Fatigue of welded joints

Fatigue design

Residual stresses and their influence

Synchrotron and neutron strain scanning

Multi-parameter fracture mechanics characterisation of crack tip stress fields

Crack shielding - modelling, photoelastic determination

Fatigue of materials

Small fatigue cracks

Variable amplitude load interactions

Finite element analysis

Failure analysis and fractography

Design and rapid prototyping

Friction stir welding

The University of Plymouth is well equipped for research in Advanced Materials & Engineering.  In the 2008 Research Assessment Exercise 100% of its research was assessed as nationally recognised, while  80% of its research was internationally excellent or internationally recognised.   Facilities include:

Instron screw driven general purpose testing machines with capacities from 5 kN to 100 kN.

Instron servohydraulic testing machine with 100 kN capacity, 1500 degree C furnace, high temperature extensometer suitable for ceramic and hardmetal specimens. Computer controlled and capable of random load testing and fracture toughness testing.

Instron electromagnetic resonant frequency testing machine of 100 kN capacity housed in an acoustic chamber (see picture above).

Avery rotating bend/torsion fatigue testing machines and an Avery tensile fatigue testing machine, suitable for S-N testing.

Facilities for SEM and TEM with analytical and image analysis capability.

Confocal laser scanning microscope.

Reflection and transmission photoelastic stress analysis equipment.

Residual stress measurement via hole drilling techniques and synchrotron/neutron radiation.

Computer aided design and numerical modelling.

Metallography and microstructure suite.

Suite of 3D rapid prototyping machines and laser cutter.

None at present

Further information regarding publications and research can be found on the UoA 28 web pages.

Multi-parameter fracture mechanics modellimg of crack tip stress fields: 2008-2010.  PhD project, collaborative with Michigan State University, USA.. 

Investigator:  Yanwei Lu

Optimisation of shot peening for 12Cr steel steam turbine blades: 2007-2009.  PhD project supported by ESKOM, South Africa. 

Investigator:  Mark Newby

Characterisation of pre- and postweld heat treatment effects on the residual stress field of friction taper stud welds in creep resistant 10CrMo910 steel: 2008. Institut Laue-Langevin, Grenoble, Experiment 1-01-26, SALSA strain imaging beamline, allocation of 5 days of beam time. 

New materials technology and manufacturing processes in the redesign and production of high quality critical flow tube fittings: 2005-2006.  Knowledge Transfer Partnership with Parker Hannifin plc, DTI Grant KTP000976). 

Performance of galvanised high performance steels for the construction market: 2005-2006.  (Research contract from International Lead Zinc Research Organisation ILZRO, North Carolina, USA). 

Improving fatigue performance of higher strength steel welds: 2006.  Institut Laue-Langevin, Grenoble, Experiment 7-01-196, SALSA strain imaging beamline, allocation of 4 days of beam time.

Optimising residual stresses in friction taper stud welds in creep resistance steel 10CrMo910: 2007. Institut Laue-Langevin, Grenoble, Experiment 1-01-8, SALSA strain imaging beamline, allocation of 5 days of beam time. 

Synchrotron diffraction investigation of residual stresses in 12Cr steel used in stationary steam turbine blades and their modification by shot peening and fatigue cycling: 2007.  European Synchrotron Radiation Facility, Grenoble, Experiment MA-326 – Beamline ID 31, allocation of 12 by 8 hour synchrotron beam shifts.

FaME38 - Exploitation and Development of the Joint Support Facility for Materials Engineering at the ILL-ESRF as a UK Collaborative Research Group: 2005-2006.  Funded by the EPSRC under grant EP/C008847/1

This was led by Plymouth, and was a collaborative research group comprising a consortium of 7 other UK Universities (Manchester, Oxford, Loughborough, Imperial College, Sheffield Hallam, Sheffield and the Open University) in partnership with the ILL and ESRF at Grenoble.  Plymouth managed the FaME38 activity in Grenoble from July 2005 to February 2008.   The laboratory manager on-site was Dr Axel Steuwer.

This project met the folllowing objectives and was ranked "Tending to Internationally Leading" in the EPSRC IGR:

1. To develop and to exploit the unique potential of FaME 38 at the ILL-ESRF for engineering applications of neutron and synchrotron radiation.  These include in-situ studies involving dynamic and thermo-mechanical loading, fast strain scanning, neutron tomography, very high-resolution X-ray nanotomographic imaging, and small angle scattering.

2. To provide fast, precise, engineering-user-friendly measurement capability, coupled with high quality data processing and visualisation, and an 'intelligent' interface between beam-line scientists, materials scientists, and engineers to ensure optimised use of beam time so as to promote efficient engineering usage and to extend the boundaries of engineering science.

