Fuselage cannot have z offset in front of the wing/fuselage junction (towards the nose)

[kccwing]

Describe the bug
as described - structural solver nonconvergence observed

To Reproduce
branch dev_fuselage_bug - simple_HALE/generate_hale_nose.py vs simple_HALE/generate_hale_nose_dip.py

Expected behaviour
convergence and time-stepping until completion

Screenshots
|==========|==========|==========|==========|==========|==========|==========|==========|==========|==========|
| iter |alpha[deg]|elev[deg] | thrust | Fx | Fy | Fz | Mx | My | Mz |
|==========|==========|==========|==========|==========|==========|==========|==========|==========|==========|
Residual is: 3.3313240830999096E-006
STOP Static equations did not converge (17235)

System Info (please complete the following information):

• OS: [e.g. CentOS7/MacOS 13.1] HPC
• SHARPy Version - latest

Additional context
prime suspect - CRV in this particular direction? Will do a bit more digging later - just starting this to keep it logged

[kccwing]

dev_fuselage_bug

[ben-l-p]

Only just looking at this now after we spoke before Christmas. I was having issues with the trim solver which I didn't get around to addressing, where the derivatives seem to be zero when they should be - do these issues also occur for dynamic cases?

[kccwing]

by derivatives you mean the gradients for the trim inputs from the Newton loop?

[ben-l-p]

Yes I do - the incorrect gradient caused some of the states for the next iteration to be something unreasonable which caused the structural solver to not converge. Might be a different issue

[kccwing]

hmmm incorrect as in diff in output/diff in input is incorrect or the result of input+output/gradient is too far away from the solution point? Assuming the former, changing the relaxation factor does help (some sort of a "step size" parameter), but only if you fix this bug

                self.input_history[self.i_iter][i_dim] = (self.input_history[self.i_iter][i_dim]*(1 - self.settings['relaxation_factor']) +
                                                          self.input_history[self.i_iter][i_dim********-1********]*self.settings['relaxation_factor'])