Trust Simulation

Professor Steve Tsai

The following is excerpted from a recent and compelling conversation between engineers at MSC Software (MSC) and Stanford University Professor Research Emeritus Stephen (Steve) Tsai regarding the work of his international Composites Design Group. 

Steve has been associated with composite materials for approximately 40 years — through his work in developing a nontrivial amount of the theory involved and in broadening the ever-widening range of practical applications.  He is well known for his contributions to the ground-breaking Tsai-Wu failure criterion in the early 1970s which is still widely used for anisotropic composite materials characterized by different strengths in tension and compression.  Today, Steve remains at the forefront in the world of composites and we are grateful for the insights he took the time to share with us based on his broad and deep experiences in this still-emerging field.

MSC: What is the mission of your international Composites Design Group?

Steve: Our ever-expanding global team is intent on pushing the boundary of composites research so we can make the use of composites more competitive in industry versus traditional metal materials.

MSC: Can you provide a brief background and history for this team – what was your strategy when forming this team?  Can new members join? 

Steve: It started in 1998 when we won a National Science Foundation (NSF) grant to research Durability and Life Prediction for composite structures.  Ours was 1 of approximately 50 proposals.  It was fortunate that Professor Yasushi Miyano, from the Kanazawa Institute of Technology in Japan, was on sabbatical at Stanford at the time – he had an excellent approach in this area which was of interest to the NSF, so we teamed with him on this work.

Our broad network enables us to tap into resources & expertise we didn’t have at Stanford.  Professors at the multiple schools in our group have access to state-of-the-art equipment as well as to a range of graduate students and post-docs who contribute broad knowledge.  We have approximately 25 people in our group with a very specific focus – the mechanics of composites. 

We have competed and won several Small Business Innovation Research (SBIR) grants too, and we are a subcontractor for a small company, Kazak Composites of Woburn, Massachusetts, which is a prime contractor for the U.S Army.  The U.S. Army wants easy-to-use life-prediction capabilities for composites.  And they would like our own specialized algorithms and methods, such as ones we developed for micro-mechanics of failure (MMF) packaged in a way that can work with commercial off-the-shelf (COTS) finite element analysis (FEA) software such as MD Nastran.  We are currently doing work in Phase 1 of that SBIR grant; we plan to make a proposal to continue our work through a Phase 2 grant.

We continue to seek members to expand our capability.  This week, for example, we had Jeffrey Fong of NIST join us both in R&D activities as well as to lead a session in our online workshop starting January 7, 2010, on the topic of Error Estimation and Reliability Theory.

Progressive Failure Analysis

MSC: Who do you see as your “target audience” – that is, as the primary beneficiaries of your team’s research?

Steve: Composites engineers in industry.  We also have some success in getting universities involved. We believe engineers in industry can benefit greatly from a more rigorous approach to composites design.  Over-reliance on physical testing is still prevalent.  Followed by correlation.  We appreciate that physical testing is important.  But a computer model is needed to better understand the behavior of composite materials under a wide range of operating conditions: “What is happening to the materials?  Why do they deform?  In what ways do they deform?  Why and how do defects propagate?”

There really is no reason not to rely more on simulation, and to trust simulation.  Even a crude model at the beginning can be helpful in understanding the effects of variations in temperature, moisture, loading, etc à for all these things, nontrivial knockdown factors are used.  Physical tests are typically only performed within a narrow range or type (such as a uni-axial tensile stress test).  Bi-axial physical tests are difficult to perform. Simulations enable us to virtually test a larger number of more realistic configurations.

Naturally, we’d also like to help young engineers get exposed to simulation earlier.  I believe numerical solutions (simulation software) must be taught early on, during their undergraduate engineering years.

MSC: Sometimes, academic research can remain just that – “academic” only.  What is the “value proposition” for your team’s research as it relates to industry?  How do businesses benefit from your work?

Steve: One of our primary goals is to get good composite design tools in the hands of users. So our group’s work is not really about research.  It is applications-oriented.  Composites are precious materials and need high technology to achieve the expected gains in commercial applications – by helping reduce weight and cost.

MSC: “Trust Simulation” is a phrase you have used on several occasions – what is the impact of simulation on the design of Composite structures?  In particular, how much do you see simulation enabling the desired goal of predictive engineering?  Similarly, how does simulation impact manufacturing, specifically the manufacturability of composite products?

Steve: I always strive to encourage more simulation work because I believe simulation is the key to accelerating innovative product development.  Coupling virtual testing with physical testing is the best-possible approach.  The people at MSC Software know that; some engineers at your customer-companies know that.  But more members of product development teams need to leverage numerical models as part of their rationale for making design decisions.  If companies can reduce the amount of physical testing they do, they can save significant amounts of money – many millions of dollars. 

Most of the large aircraft OEMs are still very experimentally focused.  Unfortunately, they don’t yet have sufficient trust of simulation; they still need simulations to help them make better predictions.  That is where our Composite Design Workshops have proven valuable and have been well received.  To-date, we have had almost 900 participants from over 100 organizations – mainly from aerospace, but some from the automotive and wind energy sectors. 

