Mailman 3 September 2016 - SfePy

Use of shaper.py
by Jan Heczko Sept. 2, 2016

Sept. 2, 2016

Is there an example of problem definition that can be used to test shaper.py? I tried to grep for its requirements (equations_direct, equations_adjoint) but couldn't find anything. Thanks in advance Jan

2 3

Pressure outlet condition in Navier Stokes
by Nikhil Vaidya Sept. 2, 2016

Sept. 2, 2016

Hi, I am trying to solve a simple incompressible steady flow in a pipe. Attached are the setup file and mesh. I want to specify a inlet velocity and an outlet pressure. But I get the following error when I do so: sfepy: left over: ['verbose', '__builtins__', '__file__', 'absolute_import', '__name__', 'data_dir', '__package__', '_filename', '__doc__'] sfepy: reading mesh [line2, tri3, quad4, tetra4, hexa8] (C:\Users\310246696\OneDrive - Philips\Thesis_Work\sfepy_testcase\pipe_flow_test\test_p.mesh)... sfepy: ...done in 0.01 s sfepy: creating regions... sfepy: Omega sfepy: Outlet sfepy: Inlet sfepy: Walls sfepy: ...done in 0.03 s sfepy: equation "balance": sfepy: + dw_div_grad.i2.Omega( fluid.viscosity, v, u ) + dw_convect.i2.Omega( v, u ) - dw_stokes.i1.Omega( v, p ) = 0 sfepy: equation "incompressibility": sfepy: dw_stokes.i1.Omega( u, q ) = 0 sfepy: using solvers: ts: no ts nls: newton ls: ls sfepy: updating variables... sfepy: ...done Traceback (most recent call last): File "C:\sfepy\simple.py", line 170, in <module> main() File "C:\sfepy\simple.py", line 167, in main app() File "C:\sfepy\sfepy\applications\application.py", line 29, in call_basic return self.call(**kwargs) File "C:\sfepy\sfepy\applications\pde_solver_app.py", line 217, in call time_solver.init_time(nls_status=nls_status) File "C:\sfepy\sfepy\solvers\ts_solvers.py", line 52, in init_time self.problem.time_update() File "C:\sfepy\sfepy\discrete\problem.py", line 629, in time_update functions, create_matrix, is_matrix) File "C:\sfepy\sfepy\discrete\problem.py", line 587, in update_equations active_only=ac) File "C:\sfepy\sfepy\discrete\equations.py", line 287, in time_update active_only=active_only) File "C:\sfepy\sfepy\discrete\variables.py", line 413, in equation_mapping problem=problem) File "C:\sfepy\sfepy\discrete\variables.py", line 1458, in equation_mapping bcs.canonize_dof_names(self.dofs) File "C:\sfepy\sfepy\discrete\conditions.py", line 129, in canonize_dof_names cond.canonize_dof_names(dofs) File "C:\sfepy\sfepy\discrete\conditions.py", line 190, in canonize_dof_names self.dofs[0] = _canonize(self.dofs[0], dofs) File "C:\sfepy\sfepy\discrete\conditions.py", line 150, in _canonize vname, dd = dofs.split('.') ValueError: need more than 1 value to unpack Commercial simulation tools, e.g. Starccm+ allow the definition of velocity inlet and pressure outlet conditions simultaneously. I would like to do the same with SfePy. Is this possible? Best regards, Nikhil

2 4

The basename argument of homogen.py
by Jan Heczko Sept. 2, 2016

Sept. 2, 2016

Hi, it seems to me that the command-line parameter -o has no influence on the outcome of the script. E.g. $ ./homogen.py examples/homogenization/linear_homogenization.py $ ls output/ coefs_le.h5 coefs_le.txt corrs_le.h5 corrs_le.vtk linear_homogenization. py and calling the same with the -o option creates the same five files in the same location: $ ./homogen.py -o homog examples/homogenization/linear_homogenization.py $ ls output/ coefs_le.h5 coefs_le.txt corrs_le.h5 corrs_le.vtk linear_homogenization. py The value of the parameter is passed to HomogenizationApp in the same way as e.g. in simple.py (in which case it works as expected), but seems to be ignored afterwards. In my opinion, the option can be removed. Jan

