Welding Modeling FeNet Nafems

8/10/2019 Welding Modeling FeNet Nafems

1/9

30/12/2014 NAFEMS Procedural Benchmarks For Common Fabrication Detail engineering analysis and simulation - FEA, Finite Element Analysis, CFD, Comput

http://www.nafems.org/about/projects/past_projects/fenet/meetings/majorca04/education/ 1/9

Descript ion Qty Ite m Priceyour bas ket is empty

sub total 0.00 $ proceed t o checkout0 Items - 0.00username login

Forgotten password?

HomeAboutJoinEventsE-learningProfessio nal DevelopmentPu blicationsBlogMembers Ar eaContact

About NAFEMS NewsTe chnic al GroupsRegional Grou psProjectsLegalVacancies

Member BenefitsMembership Types

NAFEMS EventsTraining Courses

NAFEMS World CongressWebinarsIndustry Events

All CoursesCourse ScheduleFlexipassCorporate SolutionsFAQsTestimonials
http://www.nafems.org/e-learning/corporate/http://www.nafems.org/e-learning/schedule/http://www.nafems.org/events/other/http://www.nafems.org/events/training/http://www.nafems.org/events/nafems/http://www.nafems.org/about/legal/http://www.nafems.org/about/projects/http://www.nafems.org/about/tech/http://www.nafems.org/about/media/http://www.nafems.org/about/about_nafems/http://www.nafems.org/contact/http://www.nafems.org/members/http://www.nafems.org/blog/http://www.nafems.org/publications/http://www.nafems.org/e-learning/http://www.nafems.org/join/http://www.nafems.org/about/http://www.nafems.org/http://www.nafems.org/error/passwordreset/http://www.nafems.org/http://www.nafems.org/events/nafems/http://www.nafems.org/e-learning/all_courses/http://www.nafems.org/about/tech/http://www.nafems.org/e-learning/corporate/http://www.nafems.org/e-learning/schedule/http://www.nafems.org/join/member_benefits/http://www.nafems.org/professional_development/http://www.nafems.org/http://www.nafems.org/events/congress/http://www.nafems.org/e-learning/flexipass/http://www.nafems.org/events/webinars/http://www.nafems.org/contact/http://www.nafems.org/blog/http://www.nafems.org/events/training/http://www.nafems.org/about/regional/http://www.nafems.org/about/about_nafems/http://www.nafems.org/checkout/http://www.nafems.org/join/http://www.nafems.org/publications/http://www.nafems.org/about/projects/http://www.nafems.org/http://www.nafems.org/about/legal/http://www.nafems.org/join/membership_types/http://www.nafems.org/members/http://www.nafems.org/e-learning/testimonials/http://www.nafems.org/error/passwordreset/http://www.nafems.org/e-learning/faq/http://www.nafems.org/about/http://www.nafems.org/about/jobs/http://www.nafems.org/events/other/http://www.nafems.org/e-learning/http://www.nafems.org/about/media/http://www.nafems.org/events/


2/9



NAFEMS TrainingPSE Competency Tracker PSE Certification

buy NAFEMS publications benchmark magazinee-librarygerman magazin

PostsContributors

ResourcesManage Your MembershipSubmit News & EventsCAE Jobsite

Feedback Request a Callback

search advanced search

you are hereAboutProjectsPast ProjectsFE-Netmeetings & workshops6th Technology WorkshopProcedural Benchmarks For Common Fa...

COGANFEM ForgingEASIT2SMiRT-23Past Projects

CCOPPSEngineering Skills Management ConsultationForum for Applied MechanicsAutosimILTOFFE-Net

