| 1 |
|
| 2 |
######################################################## |
| 3 |
# |
| 4 |
# Copyright (c) 2003-2008 by University of Queensland |
| 5 |
# Earth Systems Science Computational Center (ESSCC) |
| 6 |
# http://www.uq.edu.au/esscc |
| 7 |
# |
| 8 |
# Primary Business: Queensland, Australia |
| 9 |
# Licensed under the Open Software License version 3.0 |
| 10 |
# http://www.opensource.org/licenses/osl-3.0.php |
| 11 |
# |
| 12 |
######################################################## |
| 13 |
|
| 14 |
__copyright__="""Copyright (c) 2003-2008 by University of Queensland |
| 15 |
Earth Systems Science Computational Center (ESSCC) |
| 16 |
http://www.uq.edu.au/esscc |
| 17 |
Primary Business: Queensland, Australia""" |
| 18 |
__license__="""Licensed under the Open Software License version 3.0 |
| 19 |
http://www.opensource.org/licenses/osl-3.0.php""" |
| 20 |
__url__="http://www.uq.edu.au/esscc/escript-finley" |
| 21 |
|
| 22 |
import unittest |
| 23 |
import tempfile |
| 24 |
|
| 25 |
from esys.escript import * |
| 26 |
from esys.finley import Rectangle |
| 27 |
from esys.escript.models import DarcyFlow |
| 28 |
import sys |
| 29 |
import os |
| 30 |
try: |
| 31 |
FINLEY_WORKDIR=os.environ['FINLEY_WORKDIR'] |
| 32 |
except KeyError: |
| 33 |
FINLEY_WORKDIR='.' |
| 34 |
|
| 35 |
|
| 36 |
VERBOSE=False # or True |
| 37 |
DETAIL_VERBOSE=False |
| 38 |
|
| 39 |
from esys.escript import * |
| 40 |
from esys.escript.models import StokesProblemCartesian, PowerLaw |
| 41 |
from esys.finley import Rectangle, Brick |
| 42 |
|
| 43 |
from esys.escript.models import Mountains |
| 44 |
from math import pi |
| 45 |
|
| 46 |
#==================================================================================================================== |
| 47 |
class Test_StokesProblemCartesian2D(unittest.TestCase): |
| 48 |
def setUp(self): |
| 49 |
NE=6 |
| 50 |
self.TOL=1e-3 |
| 51 |
self.domain=Rectangle(NE,NE,order=2,useFullElementOrder=True) |
| 52 |
def tearDown(self): |
| 53 |
del self.domain |
| 54 |
def test_PCG_P_0(self): |
| 55 |
ETA=1. |
| 56 |
P1=0. |
| 57 |
|
| 58 |
x=self.domain.getX() |
| 59 |
F=-P1*x[1]*[1.,0]+(2*ETA-P1*x[0])*[0.,1.] |
| 60 |
mask=whereZero(x[0]) * [1.,1.] \ |
| 61 |
+whereZero(x[0]-1) * [1.,1.] \ |
| 62 |
+whereZero(x[1]) * [1.,0.] \ |
| 63 |
+whereZero(x[1]-1) * [1.,1.] |
| 64 |
|
| 65 |
sp=StokesProblemCartesian(self.domain) |
| 66 |
|
| 67 |
sp.initialize(f=F,fixed_u_mask=mask,eta=ETA) |
| 68 |
u0=(1-x[0])*x[0]*[0.,1.] |
| 69 |
p0=Scalar(P1,ReducedSolution(self.domain)) |
| 70 |
sp.setTolerance(self.TOL) |
| 71 |
u,p=sp.solve(u0,p0,show_details=DETAIL_VERBOSE, verbose=VERBOSE,max_iter=100,usePCG=True) |
| 72 |
|
| 73 |
error_v0=Lsup(u[0]-u0[0]) |
| 74 |
error_v1=Lsup(u[1]-u0[1])/0.25 |
| 75 |
error_p=Lsup(p+P1*x[0]*x[1]) |
| 76 |
self.failUnless(error_p<10*self.TOL, "pressure error too large.") |
| 77 |
self.failUnless(error_v0<10*self.TOL, "0-velocity error too large.") |
| 78 |
self.failUnless(error_v1<10*self.TOL, "1-velocity error too large.") |
| 79 |
|
| 80 |