2 3

Do I generate mesh in the wrong way?
by Ronghai Wu Sept. 1, 2016

Sept. 1, 2016

Hi Robert, I am using sfepy for static linear elasticity problem in 2D. The mesh size is 200*200 (mesh file is attached) and equations are as following: def D_function(self, ts, coors, mode, term=None, **kwargs): if mode != 'qp': return _, weights = term.integral.get_qp('2_4') self.n_qp = weights.shape[0] self.val_D = np.repeat(self.stiffness_tensor, self.n_qp, axis=0) return {'function' : self.val_D} def prestress_function(self,ts, coors, mode, term=None, **kwargs): if mode != 'qp': return _, weights = term.integral.get_qp('2_4') self.n_qp = weights.shape[0] self.val_pre = np.repeat(self.prestress_value, self.n_qp, axis=0) return {'function' : self.val_pre} self.m = Material('m', function=self.D_function) self.prestress_value = np.copy(self.stress_0) (self.prestress_value).shape = (-1, 3, 1) self.prestress = Material('prestress', function=self.prestress_function) integral = Integral('i', order=2) self.t1 = Term.new('dw_lin_elastic(m.function, v, u )', integral, self.omega, m=self.m, v=self.v, u=self.u) self.t2 = Term.new('dw_lin_prestress(prestress.function, v )',integral, self.omega, prestress=self.prestress, v=self.v) self.t3 = Term.new('dw_surface_ltr(traction_lef.val, v)', integral, self.lef, traction_lef=self.traction_lef, v=self.v) self.t4 = Term.new('dw_surface_ltr(traction_rig.val, v)', integral, self.rig, traction_rig=self.traction_rig, v=self.v) self.t5 = Term.new('dw_surface_ltr(traction_top.val, v)', integral, self.top, traction_top=self.traction_top, v=self.v) self.t6 = Term.new('dw_surface_ltr(traction_bot.val, v)', integral, self.bot, traction_bot=self.traction_bot, v=self.v) self.fem_eq = Equation('balance', self.t1 - self.t2 - self.t3 - self.t4 - self.t5 - self.t6) self.eqs = Equations([self.fem_eq]) self.ls = ScipyDirect({}) self.nls_status = IndexedStruct() self.nls = Newton({'i_max' : 1,'eps_a' : 1e-8,'problem' : 'nonlinear'}, lin_solver=self.ls, status=self.nls_status) self.pb = Problem('elasticity', equations=self.eqs, nls=self.nls, ls=self.ls) self.pb.time_update(ebcs=Conditions([fix_u_lef_corn])) self.pb.update_materials() self.vec = self.pb.solve() When I run the simulation, the information below are printed. I have several questions about it. I tried to find answers from source codes but failed. Could you please give me some help, thanks. (1) I do understand why "matrix shape: (79999, 79999)" and "matrix structural nonzeros: 1439965". what exactly this "matrix" is? what is its relation with mesh size (200*200)? (2) I do not understand why "updating materials..." did twice. (3) "solve: 6.18 [s]". I guess it should not take such a long time for a mesh size (200*200) problem. Is it because I generate mesh file in the wrong way or because of other reasons? (4) In which source code file the following information are printed? sfepy: updating variables... sfepy: ...done sfepy: setting up dof connectivities... sfepy: ...done in 0.00 s sfepy: matrix shape: (79999, 79999) sfepy: assembling matrix graph... sfepy: ...done in 0.05 s sfepy: matrix structural nonzeros: 1439965 (2.25e-04% fill) sfepy: updating materials... sfepy: m sfepy: traction_bot sfepy: traction_rig sfepy: prestress sfepy: traction_top sfepy: traction_lef sfepy: ...done in 0.05 s sfepy: updating materials... sfepy: m sfepy: traction_bot sfepy: traction_rig sfepy: prestress sfepy: traction_top sfepy: traction_lef sfepy: ...done in 0.05 s sfepy: nls: iter: 0, residual: 4.741758e+01 (rel: 1.000000e+00) sfepy: rezidual: 0.04 [s] sfepy: solve: 6.18 [s] sfepy: matrix: 0.10 [s] sfepy: nls: iter: 1, residual: 5.936719e-14 (rel: 1.252008e-15) sfepy: equation "tmp": sfepy: ev_cauchy_strain.2.Omega(u) sfepy: updating materials... sfepy: ...done in 0.00 s Regards Ronghai

2 10