By Chris Rogers, CREA Consultants Limited, Buxton, UK.
Quality Assurance has been a recurring theme within the FENet discussions over the last 2½ years or so. The subject has been raised in one form or another in just about all of the workshop and seminar sessions throughout the FENet project. Sitting in on sessions related to fatigue, multiphysics, civil engineering, marine, education and dissemination, and many others, Quality Assurance has raised its head in one way or another. Validation has not surprisingly been the most talked about topic; however, other topics have also been to the fore. These include:
The intention of the workshop was to identify one or more QA issues that can be taken forward and developed into a FENet deliverable. Preferably, the issue(s) identified should fit into the FENet overarching scope of “QA as a barrier to the further uptake of computer-aided numerical analysis”.
It should be noted that QA and analysis management are subjects that were not included in the original FENet scope, but they have emerged as key subjects over the course of the project. To a certain extent, the mini workshop was designed to remedy this situation.
The format of the workshop was as a brain storming session, where the attendees discussed QA and analysis management. One intention of the session was to attempt to expand on previous discussions within FENet where differences between the different professional structures across Europe have been aired.
As QA has not had a formal place within FENet to date, the discussions were initially broad ranging, to allow the important issues to come forward. As it is well known that verification and validation are the most significant issues to have been highlighted throughout FENet, the decision was taken to exclude V&V from discussions.
The workshop therefore examined issues based on following:
As the major intent of the workshop was to brainstorm the topic, presentations were kept to a minimum. Excluding the chairman’s introduction, there were three presentations, each of which helped to set the scene.
Olivier Tabaste presented a paper on the sensitivity analysis of weld failure, this showed the link between quality of analysis and quality of product.
Jack Reimers presented thoughts on the general subject of QA.
Althea de Souza presented on the contrast between QA in FEA and CFD.
The most important part of the workshop was the discussion on QA and how it influences the analysis process.
A discussion on the meaning of QA as distinct from product quality, as this is perceived as being an important concept. The discussion examined what is meant by quality, and what is meant by quality assurance. We all know of measures of quality. In terms of the cars we drive, there are many quality bands, but we have the basic, super luxury (Rolls Royce, Daimler, etc), the executive (VW, Audi, Jaguar, etc), family (Ford, GM, VW, Fiat) and the economy. This is a measure of quality on the basis of personal taste and wealth. However, quality also reaches into the world of reliability, how many people do we know who run around in twenty plus year old cars, where the odometer is on its second, third or even fourth time around?
The term Quality Assurance appears to imply that the process ensures that we produce Rolls Royces, but is that what it means? This focussed our discussions. Several threads of discussion centred on topics related to the quality management process.
The NAFEMS Quality System Supplement (QSS) needs updating to suit ISO 9001:2000. This is a task that will be placed on the Analysis Management Working Group of NAFEMS, see the end of this article for more information.).
We examined the different needs of FEA and CFD, and also looked at the implications of CFD being about 10 years behind FEA. We saw that CFD is facing many of the problems that FEA has experienced, including the dreaded encapsulated solvers. There is clearly a need for the FEA community to pass on to CFD the lessons learnt to close the 10- year gap. In its early life NAFEMS concentrated on FEA, hence the original meaning of the acronym.
However, we now have many diverse numerical analysis fields, ranging across science and industry, applied using paper and pencil, through calculators, to PCs and super-computers. It was clearly decided that discussions should expand to encompass numerical analysis in general, though engineering will be the underlying basis.
The conclusion of the workshop was that it is time to look beyond the analyst community, the major consideration should be:
We discussed the need to prepare procedures for design engineers and graduates, as most analysis procedures are written for experts.
And, we discussed the need for design engineers and draughtsmen to take responsibility for the output from embedded systems.
The suggested deliverable is one or more booklets aimed at:
The final suggestion is that it would be useful if time could be made available during the final stages of FENet to hold a Management and Quality seminar/ workshop to focus on the three key subjects, from which a deliverable can be produced.
Time has been made available to discuss analysis management and QA further. In consultation with John Smart it has been decided to allocate some time to analysis management and QA during the Education and Dissemination workshop at the next FENet meeting in Glasgow. As a minimum, we will get 30 minutes or so, but if the subject proves popular, this can be extended. There will also possibly be a special lunch area for the discussion of QA and analysis management. Clearly this depends on interest and available space. Thanks in advance to John for allowing the time.
The old NAFEMS QA Working Group has been resurrected, but with a new name, and a wider remit. The new name is the Analysis Management Working Group.
Anyone interested in finding out more, or just interested in a copy of the meeting minutes should contact either Chris Rogers (firstname.lastname@example.org ), or the NAFEMS office.