def test_PCG_P_small(self): |
| 81 |
ETA=1. |
| 82 |
P1=1. |
| 83 |
|
| 84 |
x=self.domain.getX() |
| 85 |
F=-P1*x[1]*[1.,0]+(2*ETA-P1*x[0])*[0.,1.] |
| 86 |
mask=whereZero(x[0]) * [1.,1.] \ |
| 87 |
+whereZero(x[0]-1) * [1.,1.] \ |
| 88 |
+whereZero(x[1]) * [1.,0.] \ |
| 89 |
+whereZero(x[1]-1) * [1.,1.] |
| 90 |
|
| 91 |
sp=StokesProblemCartesian(self.domain) |
| 92 |
|
| 93 |
sp.initialize(f=F,fixed_u_mask=mask,eta=ETA) |
| 94 |
u0=(1-x[0])*x[0]*[0.,1.] |
| 95 |
p0=Scalar(P1,ReducedSolution(self.domain)) |
| 96 |
sp.setTolerance(self.TOL*0.2) |
| 97 |
u,p=sp.solve(u0,p0,show_details=DETAIL_VERBOSE, verbose=VERBOSE,max_iter=100,usePCG=True) |
| 98 |
error_v0=Lsup(u[0]-u0[0]) |
| 99 |
error_v1=Lsup(u[1]-u0[1])/0.25 |
| 100 |
error_p=Lsup(P1*x[0]*x[1]+p) |
| 101 |
saveVTK("d.vtu",p=p, e=P1*x[0]*x[1]+p, p_ref=P1*x[0]*x[1]) |
| 102 |
self.failUnless(error_v0<10*self.TOL, "0-velocity error too large.") |
| 103 |
self.failUnless(error_v1<10*self.TOL, "1-velocity error too large.") |
| 104 |
self.failUnless(error_p<10*self.TOL, "pressure error too large.") |
| 105 |
|
| 106 |
def test_PCG_P_large(self): |
| 107 |
ETA=1. |
| 108 |
P1=1000. |
| 109 |
|
| 110 |
x=self.domain.getX() |
| 111 |
F=-P1*x[1]*[1.,0]+(2*ETA-P1*x[0])*[0.,1.] |
| 112 |
mask=whereZero(x[0]) * [1.,1.] \ |
| 113 |
+whereZero(x[0]-1) * [1.,1.] \ |
| 114 |
+whereZero(x[1]) * [1.,0.] \ |
| 115 |
+whereZero(x[1]-1) * [1.,1.] |
| 116 |
|
| 117 |
sp=StokesProblemCartesian(self.domain) |
| 118 |
|
| 119 |
sp.initialize(f=F,fixed_u_mask=mask,eta=ETA) |
| 120 |
u0=(1-x[0])*x[0]*[0.,1.] |
| 121 |
p0=Scalar(P1,ReducedSolution(self.domain)) |
| 122 |
sp.setTolerance(self.TOL) |
| 123 |
u,p=sp.solve(u0,p0,show_details=DETAIL_VERBOSE, verbose=VERBOSE,max_iter=100,usePCG=True) |
| 124 |
|
| 125 |
error_v0=Lsup(u[0]-u0[0]) |
| 126 |
error_v1=Lsup(u[1]-u0[1])/0.25 |
| 127 |
error_p=Lsup(P1*x[0]*x[1]+p)/P1 |
| 128 |
self.failUnless(error_p<10*self.TOL, "pressure error too large.") |
| 129 |
self.failUnless(error_v0<10*self.TOL, "0-velocity error too large.") |
| 130 |
self.failUnless(error_v1<10*self.TOL, "1-velocity error too large.") |
| 131 |
|
| 132 |
def test_GMRES_P_0(self): |
| 133 |
ETA=1. |
| 134 |
P1=0. |
| 135 |
|
| 136 |
x=self.domain.getX() |
| 137 |
F=-P1*x[1]*[1.,0]+(2*ETA-P1*x[0])*[0.,1.] |
| 138 |
mask=whereZero(x[0]) * [1.,1.] \ |
| 139 |
+whereZero(x[0]-1) * [1.,1.] \ |
| 140 |
+whereZero(x[1]) * [1.,0.] \ |
| 141 |
+whereZero(x[1]-1) * [1.,1.] |
| 142 |
|
| 143 |
sp=StokesProblemCartesian(self.domain) |
| 144 |
|
| 145 |
sp.initialize(f=F,fixed_u_mask=mask,eta=ETA) |
| 146 |
u0=(1-x[0])*x[0]*[0.,1.] |
| 147 |
p0=Scalar(P1,ReducedSolution(self.domain)) |
| 148 |
sp.setTolerance(self.TOL) |
| 149 |
u,p=sp.solve(u0,p0,show_details=DETAIL_VERBOSE, verbose=VERBOSE,max_iter=50,usePCG=False,iter_restart=18) |
| 150 |
|
| 151 |
error_v0=Lsup(u[0]-u0[0]) |
| 152 |
error_v1=Lsup(u[1]-u0[1])/0.25 |
| 153 |
error_p=Lsup(P1*x[0]*x[1]+p) |
| 154 |
self.failUnless(error_p<10*self.TOL, "pressure error too large.") |
| 155 |
