Inconsistency between the TF coil radial build in machine_build.f90 and sctfcoilf90

[jonmaddock]

In GitLab by @schislet on Sep 25, 2020, 10:41

Summary

There is an inconsistency between the TF coil radial build calculation and in machine_build.f90 and sctfcoil.f90 (lines 3793 to 3825 as of 25/09/2020 on develop). This was introduced with the changes implemented alongside issue #1048. The

The code in sctfcoil.f90 is comparing the diagonal length of the TF coil (r_tf_inboard_mid) with the horizontal radius and flagging an error because they are different - which one would expect in a wedge shaped coil.

The inconsistency leads to an error message in the outfile suggesting that the TF coil should have a greater radial thickness than it actually has (The error's not listed in the errorlist.json). This doesn't affect the output in any way, but it is rather distracting if the user doesn't know why this 'error' is coming up. Note that it crops up even when running test suite files (see below).

2017_baseline_test_OUT.DAT

@skahn @schislet @mkovari

Checklist

After implementing issue do the following

• Run test suite
• Does it change the output? If so, give justifications.
• Does it change the baseline 2019 run? (you may need to run on Freia)
• Merge request created

[jonmaddock]

In GitLab by @mkovari on Sep 25, 2020, 10:44

The text of the issue implies that the error message is incorrect. This has not been demonstrated. When an error message appears it is wise to start by assuming that there is indeed an error.

[jonmaddock]

In GitLab by @schislet on Sep 25, 2020, 10:50

@skahn and I are quite convinced that the error message is indeed incorrect (for instance it still appears when running test files e.g. 2017 DEMO baseline output below). The code in sctfcoil.f90 is comparing the diagonal length of the TF coil (r_tf_inboard_mid) with the horizontal radius and flagging an error because they are different - which one would expect in a wedge shaped coil.

I'll add this into the error description.

2017_baseline_test_OUT.DAT

[jonmaddock]

In GitLab by @skahn on Oct 1, 2020, 14:45

Hello Michael, this error arrises from a check on the middle inboard TF direction (the shortest) done using tfcth taken now as the maximum TF radius, and hence the diagonal direction.

Therefore, I think this error should be corrected to use a consistent TF raduial definition.

See the TF geometry issue : https://git.ccfe.ac.uk/process/process/issues/1048

[jonmaddock]

In GitLab by @skahn on Oct 1, 2020, 14:47

PS : @mkovari, is your picture profile a tokamak diagnosic window ?

[jonmaddock]

In GitLab by @mkovari on Oct 1, 2020, 14:50

I suggest you don't remove or disable the error check, but replace it with one that is correct.

A diagram here in the issue will clarify the final correct code.

@jmorris-uk

[jonmaddock]

In GitLab by @skahn on Oct 1, 2020, 15:15

Other question : Shouldn't this error be captured by the PROCESS error system rather that being a error message lost in the OUT.DAT ?

If so I would be happy to make it a lvl 2 error on the error list

[mkovari]

Will you put me down as a reviewer? @timothy-nunn