undertaken which would help in the selection of the appropriate materials, welding process, fatigue improvement techniques and inspection methods to be applied throughout the design life of the system.
To guarantee the hull integrity of a FPSO a fatigue and fracture control programme (FFCP) should be developed to mitigate the risk of brittle fracture in the hull components. During the construction of purpose-built FPSOs or new-build converted tanker the quality of the hull structure must be of a standard to ensure the vessel can remain on station during the life of the field with the minimum of repairs to critical areas of the hull structure. Examples of the hull shape of a purpose built FPSO and an intercept FPSO, illustrated in Figure 1, demonstrate the unique hull shape of the respective vessels.
Fig. 1. Purpose-built FPSO and intercept tanker modified as an FPSO
Because of the high integrity required of FPSO hulls the inspection of fatigue-critical welded joints has to be made at a level above that normally required by classification societies. Experience has shown  that several weld defects or flaws, that are tolerable for traditional shipbuilding construction, are less tolerable for FPSOs, as listed in Table 1. Weld misalignment, root opening, undercuts, buried linear defects, weld distortions are amongst the weld defects that are less tolerable in FPSOs when compared to oil tankers. These planar defects will impose additional stress concentration which will consequently decrease the fatigue life of the component, therefore affecting the ISIP and the reliability of the assessment carried out by the FFCP if they are not spotted during the construction or modification stage. Additionally, these flaws will have to be repaired in-service which in some instances may not be practical.
Table 1 - Weld defects that are less tolerable in FPSO structures.
|Hull block welds
Buried linear defects
|Long. Stiffener joints
|Fillet weld at bulkhead stringers or cargo tank brackets
|Process facilities support steel work
|Longitudinal stiffener and bulkhead intersections
To ensure optimum hull quality, additional inspection of hull critical areas using conventional techniques, e.g., radiography, ultrasonic inspection, etc., is mandatory. The information obtained from the inspection will be used by the ISIP, which will rely on the FFCP during the operation of the system. Therefore, more accurate data will contribute to more reliable assessments during the expected operation interval.
Hull materials and weld metals
Owing to environmental conditions FPSO hull structures also require a higher fracture resistance than tankers, i.e. fracture initiation and crack arrest toughness. The former will imply that higher steel grades would be more appropriate due to their superior fracture toughness properties. The latter would require not only better crack arrest toughness of the steel used but also the use of appropriate welding methods that would produce higher crack arrest toughness on the weld metal and heat affected zone (HAZ). Also, the consumables to be used in the various joints throughout the hull construction need to be selected to ensure that the properties of the weldments match the hull materials.
Fatigue and fracture control programme
Based on fracture mechanics theory a fatigue and fracture control programme (FFCP) would take into account the fatigue crack growth of a certain flaw whilst determining the tolerable flaw sizes for specific critical components of the hull. The aim of a FFCP is, therefore, to carry out engineering critical analysis (ECA) of weldments and identify the most fatigue-critical details, which would require more frequent in-service inspection.
The ECA, which is based on methods of assessment of welded structures, such as the BS7910  , uses fracture mechanics fatigue assessment based on the integration of the relationship between crack growth rate (da/dN) and the stress intensity range ( ΔK), defined by the Paris law in Eq 1:
da/dN = A( ΔK) m
For the estimation of the remaining life of the component, a realistic initial flaw size would have to be estimated, e.g., tolerable undercut dimension, and the Paris law constants (A and m) applied in accordance with the environment surrounding the component (e.g. seawater, air, etc.). The load spectrum to be applied also needs to be defined in order to calculate the increment of crack growth per stress range block of n i cycles. For this purpose, the results of a short term and long term prediction of stresses applied on the hull, derived from ship motion 3D finite element analysis, would be required. Also, definition of the stress ranges that the component is subjected to during service such as nominal stresses and hot spot stresses. To mitigate the effect of fatigue loading several stress range blocks for different loading conditions expected during service can be applied during the analysis. In the absence of detailed stress ranges the most critical load cases could be used as reference, which would give a conservative estimation of the criticality of the component. The relative criticality of the component could then be estimated as the ratio of the residual to designed life.
It is also anticipated that different geometries and weldments will produce different ranges of tolerable flaws. Therefore, the ECA will help in identifying these regions, hence mapping the critical components with their specific requirements and limitations. The results of this analysis will be used as input data for the FFCP and the ISIP, which will monitor the system throughout its designed life.
Another output of the FFCP, if carried out during the design stage, would be the identification of areas where design changes need to be made to guarantee the expected performance of the structure during the design life (e.g., bracket, longitudinal and transverse girder joints, etc.). This early stage investigation would allow designers to avoid critical geometric profiles that could potentially increase the fatigue crack growth rates of a flaw in the component due to stress or strain concentrations, therefore optimising the maintenance schedule of the structure.
