Fix Julia 1.10 CI fails due to hydrostatic tank validation

[efaulhaber]

I could not figure out where the non-determinism comes from. This is the stupid fix now.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 70.49%. Comparing base (71d42df) to head (d3bdb5b).

❗ There is a different number of reports uploaded between BASE (71d42df) and HEAD (d3bdb5b). Click for more details.

HEAD has 1 upload less than BASE

Flag	BASE (71d42df)	HEAD (d3bdb5b)
total	1	0

Additional details and impacted files

@@             Coverage Diff             @@
##             main    #1205       +/-   ##
===========================================
- Coverage   89.88%   70.49%   -19.40%     
===========================================
  Files         133      133               
  Lines       10452    10431       -21     
===========================================
- Hits         9395     7353     -2042     
- Misses       1057     3078     +2021     

Flag	Coverage Δ
total	?
unit	70.49% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[efaulhaber]

@svchb I don't understand why this is so inconsistent. macOS is passing because it's ARM, but why is Julia 1.11 Windows failing?

[svchb]

I kept the tspan and resolution as small as possible to reduce the runtime. This variation should reduce with the longer sampling time.
The simulation result is basically swinging around the analytical value slight shifts through adaptive time stepping modify the frequency of this swinging which have a larger impact on the smaller sampling time.

[svchb]

With the longer sampling time:
Errors for hydrostatic water column 2D:Dict(:wcsph => (1.9001644312676932e-6, 0.026961378735545445), :edac => (1.531217523045389e-6, 0.021726401613458318))

Before:
Errors for hydrostatic water column 2D:Dict(:wcsph => (3.086363804366519e-6, 0.043792327693287766), :edac => (2.1778726591173615e-6, 0.0309017728329325))

so the error reduced by 20-40%.

[efaulhaber]

My concern is the variation between CI runs, not the error itself.

[svchb]

We kept the error bounds currently the same so the error is now well below the current bounds in the validation check.

[efaulhaber]

We still have a problem with inconsistent and especially non-deterministic differences.