Areas of interest include deformation mechanisms and texture evolution, phase transformations and micromechanics, in-situ studies of welded and bonded joints, processing and joining high temperature materials for the aeroengine and nuclear industry, stress corrosion interactions and cracking, residual stress-crack interactions, and more reliable life prediction for joints and advanced components/materials.

3. To inform, and to embed the research potential of neutron and synchrotron facilities into the UK/European engineering research community.  FaME38 is well placed to liase between and to link UK materials science and engineering research with major European initiatives such as GENNESYS (nanoscience and nanotechnology), and to interact with other collaborative facilities at the ILL-ESRF, such as the Partnership for Structural Biology and Microfocus and Nanotechnology initiatives. The location of FaME38 at the ILL-ESRF site is highly advantageous for the UK as it will enable the UK engineering user community to benefit from interactions with the leading users from across Europe.

4. To establish linkages that will allow development and implementation of neutron and synchrotron X-ray techniques and procedures to agreed international standards to facilitate industrial engineering usage.

5. To facilitate the transition to full incorporation of FaME38 within the infrastructures of the ILL and ESRF so as to establish materials engineering as a routine activity at ILL-ESRF.
 

The aim of this experiment was to evaluate the effectiveness of SALSA in strain scanning of a fairly heavy, thick plate, multipass steel weld. It forms part of a programme of work aimed at evaluating the role of residual stresses in affecting the fatigue performance of welded higher strength steels as a function of weld process conditions and to use this to empower weld life prediction models. Weldable higher strength steels (tensile strength > 500 MPa) are specifically formulated for low cycle fatigue situations (lives typically less than 105 cycles) where significant engineering benefit derives from their higher strength. Thus, for example, at a life of 105 cycles, the allowable fatigue strength of such a steel may be 220 MPa for the as-welded condition, which is more than double the value at 107 cycles. The main interest revolves around the use of such steels in tubular structures (bridges, offshore platforms etc). Typical tube wall and hence specimen thickness ranges from 8 mm up to 100 mm. Certain joints have shown much higher fatigue performance in low cycle fatigue than other very similar ones in previous work and, whilst residual stresses were thought to dominate this behaviour, initial measurements made using hole drilling and surface X-ray were contradictory and inconclusive.

The present beam time proposal was formulated in the context of a PhD project that aims to examine the problem using the powerful strain scanning tool of neutron diffraction via the high intensity neutron source, the specimen handling capability and data processing software available for SALSA at the ILL. This instrument will greatly facilitate accurate, fast and precise acquisition of high quality information regarding the 2 and 3-dimensional residual stress field through the depth of thick Multipass welds in steel. In particular, the stresses at the critical depths between about 0.5 mm and 2.0 mm can be characterised in detail for the as-welded condition and the changes observed in peak position and magnitude with change in thickness and process parameters. This type of experiment has the potential to significantly improve understanding of the changes in residual stresses under fatigue cycling.

Influence of process conditions on dynamic performance and defects in friction stir welds:  2003-2006.  Funded by HEFCE QR and the NRF, South Africa.

Collaborative with Professor D Hattingh, Department of Mechanical Engineering, Nelson Mandela Metropolitan University, South Africa.

Investigator: Hannelie Lombard

Certain alloys are susceptible to a novel planar type of weld defect during friction stir welding.  These have been ascribed to the operation of a strain partitioning ductility drop induced during high rate loading.  This project makes used of a computer controlled and instrumented FSW machine which measures forces, torques and temperatures during welding.  These parameters can be fed back into the machine to control its speed and feed and, in principle, used to identify optimum weld process conditions in terms of dynamic performance.  Information on forces can be coupled with single-point machining models to extend model predictions to different alloys and weld conditions, and to determine optimum tool type from the polar 'force' footprint associated with welding.

Process and performance optimisation for friction stir welds using the force footprint and strain scanning: 2004.  Funded by the ESRF (Experiment ME 992: 27 x 8 hour shifts on ID31).

Collaborative with Professor DG Hattingh and Dr Axel Steuwer (FaME38 at ILL/ESRF).

This experiment had several linked aims, the first of which is to evaluate the residual stress state in friction stir welds in 5083 alloy as a function of varying process parameters, which is the purpose of the synchrotron beamtime application.  This information is essential to advance our ability in predicting optimised process conditions for FSW joints, through the other aims of linking the residual stress state with the characteristic 'force footprint' obtained during welding and of correlating these with process parameters and the dynamic fatigue performance of the welded joints.  The net outcome of this will be a significant improvement in understanding and modelling of the process-structure-performance relationship for industrially important FSW.