project overviewTechnology FindingsIndustry Findingsmeetings & workshopssearch

NUFRICSAFESAGEM
http://www.nafems.org/about/projects/past_projects/fenet/overview/http://www.nafems.org/publications/browse_buy/http://www.nafems.org/professional_development/certification/http://www.nafems.org/about/projects/past_projects/fenet/advancedsearch/http://www.nafems.org/about/projects/femforging/http://www.nafems.org/about/projects/past_projects/ccopps/http://www.nafems.org/contact/feedback/http://www.nafems.org/about/projects/past_projects/NUFRIC/http://www.nafems.org/professional_development/nafems_training/http://www.nafems.org/about/projects/past_projects/fenet/http://www.nafems.org/members/submit_news_and_events/http://www.nafems.org/blog/contributors/http://www.nafems.org/about/projects/past_projects/fenet/industry/http://www.nafems.org/professional_development/competency_tracker/http://www.nafems.org/about/projects/past_projects/fenet/technology/http://www.nafems.org/members/resources/http://www.nafems.org/about/projects/past_projects/fenet/http://www.nafems.org/about/projects/past_projects/GEM/http://www.nafems.org/about/http://www.nafems.org/about/projects/past_projects/http://www.nafems.org/about/projects/cogan/http://www.nafems.org/publications/magazin/http://www.nafems.org/about/projects/past_projects/fenet/meetings/http://www.nafems.org/publications/e-library/http://www.nafems.org/about/projects/past_projects/http://www.nafems.org/members/manage_your_membership/http://www.nafems.org/about/projects/past_projects/FAM/http://www.nafems.org/blog/posts/http://www.nafems.org/about/projects/past_projects/fenet/meetings/majorca04/http://www.nafems.org/about/projects/past_projects/fenet/meetings/majorca04/education/http://www.nafems.org/about/projects/past_projects/autosim/http://www.nafems.org/about/projects/past_projects/ILTOF/http://www.nafems.org/contact/callback/http://www.nafems.org/search/http://www.nafems.org/about/projects/past_projects/fenet/meetings/majorca04/education/http://www.nafems.org/members/caejobsite/http://www.nafems.org/about/projects/past_projects/fenet/meetings/http://www.nafems.org/about/projects/http://www.nafems.org/about/projects/past_projects/SAFESA/http://www.nafems.org/publications/benchmark/http://www.nafems.org/about/projects/smirt23/http://www.nafems.org/about/projects/easit2/http://www.nafems.org/about/projects/past_projects/esm/


3/9



Procedural Benchmarks For CommonFabrication Detail

A summary report of the FENet Education and Dissemination(E&D) workshop in Majorca, Spain, 25th March 2004.

By Dr. Jim Wood, University of Strathclyde, UK

This article provides a summary of the observations made at the Education andDissemination session in Majorca, which was reasonably well attended andresulted in 3-4 hours of stimulating discussion.

The three procedural benchmarks are shown below.

Results obtained from all contributors are available for download for each benchmark, as well as results and the benchmarks themselves. The followingobservations are based on the results contained in these spreadsheets and also fromthe more comprehensive submissions made by participants.

Fig. 1 Procedural Benchmark FENET_E&D1 Shell Intersection

Benchmark Download (PDF)

Results Download (PDF) Local Stresses

Stiffness, Overall, Field

Presentation Download (PDF)
http://www.nafems.org/downloads/FENet_Meetings/Majorca_Spain_Mar_2004/FENET_Majorca_March2004_Procedural_Benchmark_ed1.pdfhttp://www.nafems.org/downloads/FENet_Meetings/Majorca_Spain_Mar_2004/FENET_Majorca_March2004_EDU1_Local_Stresses.pdfhttp://www.nafems.org/downloads/FENet_Meetings/Majorca_Spain_Mar_2004/FENET_Majorca_March2004_EDU_Wood2_ED1.pdfhttp://www.nafems.org/downloads/FENet_Meetings/Majorca_Spain_Mar_2004/FENET_Majorca_March2004_EDU1_Stiffness_Overall_Field.pdf


4/9



Fig 2 Procedural Benchmark FENET_E&D2 Shell Reinforcement

Benchmark Download (PDF)

Results Download (PDF) Local Stresses Field Stresses

Stiffness & Overall Stresses


Fig. 3 Procedural Benchmark FENET_E&D3 Offset Shell Mid-surface

Benchmark Download (PDF)Results Download (PDF)

