The goal of the project Matthew and I are working on this semester is to create a computational model for dyssynchrony (a medical condition where part of the heart does not beat at the same time or strength as another) in the heart. We are up to our third edition of this model--each edition removes assumptions and adds variables, which increases accuracy of the model as compared to a real heart. We created this model and found that it was not giving us answers that correlate with the physiology of the heart. We thought this meant that there was either a problem with our math, or we made a mistake when constructing the computer model. We fixed a few mistakes, but for about the last ten days, we have been troubleshooting trying to find the problems in the model because it still is incorrect.
We finally brought in our instructor, Dan Ewert, to take a good look at the model. Together, we tried to trace the problems to where we thought they should be. No luck. Then, we fiddled with a few settings to get a clearer picture of some very specific parts of the model, and miraculously, the problem was fixed. The whole time we were looking for problems in the math and the model, but the program was the issue. MATLAB was trying to make logical leaps that caused several of the equations to fall apart.
Our project will now conclude with an accurate model describing a heart with dyssynchrony, and I learned to never eliminate any possible source of error--just because it is supposedly infallible.
Proofed by Dan Huju
No comments:
Post a Comment