Residual stresses in shot peened alloys and their incorporation into advanced life prediction: 2003.  Funded by the ESRF, Grenoble (Experiment ME 748: 30 x 8 hour shifts on ID31).

Collaborative with Professor J R Yates, Department of Mechanical Engineering, University of Sheffield, and Professor A Navarro, University of Seville.

The aim of this project was to evaluate the residual stresses in shot peened components and, particularly, the changes that occur during fatigue cycling.  This realistic stress information will then be built into an advanced micromechanical crack growth model for fatigue life prediction, and also input into determination of optimum peening conditions for maximum fatigue life.

A prime example of the engineering importance of residual stresses is life prediction modelling for shot peened components and, particularly, the influence of fatigue cycling on residual stresses.  Shot peening induces beneficial compressive residual stresses into the surface at critical locations in highly stressed components, eg. airframes and aircraft engines.  It also changes the microstructure in the peened region due to inhomogeneous plastic deformation.  Controversy still exists in the shot peening and fatigue community as to whether the major benefits of shot peening should be ascribed to the compressive residual stresses or to microstructural changes which occur over the same region.

 This controversy adversely affects our ability to predict the fatigue life of peened components, and arises partly because of uncertainties in accurately measuring the magnitude of residual stresses through the depth of the shot peened layer, and partly from lack of knowledge of the effect of fatigue cycling on such stresses.

Recently, sophisticated micromechanical modelling of small fatigue cracks has been attempted, that includes consideration of residual stresses.  Realistic life prediction for shot peened components, must take account of all relevant factors, e.g. hardness, grain size, plastic deformation, small crack growth mechanics and residual stress magnitude.  The potential improvement offered by this technique could not be properly assessed against experimental data because of poor knowledge of the residual stress field.  The stresses must be measured over a depth of at least several millimetres below the surface, as this region has a major influence on fatigue life for small cracks.  This means that laboratory x-ray diffraction is not useful, while incremental hole drilling is unreliable, and semi-destructive, hence it will shift crack initiation sites to the hole under subsequent fatigue cycling.

The proposed work is thus timely, and in engineering terms, highly innovative in its linkage between advanced micromechanical modelling of crack growth for shot peened components, synchrotron measurement of residual stresses, and the effect of fatigue cycling.  It also has a very high level of technological importance.  The work aims at demonstrating the potential improvements to life prediction procedures for shot peened components that would follow from micromechanical life modelling, using reliable residual stress information.  This requires performing detailed residual stress measurement in the as-peened condition and after various numbers of applied fatigue cycles.  The data obtained from this work will provide fundamental understanding of the effect of fatigue cycling on shot peening stresses, and allow separation of the influences of plastic deformation/hardening and residual stresses.  This enhanced understanding of such influences as type of applied loading and work hardening characteristics of different alloys, and of optimisation of shot peening parameters, can be incorporated into a refined and improved version of the microstructural fracture mechanics model for life prediction of shot peened specimens.

Enhancing the fatigue performance of welds in higher strength steels: 2003-2005.  Supported by Corus Research & Development, Rotherham.

Investigator:  Sii-Pin Ting

The aim of this project is to evaluate the potential to consistently fabricate butt welded joints in higher strength steels, that possess high fatigue strength.  Weldable higher strength steels (tensile strength > 600 MPa) are specifically formulated for low cycle fatigue situations (lives typically less than 10⁵ cycles), where as seen in Figure 1, significant engineering benefit derives from their higher strength.   Thus, for example, at a life of 10⁵ cycles, the allowable fatigue strength is 220 MPa for the as-welded condition, which is more than double the value at 10⁷ cycles.  The main interest revolves around the use of such steels in tubular structures (bridges, offshore platforms etc). Figure 2 indicates how typical 4-point bend fatigue specimens with fillet welded attachments, can simulate the joint between chord and brace sections in such tubular structures.  Typical tube wall and hence specimen thicknesses range from 8 mm up to > 25mm.  Figure 1 also indicates, clearly, that certain joints have much higher fatigue performance in low cycle fatigue than other very similar ones (compare the open circle data with the dashed and solid lines for example).

Figure 1                                                       Figure 2

If the reasons for such variation could be understood, it would allow weld design of higher strength steel structures to become more efficient.  This would have a substantial beneficial economic impact on European steel manufacturing, and on structural design and fabrication.  It would also assist in more efficient use of structural steel, and hence improve utilisation of natural resources.  These potential benefits were recognised in previous research programmes to examine the effect of various weld parameters on fatigue performance of higher strength steels.  These research programmes produced specimens that occasionally displayed much enhanced life performance (life improvement by a factor of 2-4; see Figure 1), but did not accomplish the aim of identifying which weld or metallurgical factors were influential.  Equally, they were unable to clarify the role of residual stresses on the phenomenon, using laboratory x-ray data.