Local Stresses Stiffness, Overall & Field Stresses


Generic observations
http://www.nafems.org/downloads/FENet_Meetings/Majorca_Spain_Mar_2004/FENET_Majorca_March2004_EDU3_Stiffness_Overall_Field.pdfhttp://www.nafems.org/downloads/FENet_Meetings/Majorca_Spain_Mar_2004/FENET_Majorca_March2004_EDU3_Local_Stresses.pdfhttp://www.nafems.org/downloads/FENet_Meetings/Majorca_Spain_Mar_2004/FENET_Majorca_March2004_Procedural_Benchmark_ed2.pdfhttp://www.nafems.org/downloads/FENet_Meetings/Majorca_Spain_Mar_2004/FENET_Majorca_March2004_EDU_Wood3_ED2.pdfhttp://www.nafems.org/downloads/FENet_Meetings/Majorca_Spain_Mar_2004/FENET_Majorca_March2004_EDU2_Local_Stresses.pdfhttp://www.nafems.org/downloads/FENet_Meetings/Majorca_Spain_Mar_2004/FENET_Majorca_March2004_EDU2_Stiffness_Overall.pdfhttp://www.nafems.org/downloads/FENet_Meetings/Majorca_Spain_Mar_2004/FENET_Majorca_March2004_Procedural_Benchmark_ed3.pdfhttp://www.nafems.org/downloads/FENet_Meetings/Majorca_Spain_Mar_2004/FENET_Majorca_March2004_EDU_Wood4_ED3.pdfhttp://www.nafems.org/downloads/FENet_Meetings/Majorca_Spain_Mar_2004/FENET_Majorca_March2004_EDU2_Field_Stresses.pdf


5/9



1. Significant variation in the modelling, results, assessment and conclusions relating to fitness for purpose of such detail is apparent across analysts and industry sectors, for both static and fatiguesituations.

2. Human error is apparent, including:

a. Mis-interpretation of boundary conditions;

b. Incorrect use of finite element functionality that altered physical response;c. Reporting wrong results;

d. Using stress output directly at singularities;

e. Using averaged stresses at shell intersections.

3. Lack of engineering common sense is apparent. For example, not all contributors checked that thefield stresses compared well with hand calculations, where appropriate.

4. From this limited linear elastic exercise:

a. The need for finite element knowledge is confirmed;

b. The need for general engineering education is confirmed;

c. The need for industry specific knowledge is confirmed;

d. The need for validation is confirmed;

e. The need for adequate QA procedures is also confirmed.

5. Established common best practice across the various industry sectors is not apparent.

6. Various guidelines exist for weld modelling and assessment in few industry sectors.

7. Use of experimentally derived results on real weld geometries is recognised as a necessary part of theassessment process for fatigue. It is how FEA results are obtained (often at locations where singularitiesexist) that provides variations in approach.

8. As with all analyses, the adequacy of any idealisation must be judged in terms of the purpose of theanalysis being conducted. The idealisation of such fabrication detail will affect static and fatigue (anddynamic, buckling, limit and fracture assessments) to different degrees.

9. The influence of fabrication detail can be local or global in nature and this fact should be consideredwhen judging adequacy. The details selected for consideration as procedural benchmarks have bothlocal and global measures selected as targets and all showed variation.

10. It is possible to use shell models to obtain necessary stress data for fatigue assessment of such details, but care and understanding is necessary. It should also be recognised that shell models will produce finiteconverged results at intersections.

11. There would appear to be two distinct approaches to obtaining hot-spot stresses from finite elementmodels:


6/9



a. A stress linearization procedure (not required with shell representations), designed to removethe peak stress component and leave the membrane and bending stress components. Such anapproach will include gross geometric stress concentration effects. Some finite element systems

provide post-processing tools for defining the assessment section in both 2D and 3Drepresentations and for linearising the results. This approach is common in the Pressure Vesselindustry.

b. A number of extrapolation approaches, with slight variations in the extrapolation procedures,

are in use. These approaches invariably involve element sizes of the order of 0.4t and differ inwhether linear or quadratic extrapolation is used to the hot-spot location (which is often asingularity in nonshell models). Some of these procedures also provide details of how the stressdistributions from shell models should be displaced by up to half a shell thickness (depending onthe angle of the intersection), before extrapolating to the hot-spot. However, it is recognised thatspecific Codes of Practice may not give the analyst any choice in which procedure to adopt.

