Hey Everyone,
I have been working with the THA-KneeBendDemo from the AMMR to introduce patient specific scaling. But I noticed a strange dissbalance between the Joint Space in the left and right knee, as shown in the pictures below.
The top picture shows the Gap in the bend Knee, while the lower one shows the Gap from the standing model. In the standing model we can see that the Gap is nearly not there at all. When the Knee is bend it looks better but than the right Gap is really big. That phenomenon is there in the standard AMMR model, whether it is with FDK or without them. It also occurs when I scale the bones to my needs. Is that just some visual thing or is that something that has to be considered in our results?
You are right there is a difference in the gap between the left and right knees. The right femur is exchanged for a cut femur. Apparently, the cut-bone is not the same as the AMMR bone. This creates the difference. Also, this is purely visual and is introduced in the model through the BoneExchange.any file. Ideally you would like to take the AMMR bone and cut away the part for the hip replacement. Or you can hide the cut bone and visualize the standard AMMR bone. Then, you would see the same gap at the knee.
Please note that this is a demo model that is intended to illustrate some concepts related to FDK. You should be careful about using this model directly as the model is not based on any anatomical data. This is described in the model description also.
Hi Dave,
Thank you for clarifying that. I have two more questions.
We are using the model to introduce custom scaling to both of the femur and tibia. We then want to use the Forces of the knee and the kinematics to compute forces in Abaqus FE. In the desciption of the model they were talking about the lak of anatomical data for the model. Is that just for the placement of the hip implant or is that regarding the whole model?
During the motion of the simulation I noticed a strange sliding movement of both feet. So during the KneeBend both feet slide in the medial direction. Is there any Explanation where that is comng from?
I would recommend you use another model and implement and adapt the relevant code from the THA-KneeBend model. The THA Knee Bend model uses the TLEM1 leg model, which is outdated now. For custom scaling, I would recommend the approach in our tutorial on custom scaling.
Yes, there is an explanation to the sliding feet. It is not real motion, but some combination of mannequin drivers and constraints on center of mass and feet. The motion can definitely be improved, and you are welcome to update the motion. You can look at the Squat model, for example.
Hi Dave,
Thank you for the information.
At first we wanted to use the Squat Model from the AMMR, but the JointReactionForces didn't really make sense for us. For Example we looked at the MedioLateral ReactionForce in the Knee and compared that to experimental data from Orthoload.com. And the Data from AnyBody was not what we expected.
I am not sure about the results for the knee medio-lateral force that you have posted. I just ran the simulation on my pc and here are the results for the medio lateral force that I have from the Squat model. This is on AMMR 3.1.3:
I suspect something has changed in your model and I would recommend you try reinstalling the AMMR to see if you get similar results to what I have posted here.
It's not possible to say if the model is suited for your use case, because I don't know what you want to do with results of the simulation. You have mentioned about using the model for custom scaling. This is alright and you can use the Squat model. Secondly, you mentioned about using the results in Abaqus for finite element analysis. Here, I suppose you are interested in accurate results. This is probably not the case as the motion is parametric and you would be better off with real motion. But, if you are interested in some kind of parametric study, then perhaps parametric motion makes sense as you can have everything in control.
As far as comparison to Orthoload is concerned, a couple of observations. First of all, the motion will not be consistent. You would need to have the same motion data before you can compare the results in detail, but the general trends should be comparable. For example, you are right to question the difference in medio-lateral force that you noted. Secondly, you have to be aware of the coordinate system in which forces are compared. In Orthoload, it looks to be on the shank. In the SelectedOutput folder in AMMR body model, the knee reaction force is expressed in the thigh coordinate system.
