News
Re-Release
The version published two days ago was been replaced by a better one.
The version released two days ago had a couple of annoying bugs (thanks to Katie Penning and Justin Wolfe for finding and help fix them).
It also had major omission which is a new solution to Ainslie's eddy-viscosity model which was kindly contributed by Dr. Mike Anderson of RES. Whilst this new version is not as fast as the Park-based models, it does away with the Crank-Nicolson method and so requires significantly less computation. The bottle-neck now appears to be the summing over the gaussian profile so if anyone has any ideas of how to speed that up then we might be looking at optimising with the EV model in the future (especially now that the energy capture is 8 threaded). Please bear in mind that the fast EV model does give different answers to our regular EV model
and it is quite sensitive to the axial resolution of the calculation. I will leave it to Mike to fill in more details of the derivation of this model.
You may have noticed that some of the more experimental code has been removed. This is now sitting in other packages which I do not distribute as part of the open-source version. This keeps some of my more inane ramblings to myself but it was also the reason that the new EV model was not in Wednesday's release.
I hope you enjoy the new features. Please keep the feedback coming.
Cheers,
Nick