12. Some fatigue assessment procedures require the use of the stress range on the weld throat area.Stresses plotted across the throat will show a highly non-linear variation. Although not always clear, it islikely that an average value of stress is required from the results, to provide consistency with handcalculations. Although not always clear, it is likely that a simple membrane + bending value of stress,with peak component removed, is required from the results, to provide consistency with handcalculations.

13. Some analysts used thick shell elements. The fact that most did not, would perhaps indicate that participants from different disciplines do not have a common understanding of when plates and shells become thick.

14. Given the nature of the Displacement Finite Element Method and the details examined, it is perhapsnot surprising that greatest agreement is apparent for global stiffness ( as measured by overalldisplacements), closely followed by field stresses (by definition away from local stress concentrations).Greatest variation is apparent for local stresses, which include finite values derived from distributions inthe vicinity of singularities. In addition, as would be expected, greatest variation is apparent for verysmall target values, with best agreement generally for large values.

Preliminary observations for E&D1

1. For this detail, the practice of displacing shellstress distributions by half a shell wall thickness

before interpolating values at the notional position of weld toes (hot-spot), would seemunnecessary.

2. The various results provided by AnalystIdentifier 1 show remarkably little variationamongst 2D-Axi and shell models (with andwithout weld representation). Whilst inclusion of the weld stiffness (by whatever means) improvesthe comparison with the highly refined 2D-Axiresults, it is not apparent that this additional complexity is merited over a simple shell representation and


7/9



use of stresses at the location corresponding to the weld toe.

3. Although not considered in these benchmarks, it is noted that a simple shell intersection representationalready has too much mass, without the addition of any measures designed to include the effect of theweld. This will have some bearing on dynamic analyses and weld models will result in a greater need toreduce the density of local elements for accurate representation.


1. The reason for the large variationin contact radius results for loadcase 2 is not apparent. This isclearly a function of global stiffnessrepresentation as well as theeffectiveness of the contact methodsused. Given the relatively good

agreement on overall deflections inmost cases, it must be assumed thatthe differences are largely due to thecontact methods. This factemphasises the need for adequate contact benchmarks.

2. The various results provided by Analyst Identifier 1 show remarkably little variation amongst D-Axiand shell models for deflections, field stresses and weld-toe stresses, for load case 1. The poor comparisons for load case 2 are due to the lack of contact simulation in the shell model for load case 2.

3. Assuming the reinforcing plate to be integral did not provide good agreement for field stresses at the

plate centre. The comparison of local stresses in the region of the weld are reasonable, particularly for those of larger magnitude. It is clear therefore, that if such an assumption is to be made, then care must

be taken to ensure that both plates effectively act as one through use of a suitable number of spot or puddle welds. The results for the central spot-weld idealisation, would indicate that an integral behaviour may be possible with relatively few plate connections.

4. Neglecting the offset due to the reinforcing plate and assuming a double thickness integralrepresentation over the reinforced area produced similar results to the integral idealisation with offset.


1. Only two participants showedthat the problem was largedisplacement (subsequentlyconfirmed by the coordinator), inspite of the tip deflection being lessthan the thickness of the plate.

Reductions in deflections and


8/9



stresses are significant. The rules of thumb commonly used as a guide to when large displacement effects become significant for beams, plates and shells are clearly not applicable for this problem. The reason for this is apparent when the source of the non-linearity is given due consideration.

2. 3D models (shells and bricks) show variations in results across the width, which are obviously absentfrom 2D results. Not all contributors commented on this effect.

3. Neglecting the offset, even with correct plate thicknesses, fails to predict adequate values for overall

stiffness, field stresses and local stresses. Analysts should therefore think carefully before neglectingoffsets in plate/shell mid-surfaces, as their effects can have a global nature as well as local. For thinner plates/shells, large displacement effects may act to reduce the global effect of the offset, through local bending of the joint and effective realignment of the midsurfaces.

4. The modelling of contact between the lapped plates appears irrelevant for the relative joint sizesconsidered. The results for separate plates, with and without contact, and models where the plates wereassumed integral appear similar. The latter model however, fails to pick up the stress singularity thatexists at either end of the lap running between fillet weld roots. Almost all contributors failed to highlightthis singularity, in any model.

