bug in core.rz2psinorm

[nicolavianello]

Hi I think I found a bug in the computation of psinorm. It actually occurs whenever operating in reverse Ip case (for example whenever we operate with positive Ip and Bt which means counter clock-wise from the top)
Indeed in rz2psi the flux is multiplied by the factor -1_self.getCurrentSign(). After that the computation
is done getting the value of the flux on axis with self.getFluxAxis() which is not multiplied by the same factor. Indeed if you try to evaluate the flux at the radial vertical position of the axis as
self.rz2psi(Rax,Zax,time) in case you are running with positive Ip you get an opposite value respect to self.getFluxAxis().
I guess it can be simply solved by multiplying in line 826 of core.py both psi_boundary and psi_0 for -1_self.getCurrentSign().

[markchil]

Thank you for looking into this. This (and a few other locations we use getCurrentSign() in core.py) are done to correct for some inconsistencies in the signs of EFIT's output: some things it gives with the correct sign, other things it always puts out positive, even when the correct sign is negative. I think I should probably figure out a way to specify where the corrections are necessary in EFIT.py so that core.py gives the correct outputs for inputs with the correct sign.

[jcwright77]

This is a constant headache for all EFIT users. We’re dealing with this issue right now in one of our codes. AKAIK, many if not all plasma simulations using EFIT equilibria have internal consistency checks on signs to correct them.
-john

On May 31, 2016, at 12:33 AM, markchil [email protected] wrote:

Thank you for looking into this. This (and a few other locations we use getCurrentSign() in core.py) are done to correct for some inconsistencies in the signs of EFIT's output: some things it gives with the correct sign, other things it always puts out positive, even when the correct sign is negative. I think I should probably figure out a way to specify where the corrections are necessary in EFIT.py so that core.py gives the correct outputs for inputs with the correct sign.

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub #3 (comment), or mute the thread https://github.com/notifications/unsubscribe/AH8lzUZ7oRFisHX58SmcZz7yE_pvkcF1ks5qG7owgaJpZM4Iohus.