Recent investigation undertaken by TWI of an intercept tanker modified to be a FPSO 
has demonstrated the advantages of an extensive fatigue and fracture control programme specifically developed to support the in-service inspection programme for continuous operation on the central North Sea.
Originally destined for service as a 105,000 DWT Aframax tanker, the ship construction was intercepted and design modified for conversion to a FPSO. Because of specific and more stringent service requirements of the FPSO hull, changes were made to both materials and welding consumables for the outer hull to a grade that would provide improved fracture resistance from crack-like fabrication defects and subsequent fatigue cracks.
Although vessel hulls are constructed in accordance with class rules, it was agreed with the shipyard to allow critical and fatigue sensitive areas to be inspected, beyond class requirements, by the site inspection team consisting of the engineering contractor and operator personnel. This additional inspection, in conjunction with the improved hull toughness levels, would ensure that any future FFCP applied to the ISIP could be implemented with confidence.
Selection of steel grades and welding consumables to meet design requirements.
Selection of steel grades for vessel hulls is dictated by the classification society's rules, although service and environmental conditions will influence the selection. The original vessel design specified steel grades A and AH32 for the double hull. Since the vessel was being converted to a FPSO, different environmental and service conditions to be taken into account, the hull outer skin materials were changed to steel grades D and DH32. The higher grades of hull materials were selected for improved fracture resistance, i.e., fracture initiation toughness and crack arrest toughness. Table 2 shows a comparison of the required Charpy impact properties for the as designed and modified hull material. In addition to changing the material grades, the thickness of the outer hull plates was increased from 16.5 to 19.5 mm. All other plates remained as previously designed.
Table 2 - Required Charpy impact properties
| ||Test Temperature|
|* Not required
In order to obtain the desirable fracture toughness and crack arrest properties required for the hull joints in every construction stage, i.e., assembly, pre-erection and erection, a careful selection of the consumables to be used in the joints needs to take into account the required properties of the hull material. Also the quality of the weldments, e.g. fillet welds fit-up and weld profiles, etc., will require special attention to the specified quality required for an FPSO construction. For each construction stage, one or more welding process can be used which would require specific consumables, as shown in Table 3.
Table 3 - Example of the consumable grades used for fabricating grade D and DH32 steels.
|Welding Process||Filler Grade||AWS Grade||Welding Position||Yield N/mm 2||Charpy Impact|
|Temp °C||Joules (Average)|
|All Positions and Downhill
|All Positions and Downhill
|(*) GFW - Gravity Feed Welding
In order to generate data to be used in the FFCP an extensive experimental programme to investigate the fracture toughness and fatigue life of specific joints in various hull locations, see Fig.2
, was undertaken. Additionally Pellini drop-weight tests 
were undertaken to provide a measure of the crack arrest capacity of the weldments.
Fig. 2. Hull cross section with investigated areas indicated as A to E
The experimental programme generated fracture toughness data via both Charpy impact tests and fracture toughness tests (CTOD) for the parent metal, weld metal and HAZ. Charpy impact results, as shown in Table 4, were undertaken according to the classification society's specification and were found to satisfy the minimum requirements. Fracture toughness tests (CTOD) were undertaken according to BS7448 Parts 1&2  . The range of fracture toughness test results (CTOD) for a number of hull test panels is shown in Fig.3.
Table 4 - Summary of Charpy impact toughness tests results for 25mm hull plates at -40°C.
|Weld region||Test panel||Material||CVN, J at|
|CVN, J at|
|ABS Gr. DH32/BS 7191 355 EMZ + E7016
ABS Gr. DH32/API 2H 50T + E7016
|ABS Gr. DH32
BS 7191 355 EMZ
API 2H 50T
|M03 side (ABS Gr. DH32)
M04 side (BS 7191 355 EMZ)
M03 side (ABS Gr. DH32)
M05 side (API 2H 50T)
|(1) Average of three results
(2) Estimated energy at -20°C, based on energy-temperature correlation of 1.5J per °C
Fig. 3. Summary of results of CTOD tests performed at -10°C, for three weldment regions investigated. The horizontal bar indicates the second lowest CTOD test results for a set of six tests, i.e. δ mat a)
The scatter of results within the non-homogeneous HAZs and weld metals is immediately apparent in Fig.3 (compared to the relative lack of scatter found in the homogeneous parent material). The δ mat values for all weld regions are above 0.10 mm. The highest δ mat values were measured in the parent materials (i.e. δ mat > 0.50 mm) and both parent materials exhibited fully plastic behaviour. The weld metals have the lowest measured δ mat and exhibited a combination of fully plastic behaviour and brittle crack extension.