Closure

The general consensus of those participating was that this was a worthwhile exercise, that should be of interest to the wider FE community. Given the level of participation, this fact was clearly not alwaysrecognised. It should be recognised that the relevance of the general observations clearly have broadinterest to the entire finite element community, it should also be recognised that the relevance of thedetail of the modelling strategies used is not confined to those involved in welding 15mm thick steel

plate! It is now my task to try and pull together some sort of conclusions from this exercise. To this end, I

have asked contributors to have a look at the final collated results and take the opportunity to commenton the results that they provided in relation to the rest. In addition, they have also been asked to have alook at the preliminary observations and to comment on these and add their own if necessary.

On receipt of everyones comments, I will then document the exercise. The final document is likely toconsist of several parts. An Introduction providing the background and aims of the project; theProcedural Benchmark Specifications; the Round- Robin results (with anonymity maintained);observations from the exercise and finally some conclusions regarding good and bad practice in theanalysis and assessment of such details.

This final report is now available to download, as well as being available from NAFEMS in hard-copy format.

The download is available in a lo-res PDF format (1.8MB), as well as a hi-res format, in a zippedpackage (22MB).

AboutAbout NAFEMS

NewsTechnical GroupsRegional GroupsProjects
http://www.nafems.org/about/regional/http://www.nafems.org/downloads/FENet_Deliverables/Education_and_Dissemination/FENET_D5603_Hi_Res_22MB.ziphttp://www.nafems.org/about/http://www.nafems.org/downloads/FENet_Deliverables/Education_and_Dissemination/FENET_D5603-Proc_Benchmarks_for_Common_Fabrication_Details.pdfhttp://www.nafems.org/about/tech/http://www.nafems.org/about/media/http://www.nafems.org/about/about_nafems/http://www.nafems.org/about/projects/


9/9


http://www nafems org/about/projects/past projects/fenet/meetings/majorca04/education/ 9/9

LegalVacancies

JoinMember BenefitsMembership Types

Events

NAFEMS EventsTraining Courses NAFEMS World CongressWebinarsIndustry Events

E-learningAll CoursesCourse ScheduleFlexipassCorporate Solutions

FAQsTestimonials

Professional Development NAFEMS TrainingPSE Competency Tracker PSE Certification

Publications buy NAFEMS publications benchmark magazinee-librarygerman magazin

BlogPostsContributors

Members AreaResourcesManage Your MembershipSubmit News & EventsCAE Jobsite

ContactFeedback Request a Callback

Nafems 2014developed by duo design
http://www.nafems.org/blog/http://www.nafems.org/contact/callback/http://www.nafems.org/events/nafems/http://www.nafems.org/events/congress/http://www.nafems.org/publications/browse_buy/http://www.nafems.org/contact/feedback/http://www.nafems.org/members/caejobsite/http://www.nafems.org/members/submit_news_and_events/http://www.nafems.org/contact/http://www.nafems.org/publications/benchmark/http://www.nafems.org/professional_development/competency_tracker/http://www.nafems.org/join/http://www.nafems.org/e-learning/all_courses/http://www.nafems.org/about/legal/http://www.nafems.org/e-learning/schedule/http://www.nafems.org/events/other/http://www.nafems.org/members/resources/http://www.nafems.org/publications/magazin/http://www.nafems.org/events/http://www.nafems.org/join/member_benefits/http://www.nafems.org/e-learning/testimonials/http://www.nafems.org/e-learning/http://www.nafems.org/e-learning/flexipass/http://www.nafems.org/professional_development/nafems_training/http://www.nafems.org/professional_development/http://www.nafems.org/join/membership_types/http://www.nafems.org/about/jobs/http://www.nafems.org/professional_development/certification/http://www.nafems.org/e-learning/corporate/http://www.nafems.org/members/manage_your_membership/http://www.nafems.org/e-learning/faq/http://www.nafems.org/publications/http://www.nafems.org/events/training/http://www.nafems.org/blog/contributors/http://www.nafems.org/publications/e-library/http://www.nafems.org/members/http://www.nafems.org/blog/posts/http://www.duodesign.co.uk/http://www.nafems.org/events/webinars/

Documents

Welding Modeling FeNet Nafems