Direct analysis of the stresses around a crack under the influence of a plastic enclave around crack tip and wake: 2001.  Funded by the EPSRC (Reference 37131) under the Direct Access SRS beamtime policy.  10 days were allocated on station 16.3 of the Synchrotron Radiation Source at the Daresbury Laboratory.

Collaborative with Professor PJ Withers, Manchester Materials Science Centre, University of Manchester/UMIST and Professor EA Patterson, Department of Mechanical Engineering, University of Sheffield.

Output from the work in grant GR/L42391 indicates that a fundamental re-assessment is required of the interpretation of plasticity-induced closure, and of potential problems in assessing crack closure via models which incorporate crack wake contact forces (as is routinely done, for example, in the aerospace industry). The results shed, apparently, important new light on the use of compliance to characterise closure, and demonstrated that fracture surface contact loads did not correlate with effective stress intensity factor in polycarbonate CT specimens. Attention is being given to a direct transfer of the principles and techniques developed in the work to metallic specimens, which are the cases of prime industrial and engineering importance. Preliminary work on aluminium specimens has indicated that similar photoelastic fields are obtained in reflection from birefringent coatings, and that the crack tip stress fields in the coatings are, in principle, analysable. When interpreting results of such a potentially influential nature, however, it is very desirable to cross-correlate the data with that obtained from a different technique.

Such cross-correlation is within the capabilities of synchrotron diffraction and a feasibility experiment was carried out by Professor Withers. Results from this suggest that for thin specimens (2 mm thick), as have been used in the work under grant GR/L42391, the use of synchrotron radiation gives results for aluminium specimens, which are directly comparable with the transmission photoelastic data obtained from polycarbonate specimens, and which mesh well with reflection photoelastic work on coated aluminium specimens.

The feasibility experiment has indicated that useful residual strain (stress) line scan data can be obtained using synchrotron radiation from relevant aluminium specimens. However, in a steeply varying stress field around a crack tip, single line scans provide little insight and can give a misleading impression of the overall stress field, being highly sensitive to the exact location. The present experiment aims to refine the use of line scan techniques, and to extend the technique to area maps. Such full field stress data is provided by the transmission photoelastic techniques which have been developed for polycarbonate specimens, but cannot be obtained in transmission for aluminium except by synchrotron radiation.

Residual stress field under fatigue loading in MIG welded marine grade aluminium alloy: 2001/2002.  Funded by the ESRF, Grenoble (Experiment ME 282: 24 x 8 hour shifts on BM16).

Collaborative with Professor PJ Webster, Centre for Materials Research, Division of Civil & Environmental Engineering, University of Salford.

Fatigue and corrosion properties are the life-limiting factors in aluminium alloys used for marine applications. Current codes used for life prediction have to be highly conservative, as the combined effect of residual stress field, local stress concentrations and applied fatigue load on performance of the parent plate is largely unknown. This means that current generation aluminium structures and vehicles are not as light and efficient of resource use as is desired. Whilst new alloys are being developed with improved corrosion resistance and higher parent plate fatigue strength, these improvements cannot, as yet, be utilised in structures with lower design margins against failure.

One of the most significant factors which has hindered greater structural use of lightweight aluminium alloys is uncertainty regarding the fatigue performance of welds. Current fabrication practice uses metal-inert gas (MIG) fusion welding processes. MIG welds in typical marine grades of aluminium, such as the 5000 series of alloys, both anneal the plate in the weld region and introduce high residual stress fields. The required tensile properties in these alloys are obtained through strain hardening, so there is no practicable way of restoring parent plate properties to the weld region, although the effect can be reduced by lower heat input. The residual stresses in fusion welding are typically of yield strength magnitude, and they tend to change the mean stress in applied fatigue cycles and therefore to reduce the allowable stress amplitude for a given cyclic life. Full information on residual stresses is currently lacking, although their importance to fatigue and fracture performance is widely recognised [1-3].

To enable the advantages of the new alloys to be incorporated in more efficient design, better physical understanding of the residual stress field around welds, and its interaction with applied loads (stress relaxation), is required. This would allow a re-assessment of existing conservative fatigue design codes and advance the design envelope for lightweight structures. To date, very limited information on the 3-dimensional nature of the residual stress field is available, and even less is known about the real interaction between residual and applied stress fields.