The fracture events of the weld HAZs were fully plastic, except for a few results from the HAZ regions of panels A2 Flux Cored Arc Welding (FCAW) and A5 Electro Gas Welding (EGW). The δ mat results of each weldment region also tend towards the bottom of their respective ranges, except for the HAZ δ mat of panels A2 (FCAW) and A5 (EGW).
Crack arrest capability
Pellini drop-weight tests 
to determine the nil-ductility transition temperature (NDTT) were undertaken on all sample panels in the various regions of the weldments, e.g. parent metal, weld metal and HAZ. The HAZ was found to have the highest NDTTs (i.e. the worst crack arrest capabilities) of the three weldment regions, ranging from -20°C to +15°C. The weld metal in all test panels was found to have the lowest NDTT of the three weldment regions, ranging from -55°C to -90°C.
In the specific case of the EGW welds tested, if a crack were to initiate from a pre-existing welding flaw and subsequently propagate exclusively in the HAZ regions of the hull butt welds, under the pessimistic circumstances of yield strength operating stress levels (due to welding residual stresses, or accidental impact loading) and at the minimum anticipated service temperature of 0°C, it is unlikely that the crack will arrest in the HAZ regions of hull butt welds. This assessment is, however, believed to be conservative because the beneficial effects of load shedding (i.e. reductions in both operating and residual stress levels as the crack propagates in the structure) have not been considered. It is likely that load shedding would occur in the hull in the vicinity of the crack tip on crack propagation, thereby reducing the crack driving force and assisting crack arrest. Nevertheless, it should be noted that an exhaustive evaluation of the hull in terms of the crack arresting ability of the hull weldments would require consideration of quantifying load shedding and consideration of the ductile tearing stability of long (several metres) arrested cracks within the structure.
Engineering critical assessment
In order to demonstrate that welding flaws remaining in the hull of the FPSO after pre-service inspection would not lead to total failure due to the fatigue loading during the vessel service life, an ECA was undertaken to estimate the fatigue crack growth in selected components. The ECA was used to demonstrate that fatigue crack growth would not lead to total failure, i.e. local fracture or leakage would happen instead during the service life of the hull.
The ECA was undertaken according to BS7910  using as input data the results of the experimental programme and stress ranges spectrum associated with the specific components that were provided by the FPSO structural assessment performed by the vessel designer. In accordance with BS7910, an assumed applied stress ratio of R>0.5 was used, i.e. the applied or effective mean stress was ignored and the stresses for the stress range spectrum of the hull were assumed to be fully damaging, regardless of the level of the applied stress ratio.
Other assumptions such as initial flaw size, Paris law constants, stress concentration factors due to axial misalignment of abutting plates, etc. were made using available data and by calibration methods as in BS7608  .
Additionally, estimations of the maximum tolerable flaw size were made according to BS7910  and using the software CRACKWISE 3. The results of the analyses were plotted as a locus of maximum tolerable flaw sizes, as shown in Fig.4, for the FCAW and EGW welds. The results for the FCAW and EGW welds means that a surface-breaking flaw of dimensions a x 2c that falls above the respective failure locus would be unacceptable. Figure 5 shows the predicted crack depth plotted against the number of cycles during the fatigue crack growth. Cracks found during in-service inspection can be checked against the tolerable flaw size locus ( Fig.4) and, if found to be in the region of tolerable flaw sizes, additional estimation of the fatigue crack growth can determine the remaining time for total failure ( Fig.5). Therefore, the estimated remaining time to total failure will define the next inspection interval according to the operator ISIP.
Fig. 4. Idealised failure loci for tolerable surface-breaking flaws for specific EGW and FCAW weldments
Fig. 5. Predicted fatigue crack growth (crack depth) against number of cycles
Fatigue improvement techniques
In a complex welded structure such as an FPSO, fatigue life is governed by the performance of welded details with poor fatigue strength, e.g. brackets between longitudinal hull stiffeners and transverse webs (frames). The performance of these details can be explained in terms of the overall geometric stress concentration at the joint, surface profile or shape of the weld, the presence of microscopic crack-like discontinuities at the weld toes and the development of welding residual stresses, ordinarily of yield strength magnitude.
