Cleveland Clinic - University of Utah - Columbia University Meeting @ SB3C 2015

Date: June 18, 2015

Time: 4:30 PM EST

Means: In person meeting

Attendees:

1. Ahmet Erdemir (Cleveland Clinic)
2. Craig Bennetts (Cleveland Clinic)
3. Ben Ellis (University of Utah)
4. Steve Maas (University of Utah)
5. Gerard Ateshian (Columbia University)

Agenda:

This was a free form meeting to bring team members from collaborating institutes for face-to-face interactions. There was no set agenda; rather, topics on development of FEBio features for Open Knee(s) were discussed.

Immediate Action Items:

• Cleveland Clinic (Ahmet)
  • Reproduce Grood and Suntay linkage system in Abaqus for cross-verification.
• University of Utah (Steve)
  • Complete in situ strain feature publication.
• Columbia University (Gerard)
  • Provide Cleveland Clinic team FEBio input file for Grood and Suntay linkage system.

Notes:

1. Ahmet described Gerard and other participants FEBio feature requests from the Open Knee(s) team. Status of the development in each feature were noted along with the necessary steps to test the features. Many of the primary features were implemented. In situ strain feature is mature and Steve has been working on a related publication and would present this work in SB3C in upcoming days. The recent FEBio (version 2.3) provides the plug in feature to activate this feature. The necessary plug in is not disseminated to public yet but it is available for the Cleveland Clinic team to test. Many convenience features were already implemented, e.g. sets. The Cleveland Clinic team needs to conduct testing to evaluate these features.
2. Gerard recently completed implementation of general purpose joints to prescribe relative kinematics/kinetics between rigid bodies. He provided many examples illustrating the use of these joints by prescription of joint axes and kinematic constraints. His implementation relies on penalty formulation to ensure kinematic constraints are enforced. He also implemented the Grood and Suntay linkage system that will be helpful for knee simulations. Test simulations with this system works when linkage rotations are prescribed on an individual basis but sometimes convergence difficulties are observed when combined rotations are prescribed. Ahmet noted that this may be a result of kinematic linkage instability, e.g. gimbal lock. He will implement the same linkage chain in Abaqus to see if similar convergence problems exist. Gerard will provide the FEBio input file for reference. Gerard also informed that these simulation can be conducted with FEBio version 2.3.
3. Many secondary feature requests were also discussed. Ahmet asked for the capability for penetration based contact, tension only springs, wrapping of spring elements, and Python scripting. FEBio development team informed the Cleveland Clinic group that features for penetration based contact (between rigid bodies) and tension only springs already exist. Steve has an interest in Python scripting, particularly from the perspective of getting people use others' models more easily. Gerard and Ben discussed potential strategies to implement spring wrapping and the need for it in simulation cases other than Open Knee(s).
4. Craig described convergence optimization studies he has been conducting. These simulations are done in a scripted fashion on a high performance computing infrastructure allowing for large scale parameter sweeps. They will allow fine tuning of FEBio simulation parameters to decrease simulation cost and/or to allow convergence of difficult problems.