Model Benchmark Summary

This section summarizes the steps involved in benchmarking a finite element knee model. The summary includes links to other chapters where specific details are provided.

These steps apply to the OpenKnee(s) data set. The documentation for the DU data set can be found in a separate document.

Knee Model - Initial Deviations

The knee model that was delivered for the Model Calibration phase will be recalibrated and used for the benchmarking phase. This step is being taken to minimize the effects of group-specific interpretation of the calibration data, where new processed experimental data was made available at the beginning of the Model Benchmarking phase. If modeling assumptions were found to be inconsistent with the intended use of the calibration data the teams were asked to recalibrate their models using these processed data. In our instance, there will be changes to the model’s fixed femoral coordinate system, and corresponding changes to the experimental kinematics that are used. These changes are summarized below and documented in more detail in the Initial Deviations section.

Coordinate Systems

There are two primary ways the knee model’s fixed coordinate systems will be updated. The first will change the way that the coordinate systems were defined. Rather than using the transforms reported in the experimental state.cfg file, digitized landmarks will be used to define the fixed femoral and tibial coordinate systems. The second change is in the selection of the fixed femoral coordinate system. The state.cfg file reports two femoral coordinate systems, one which is used to determine kinematics using Optotrak sensor positions, and the other is used to determine kinematics using the robot’s end effector position. Due to these differences, two sets of kinematics are reported in each .tdms file. The teams were asked to use kinematics defined using the Optotrak sensor positions, which differs from the kinematics used by the CSU/WSU team in the Model Calibration phase. Due to this, the knee model’s fixed femoral coordinate system was changed from that used in the Model Calibration phase. See the Coordinate Systems section for more information.

Experimental Data

There are two primary ways that the experimental data used for model calibration will change. First, rather than parsing data from the given .tdms files, the experimental data has already been parsed and given in a .csv file. Second, the kinematics that will be used to calibrate the knee model will use a different set of experimental kinematics than was used in the Model Calibration phase. See the Experimental Data section for more information.

Model Benchmarking

Benchmarking Simulations

The processed data delivered by the Cleveland Clinic team for “combined loading” will be run using the model that was recalibrated following the application of the initial deviations. This data consists of measured kinematics at 0, 30, 60 and 90 degrees flexion for simultaneously applied moments of 10 Nm valgus and 5 Nm internal (i.e. “combined” loads). These loads are defined in the joint coordinate system (JCS) and were parsed from the .tdms file and provided by the Cleveland Clinic team in a .csv file (see the download package on the KneeHub SimTK site).

The recalibrated model will kinematically control the flexion angle and apply the experimentally measured forces to the tibia’s fixed coordinate system in all six degrees of freedom. See the Benchmark Simulations section for more information. The model’s predicted kinematics, contact mechanics, and ligament loads will be recorded and kinematics results will be compared to the measured values using root-mean-square (RMS) errors for each degree of freedom.