Blowing and suction on wall boundary conditions

[Anthony Larroque]

Hello,

I would like to simulate with PyFR the flow around a cylinder with blowing and suction on the wall. I saw that in the user guide of PyFR, for the wall boundary conditions, it was only possible to specify the velocity with a float.

Thus, is there a way to specify blowing/suction conditions on the wall which could depend on the position and the time ?

Kind regards,

Anthony Larroque

[Freddie Witherden]

Hi Anthony,

Yes. Expressions, which can be functions of space and time, are
permitted in boundary conditions in addition to floats.

Regards, Freddie.

[Anthony Larroque]

Dear Freddie,

Thank you very much for your answer.

However, it does not seem to work with my setup.

I am using the last version of PyFR that I found on: https://github.com/vincentlab/PyFR and run the example of the cylinder with simply adding in the inc_cylinder_2d.ini the velocities u = 0.1*x and v=0.1*x in the [soln-bcs-wall] to try an expression that depends on space.

When I set a constant for the velocities like u = 0.1 and v = 0.1, I do not have a problem to run the simulation. However, with the expressions mentioned before, I have the following error message:

Traceback (most recent call last):
  File "/home/anthony/anaconda3/bin/pyfr", line 11, in <module>
    load_entry_point('pyfr==1.8.0', 'console_scripts', 'pyfr')()
  File "/home/anthony/anaconda3/lib/python3.6/site-packages/pyfr/__main__.py", line 110, in main
    args.process(args)
  File "/home/anthony/anaconda3/lib/python3.6/site-packages/pyfr/__main__.py", line 235, in process_run
    args, NativeReader(args.mesh), None, Inifile.load(args.cfg)
  File "/home/anthony/anaconda3/lib/python3.6/site-packages/pyfr/__main__.py", line 216, in _process_common
    solver = get_solver(backend, rallocs, mesh, soln, cfg)
  File "/home/anthony/anaconda3/lib/python3.6/site-packages/pyfr/solvers/__init__.py", line 16, in get_solver
    return get_integrator(backend, systemcls, rallocs, mesh, initsoln, cfg)
  File "/home/anthony/anaconda3/lib/python3.6/site-packages/pyfr/integrators/__init__.py", line 36, in get_integrator
    return integrator(backend, systemcls, rallocs, mesh, initsoln, cfg)
  File "/home/anthony/anaconda3/lib/python3.6/site-packages/pyfr/integrators/dual/phys/controllers.py", line 8, in __init__
    super().__init__(*args, **kwargs)
  File "/home/anthony/anaconda3/lib/python3.6/site-packages/pyfr/integrators/dual/phys/steppers.py", line 8, in __init__
    super().__init__(*args, **kwargs)
  File "/home/anthony/anaconda3/lib/python3.6/site-packages/pyfr/integrators/dual/phys/base.py", line 19, in __init__
    initsoln, cfg, self._stepper_coeffs, self._dt
  File "/home/anthony/anaconda3/lib/python3.6/site-packages/pyfr/integrators/dual/pseudo/__init__.py", line 16, in get_pseudo_integrator
    initsoln, cfg, tcoeffs, dt)
  File "/home/anthony/anaconda3/lib/python3.6/site-packages/pyfr/integrators/dual/pseudo/multip.py", line 112, in __init__
    initsoln, cfg, tcoeffs, dt)
  File "/home/anthony/anaconda3/lib/python3.6/site-packages/pyfr/integrators/dual/pseudo/pseudocontrollers.py", line 98, in __init__
    super().__init__(*args, **kwargs)
  File "/home/anthony/anaconda3/lib/python3.6/site-packages/pyfr/integrators/dual/pseudo/pseudocontrollers.py", line 12, in __init__
    super().__init__(*args, **kwargs)
  File "/home/anthony/anaconda3/lib/python3.6/site-packages/pyfr/integrators/dual/pseudo/pseudosteppers.py", line 231, in __init__
    super().__init__(*args, **kwargs)
  File "/home/anthony/anaconda3/lib/python3.6/site-packages/pyfr/integrators/dual/pseudo/pseudosteppers.py", line 190, in __init__
    super().__init__(*args, **kwargs)
  File "/home/anthony/anaconda3/lib/python3.6/site-packages/pyfr/integrators/dual/pseudo/base.py", line 51, in __init__
    nregs=self.nregs, cfg=cfg)
  File "/home/anthony/anaconda3/lib/python3.6/site-packages/pyfr/solvers/base/system.py", line 60, in __init__
    bc_inters = self._load_bc_inters(rallocs, mesh, elemap)
  File "/home/anthony/anaconda3/lib/python3.6/site-packages/pyfr/solvers/base/system.py", line 157, in _load_bc_inters
    self.cfg)
  File "/home/anthony/anaconda3/lib/python3.6/site-packages/pyfr/solvers/acnavstokes/inters.py", line 98, in __init__
    self._tpl_c['v'] = self._eval_opts('uvw'[:self.ndims], default='0')
  File "/home/anthony/anaconda3/lib/python3.6/site-packages/pyfr/solvers/baseadvec/inters.py", line 96, in _eval_opts
    return [npeval(cfg.getexpr(sect, k, default), cc) for k in opts]
  File "/home/anthony/anaconda3/lib/python3.6/site-packages/pyfr/solvers/baseadvec/inters.py", line 96, in <listcomp>
    return [npeval(cfg.getexpr(sect, k, default), cc) for k in opts]
  File "/home/anthony/anaconda3/lib/python3.6/site-packages/pyfr/nputil.py", line 81, in npeval
    return eval(expr, _npeval_syms, locals)
  File "<string>", line 1, in <module>
TypeError: 'NoneType' object is not subscriptable

Am I doing something wrong ? Do the expressions depending on space and time work with all the types of wall boundary conditions or with only a few of them ? 

Kind regards,

Anthony Larroque

[Freddie Witherden]

Hi Anthony,

On 20/05/2019 11:44, Anthony Larroque wrote:
> Thank you very much for your answer.
>
> However, it does not seem to work with my setup.
>
> I am using the last version of PyFR that I found on:
> https://github.com/vincentlab/PyFR and run the example of the cylinder
> with simply adding in the inc_cylinder_2d.ini the velocities u = 0.1*x
> and v=0.1*x in the [soln-bcs-wall] to try an expression that depends on
> space.
>
> When I set a constant for the velocities like u = 0.1 and v = 0.1, I do
> not have a problem to run the simulation. However, with the expressions

So looking through the code it appears as if the no-slp-wall boundary
condition for the incompressible Navier-Stokes solver has not yet been
updated to include support for expressions.

This should not be too difficult to fix and I will see if I can get it
into the next release.

For the moment your best bet is to switch to the compressible solver
which does support this.

Regards, Freddie.

[Anthony Larroque]

Hi Freddie,

Thank you very much for your answer.

Indeed, it works well with the compressible solver.

Regards,

Anthony Larroque