MIG welding is the current industry standard process and is likely to remain so for the foreseeable future, although newer techniques, such as friction stir welding are expected to make inroads into modern fabrication practice. This experiment is designed to:

a)     Obtain detailed residual stress information for plates of MIG welded 5383-H321 marine grade alloy which are large enough to be representative of real structural plates.

b)     Examine the re-distribution of stresses which takes place when fatigue cycles are applied. Of particular interest is the effect of underload and overload cycles, and of the re-distribution achieved as a function of number of applied load cycles.

c)    Use information on effective fatigue cycle at the MIG welds to assess the extent that this can account for slope changes in S-N curves.

Residual stress field in friction stir welded marine grade aluminium alloy: 2001.  Funded by ESRF, Grenoble (Experiment ME197: 24 x 8 hour shifts on BM16).

Collaborative with Professor PJ Webster, Centre for Materials Research, Division of Civil & Environmental Engineering, University of Salford.

This work investigated the 2-dimensional stress field associated with single pass and double pass friction stir welded specimens of 5383-H321 aluminium alloy, as a function of applied fatigue cycles.   Full through-thickness maps of the residual stresses were acquired in both the longitudinal and transverse directions to the weld run.  The effect of fatigue loads on this stress distribution was characterised.

Numerical modelling of fatigue crack closure : 1998-2002.  HEFCE CollR funded

Investigator: Dr LiWu Wei

FE modelling of closure for inclined and kinked cracks was performed.

Fatigue Performance of Welded Joints in 5383 Aluminium alloy : 1999-2002.  Funded by CORUS Research, Development & Technology, IJmuiden and Rotherham

Investigator: Gareth Bradley

This project aims at developing better life prediction models for friction stir (single and double pass) and MIG (high and low heat input) welded butt joints in 5383 aluminium alloy.  Residual stress distributions and microstructure will be characterised, and an attempt made to rationalise small crack growth data for the various weld processes and conditions using a strain-based model.  Incorporation of residual stress information and size effect should allow modelling of S-N data butt-welded specimens to be performed.  The residual stress report is available as a PDF file.   A microstructural report is also available as a pdf file.

Investigation of Tool Travel Speed Effect of Defects in Friction Stir Welding : 2001  Funded by PE Technikon and the National Research Foundation, South Africa.

Investigator: Dr D G Hattingh

This project considered the effect of weld tool travel speed on the defect population and on fatigue life of FS welds in 5083-H321 aluminium alloy.

Investigation of Wake Contact Stresses Developed During Fatigue Crack Growth : 1998-2000.  EPSRC Grant GR/L42391

Collaborative with Professor E A Patterson, Department of Mechanical Engineering, University of Sheffield.

Investigator: Dr Mark Pacey

This project is concerned with a fundamental investigation of plasticity-induced fatigue crack closure.  A Muskhelishvili complex potential stress function for the case of a crack with remote applied stress and crack surface contact, has been set-up and solved.  Full field photoelastic stress patterns for polycarbonate CT specimens are then fitted to the theoretical crack tip stress solution, using genetic algorithm optimisation, to yield values for crack tip K and crack surface contact forces.  Results have been compared with a finite element numerical model of closure in a CT specimen and with compliance-based closure measurements.  Indications are that the work is giving valuable new insight into the origins and import of plasticity-induced closure.  Further information about the work is available, including photoelastic images and video clip from the work.  The final EPSRC report is available in pdf format.

The overall peer review rating by the EPSRC of this outcomes of this research project was "Tending to Outstanding".

Metallurgical Aspects of Fatigue Resistance and Cold Work Embrittlement of Thin Steel Sheets: 1998-2001.   Sub-contractor to Hoogovens Research & Development on ECSC sponsored research project (partners Hoogovens R & D, Thyssen Krupp Stahl, CRM Ghent, IRSID, Voest Alpine).

This project considered the conditions under which intergranular fatigue occurs in low carbon interstitial-free thin sheet steels (1 mm).

Effect of Forming Processes on Fatigue Life of the Centre Disc of an Automotive Wheel:   1998-2000.  Funded by the National Research Foundation of South Africa and supported by Guestro Wheels.  Research project in the Department of Mechanical Engineering at the Port Elizabeth Technikon, South Africa.

Investigator: Pat McGrath

Prediction of fatigue life for new designs of automotive wheel centre discs is difficult to perform because of the plastic deformation and residual stresses induced by wheel forming processes.   Equally, when the production process varies, batches of wheels can be produced which do not meet the empirical fatigue test requirements.  This project examines the effect of the various production stages on residual stress distribution in the centre discs, and on the fatigue performance of the dual-phase steel used for the wheel.