SALOME platform Git repositories - tools/solverlab.git/blob

   1 #!/usr/bin/env python
   2 # -*-coding:utf-8 -*
   3
   4 import validationStationaryDiffusionEquation
   5 import cdmath as cm
   6 import matplotlib.pyplot as plt
   7 import numpy as np
   8 from math import log10, sqrt
   9 import time, json
  10
  11 convergence_synthesis=dict(validationStationaryDiffusionEquation.test_desc)
  12
  13 def convergence_StationaryDiffusion_2DFV_Neumann_RightTriangles():
  14     start = time.time()
  15     ### 2D FV right triangles mesh
  16     method = 'FV'
  17     BC = 'Neumann'
  18     meshList=[5,20,50,100, 200,400]
  19     nbMeshes=len(meshList)
  20     error_tab=[0]*nbMeshes
  21     mesh_size_tab=[0]*nbMeshes
  22     mesh_name='squareWithTriangles'
  23     meshType="Regular_RightTriangles"
  24     diag_data=[0]*nbMeshes
  25     time_tab=[0]*nbMeshes
  26     resolution=100
  27     curv_abs=np.linspace(0,sqrt(2),resolution+1)
  28     plt.close('all')
  29     i=0
  30     testColor="Orange \n (suspicious order 0 convergence), singular matrix"#Convergence of the linear solver if direct solver. Scheme seems to diverge (order -0.005)
  31     # Storing of numerical errors, mesh sizes and diagonal values
  32     for nx in meshList:
  33                 my_mesh=cm.Mesh(0,1,nx,0,1,nx,1)
  34                 error_tab[i], mesh_size_tab[i], diag_data[i], min_sol_num, max_sol_num, time_tab[i] =validationStationaryDiffusionEquation.SolveStationaryDiffusionEquation(my_mesh,resolution,meshType,method,BC)
  35
  36                 assert min_sol_num>-1.22
  37                 assert max_sol_num<1.
  38                 plt.plot(curv_abs, diag_data[i], label= str(mesh_size_tab[i]) + ' cells')
  39                 error_tab[i]=log10(error_tab[i])
  40                 time_tab[i]=log10(time_tab[i])
  41                 mesh_size_tab[i] = 0.5*log10(mesh_size_tab[i])
  42                 i=i+1
  43     end = time.time()
  44
  45
  46
  47     # Plot over diagonal line
  48     plt.legend()
  49     plt.xlabel('Position on diagonal line')
  50     plt.ylabel('Value on diagonal line')
  51     plt.title('Plot over diagonal line for finite volumes for the Poisson problem \n on 2D right triangles meshes with with Neumann BC')
  52     plt.savefig(mesh_name+"_2DFV_StatDiffusion_Neumann_PlotOverDiagonalLine.png")
  53
  54     # Least square linear regression
  55     # Find the best a,b such that f(x)=ax+b best approximates the convergence curve
  56     # The vector X=(a,b) solves a symmetric linear system AX=B with A=(a1,a2\\a2,a3), B=(b1,b2)
  57     a1=np.dot(mesh_size_tab,mesh_size_tab)
  58     a2=np.sum(mesh_size_tab)
  59     a3=nbMeshes
  60     b1=np.dot(error_tab,mesh_size_tab)
  61     b2=np.sum(error_tab)
  62
  63     det=a1*a3-a2*a2
  64     assert det!=0, 'convergence_StationaryDiffusion_2DFV_Neumann_RightTriangles() : Make sure you use distinct meshes and at least two meshes'
  65     a=( a3*b1-a2*b2)/det
  66     b=(-a2*b1+a1*b2)/det
  67
  68     print "FV for diffusion on 2D right triangle meshes: scheme order is ", -a
  69     assert abs(a-0.005)<0.01
  70
  71     # Plot of convergence curve
  72     plt.close()
  73     plt.plot(mesh_size_tab, error_tab, label='log(|numerical-exact|)')
  74     plt.plot(mesh_size_tab, a*np.array(mesh_size_tab)+b,label='least square slope : '+'%.3f' % a)
  75     plt.legend()
  76     plt.plot(mesh_size_tab, error_tab)
  77     plt.xlabel('log(sqrt(number of cells))')
  78     plt.ylabel('log(error)')
  79     plt.title('Convergence of finite volumes  for the Poisson problem \n on 2D right triangles meshes with with Neumann BC')
  80     plt.savefig(mesh_name+"_2DFV_StatDiffusion_Neumann_ConvergenceCurve.png")
  81
  82     # Plot of computational time
  83     plt.close()
  84     plt.plot(mesh_size_tab, time_tab, label='log(cpu time)')
  85     plt.legend()
  86     plt.xlabel('log(sqrt(number of cells))')
  87     plt.ylabel('log(cpu time)')
  88     plt.title('Computational time of finite volumes for the Poisson problem \n on 2D right triangles meshes with with Neumann BC')
  89     plt.savefig(mesh_name+"_2DFV_StatDiffusion_Neumann_ComputationalTime.png")
  90
  91     plt.close('all')
  92
  93     convergence_synthesis["Mesh_names"]=meshList
  94     convergence_synthesis["Mesh_type"]=meshType
  95     #convergence_synthesis["Mesh_path"]=mesh_path
  96     convergence_synthesis["Mesh_description"]=mesh_name
  97     convergence_synthesis["Mesh_sizes"]=[10**x for x in mesh_size_tab]
  98     convergence_synthesis["Space_dim"]=2
  99     convergence_synthesis["Mesh_dim"]=2
 100     convergence_synthesis["Mesh_cell_type"]="Triangles"
 101     convergence_synthesis["Errors"]=[10**x for x in error_tab]
 102     convergence_synthesis["Scheme_order"]=round(-a,4)
 103     convergence_synthesis["Test_color"]=testColor
 104     convergence_synthesis["PDE_model"]='Poisson'
 105     convergence_synthesis["Num_method"]=method
 106     convergence_synthesis["Bound_cond"]=BC
 107     convergence_synthesis["Comput_time"]=round(end-start,3)
 108
 109     with open('Convergence_Poisson_2DFV_'+mesh_name+'.json', 'w') as outfile:
 110         json.dump(convergence_synthesis, outfile)
 111
 112
 113 if __name__ == """__main__""":
 114     convergence_StationaryDiffusion_2DFV_Neumann_RightTriangles()