To deal with these stress enhancing factors, and to achieve increased fatigue life for some structural details on the FPSO, fatigue life improvement techniques have been investigated, i.e. weld toe grinding and weld toe needle peening. Fatigue testing were conducted in air at ambient temperature, under uni-axial amplitude loading, and at frequencies of 4 to 8 Hz. The specimens were loaded until complete separation was achieved. The load spectrum applied to the fatigue tests was a simplified wave action standard history (WASH-W) [Ref.10] , i.e. excluding loading frequency and calmer sea-states 1 to 3. In order to avoid buckling of the fatigue specimens large compressive loads were also discarded. To compensate for the latter, additional compressive loads were added to the spectrum at 1,000 cycle intervals. This modification is partly conservative since the repetition of compressive loads introduced in this way should exceed the occurrence of compressive overloads under service loading conditions.
The results of the fatigue testing are shown in Fig.6. The results show that there is no significant difference between the fatigue lives of the two techniques investigated. Therefore, the operator could choose the most suitable technique based on practical considerations (e.g., speed of application, cost, quality control, etc.)
Fig. 6. Fatigue improvement technique test results for as welded, ground and peened conditions of steel grades D and DH32 respectively
Class requirements, as applied to tankers, are superseded by the UK HSE's DCR  requirements for FPSO operated in the UK sector of the North Sea. An In-Service Inspection Programme (ISIP) was therefore established for the FPSO based on maintenance and inspection philosophy developed by the operator and contractor. The strategy adopted included the use of a FFCP. An in-service inspection strategy for the main hull plate butt welds was subsequently obtained from the FFCP, which identified the hull welds most at risk of failure by fatigue and fracture. Furthermore, on applying the results of the FFCP, critical fatigue crack dimensions could be established. However, the results of the FFCP and subsequently of the ECA can only be used with confidence if a competent inspection management system is in place during the vessel construction and a detailed ISIP is established onboard the FPSO.
The investigation undertaken on an intercept FPSO construction has applied fracture mechanics to develop a Fracture and Fatigue Control Programme that formed the basis for the In-Service Inspection Programme adopted by the operator. The results of the experimental programme and the structural analysis performed by the designer were used as input for the FFCP, which estimated maximum tolerable flaws sizes and the expected fatigue crack growth of selected critical components. An experimental investigation of fatigue life improving techniques also demonstrated that there was no particular evidence to support the selection of one or another of the two techniques investigated in terms of the expected improved life, therefore the operator could choose the most convenient technique in terms of quality control, speed of application and costs.
The work undertaken has demonstrated that a FFCP has many advantages on the selection of the appropriate material, welding process and filler metals for the construction of a FPSO hull, which could also be applied for oil tankers and any other sea-going structure. It has also been demonstrated that a FFCP can be used effectively from the design stage to construction of the vessel. The accuracy of the methodology developed depends on the quality of the data available for the analysis, therefore the data collected from the earlier stages of design, construction and in-service inspection will improve its effectiveness throughout the service life of the vessel. The full benefits of a combined ISIP and FFCP to mitigate the risk of catastrophic failure, improve the maintenance and repair of the hull structure, in order to increase the operation period without the need for docking, remains to be seen during the operation of the vessel.
- BS7910: 1999 - amendment 1 - Guide on methods for assessing the acceptability of flaws in metallic structures.
- API 579: Part 1 & 2 - Recommended Practice For Fitness-For-Service, issue 12, Sep. 1999.
- WES2805 - Methods of Assessment for Defects in Fusion Welded Joints with Respect to Brittle Fracture and Fatigue Crack Growth, JWES, 1995.
- Lotsberg I, Background and Status of the FPSO Fatigue Capacity JIP, in Proceedings of Offshore Technology Conference 2000.
- Technical Guide Regarding the Strength Evaluation of Hull Structures, Nippon Kaiji Kyokai (ClassNK), Dec 1999.
- Still J R and Speck J B, Hull Weld Quality Critical for Offshore Oil Production Vessels Welding Journal (AWS) August 2000, Vol79 pages 33 to 38.
- ASTM E208-95a: Standard test methods for conducting drop-weight test to determine nil-ductility transition temperature of ferritic steels, ASMT, 1995.
- Literature Review of Fracture Properties of Grade A Ship Plate, Offshore Technology Report, Health and Safety Executive, OGT 95 489.
- BS7608:1993 - Fatigue design and assessment of steel structures.
- Schutz W and Pook LP, WASH (Wave action standard history), A standard stress-time history for offshore structures, Steel in Marine Structures, ED Noordhork C and de Back J, Elsevier Science Publishers,Amsterdam, 1987.
- SI 1996, No 913: A guide to the installation, and miscellaneous aspects of amendments to the Offshore Installations (Safety Case) Regulations 1992, Offshore Installations and Wells, Design andConstruction Regulations, 1996.