Carbon epoxy is high-tech; it offers the highest performance and represents the ultimate performance that you can achieve.  Naturally, that’s of interest to engineers building products that push the envelope (such as new aircraft and race cars).  The use of composites can also certainly lead to more energy-efficient and long-lasting designs; durability is a key quality of composite material because it doesn’t corrode.

Manufacturing tools should be based on simulation as well.  Ideally, every part of the product development cycle should be backed up by simulation to ensure allowable designs.

Virtual Crack Closure Technique (VCCT)

MSC: For a few years now, your team has conducted periodic web-based Tutorials or Workshops on Composites for engineers in industry every 3-4 months (see http://compositesdesign.stanford.edu or http://www.stanford.edu/group/composites/).  What are the objectives, how have they been received thus far, and what are your future plans for them?

Steve: I have always been interested in doing workshops for composites design.  In the 1970s, I started a series of “Composites Computation Workshops” at UC Berkeley which endured for 17 years.  Then I started to host SAMPE (Society of Aerospace Materials and Processing Engineers) tutorials for a 1-day training session that attracted more than 400 participants in-person.  About 3 years ago, to enable increased attendance, we decided to switch to a web-based format for our Workshops.

Since then, we have conducted 4 Tutorials and 1 Workshop, continually improving the quality of these sessions.  The most recent one, conducted in September 2009, was the best so far judging from the feedback we received from participants.  We’re striving to make the next one even better.

We do not believe that we have saturated the market – we are the only ones conducting live online training in this area and we think there are hundreds, if not thousands, of interested engineers out there.  So we expect to continue at our current pace, with our second Workshop of 2010 scheduled for July.

MSC: How did you get connected with MSC Software?

Kim Parnell, MSC

Steve: We started out using another COTS software tool for FEA, but several of the largest, most prestigious aircraft OEMs participating in our Composite Design Workshops asked us to couple our software with MSC’s MD Nastran software, since that is what they use at their companies.  I reached out to MSC through an old acquaintance of mine in Europe, Yves Lombard, who was able to connect me with the appropriate people in MSC’s Academic Program – they were very eager to help us and developed a special “MD Nastran Edition for Stanford” to enable our Workshop participants to use that software as part of our tutorials. 

In addition, we were introduced to Kim Parnell, a senior technical specialist at MSC who works near Stanford at their Sunnyvale, California, office and is a Stanford alumnus.  Kim has invested significant time to help our group successfully couple our MMF software with MD Nastran for cohesive elements and composite delamination/failure simulations.  It is very valuable to have the codes developed by the Stanford Research team coupled with industrial-strength FEA tools like MD Nastran.  Kim is now a member of our group, he has created several example applications and presented them during our Workshops, and he will participate again during the sessions coming up in January, 2010.  He is encouraging more MSC staff and customers to get involved in our online Workshops.  His focus on delamination is very critical and perhaps the most difficult phenomenon to model at present. 

The relationship with MSC to-date has been really great; we are making continual progress and improvements to the way we collaborate to ensure win-win-win (for both of our groups and for our mutual customers in industry).  Another of our team members, Sung Ha, and I were invited to participate on MSC’s University Advisory Board, newly formed this past year, and we are happy to share our thoughts and recommendations through this group.

MSC: What do you see as the most important next steps for the 3-way collaboration between your team, MSC Software, and your mutual stakeholders in industry?

Steve: We expect to continue our web-based Workshops due to the low costs involved, both from our group’s hosting perspective and from our participants’ perspective (since no travel is required by them).  Our first 4 sessions were billed as “Tutorials” and went on for 12 weeks (with live interactions occurring every few days), but participation tended to fall off during that extended timeframe.  With our 5^th session, we shortened it to 2 weeks for 4 hours each day and people seemed happier with that compressed schedule. This January will be our 6^th session, and we’re now calling them  “Workshops”.  The 7^th session will be in July 2010; we expect interest to continue and expand as the software gets more capable and usage expands.

I would also like us to work together more to raise awareness for the commercial value proposition associated with the simulation of composite structures – there is still so much more value to be gained in this area.  To help with this, we have published a book in October 2008 entitled “Strength & Life of Composites” which supplements the Workshops; we may publish an updated version in the near-future.  We have also teamed with a French organization, JEC Composites, which is an international web platform for the composites industry that hosts related exhibitions twice a year: one in France and one in Singapore, typically in March/April  and October, respectively.

One of our group’s newest members, Jeffrey Fong, has influenced our strategy by encouraging us to pay more attention to experimental data and validation, so we have invited him to present on this topic in our upcoming Workshop.  Additionally, two engineers from industry will talk about optical systems to track full-field deformation.  Fresh topics like these will undoubtedly make our Workshops more compelling and simulation more believable and trusted.

For more information about the upcoming Stanford Composites Design Workshop and the role of MSC Software, see:  http://compositesdesign.stanford.edu

December 22nd, 2009 | Tags: CAE, finite element analysis, MSC, simulation, university, Virtual Prototyping | Category: Academia, Auto, CAE, Industries, Virtual Build & Test Management, Wind Energy