self.failUnless(error_v0<10*self.TOL, "0-velocity error too large.") |
| 156 |
self.failUnless(error_v1<10*self.TOL, "1-velocity error too large.") |
| 157 |
|
| 158 |
def test_GMRES_P_small(self): |
| 159 |
ETA=1. |
| 160 |
P1=1. |
| 161 |
|
| 162 |
x=self.domain.getX() |
| 163 |
F=-P1*x[1]*[1.,0]+(2*ETA-P1*x[0])*[0.,1.] |
| 164 |
mask=whereZero(x[0]) * [1.,1.] \ |
| 165 |
+whereZero(x[0]-1) * [1.,1.] \ |
| 166 |
+whereZero(x[1]) * [1.,0.] \ |
| 167 |
+whereZero(x[1]-1) * [1.,1.] |
| 168 |
|
| 169 |
sp=StokesProblemCartesian(self.domain) |
| 170 |
|
| 171 |
sp.initialize(f=F,fixed_u_mask=mask,eta=ETA) |
| 172 |
u0=(1-x[0])*x[0]*[0.,1.] |
| 173 |
p0=Scalar(P1,ReducedSolution(self.domain)) |
| 174 |
sp.setTolerance(self.TOL*0.1) |
| 175 |
u,p=sp.solve(u0,p0,show_details=DETAIL_VERBOSE, verbose=VERBOSE,max_iter=20,usePCG=False) |
| 176 |
|
| 177 |
error_v0=Lsup(u[0]-u0[0]) |
| 178 |
error_v1=Lsup(u[1]-u0[1])/0.25 |
| 179 |
error_p=Lsup(P1*x[0]*x[1]+p) |
| 180 |
self.failUnless(error_p<10*self.TOL, "pressure error too large.") |
| 181 |
self.failUnless(error_v0<10*self.TOL, "0-velocity error too large.") |
| 182 |
self.failUnless(error_v1<10*self.TOL, "1-velocity error too large.") |
| 183 |
|
| 184 |
def test_GMRES_P_large(self): |
| 185 |
ETA=1. |
| 186 |
P1=1000. |
| 187 |
|
| 188 |
x=self.domain.getX() |
| 189 |
F=-P1*x[1]*[1.,0]+(2*ETA-P1*x[0])*[0.,1.] |
| 190 |
mask=whereZero(x[0]) * [1.,1.] \ |
| 191 |
+whereZero(x[0]-1) * [1.,1.] \ |
| 192 |
+whereZero(x[1]) * [1.,0.] \ |
| 193 |
+whereZero(x[1]-1) * [1.,1.] |
| 194 |
|
| 195 |
sp=StokesProblemCartesian(self.domain) |
| 196 |
|
| 197 |
sp.initialize(f=F,fixed_u_mask=mask,eta=ETA) |
| 198 |
u0=(1-x[0])*x[0]*[0.,1.] |
| 199 |
p0=Scalar(P1,ReducedSolution(self.domain)) |
| 200 |
sp.setTolerance(self.TOL) |
| 201 |
u,p=sp.solve(u0,p0,show_details=DETAIL_VERBOSE, verbose=VERBOSE,max_iter=100,usePCG=False) |
| 202 |
|
| 203 |
error_v0=Lsup(u[0]-u0[0]) |
| 204 |
error_v1=Lsup(u[1]-u0[1])/0.25 |
| 205 |
error_p=Lsup(P1*x[0]*x[1]+p)/P1 |
| 206 |
self.failUnless(error_p<10*self.TOL, "pressure error too large.") |
| 207 |
self.failUnless(error_v0<10*self.TOL, "0-velocity error too large.") |
| 208 |
self.failUnless(error_v1<10*self.TOL, "1-velocity error too large.") |
| 209 |
#==================================================================================================================== |
| 210 |
class Test_StokesProblemCartesian3D(unittest.TestCase): |
| 211 |
def setUp(self): |
| 212 |
NE=6 |
| 213 |
self.TOL=1e-4 |
| 214 |
self.domain=Brick(NE,NE,NE,order=2,useFullElementOrder=True) |
| 215 |
def tearDown(self): |
| 216 |
del self.domain |
| 217 |
def test_PCG_P_0(self): |
| 218 |
ETA=1. |
| 219 |
P1=0. |
| 220 |
|
| 221 |
x=self.domain.getX() |
| 222 |
F=-P1*x[1]*x[2]*[1.,0.,0.]-P1*x[0]*x[2]*[0.,1.,0.]+(2*ETA*((1-x[0])*x[0]+(1-x[1])*x[1])-P1*x[0]*x[1])*[0.,0.,1.] |
| 223 |
x=self.domain.getX() |
| 224 |
mask=whereZero(x[0]) * [1.,1.,1.] \ |
| 225 |
+whereZero(x[0]-1) * [1.,1.,1.] \ |
| 226 |
+whereZero(x[1]) * [1.,0.,1.] \ |
| 227 |
+whereZero(x[1]-1) * [1.,1.,1.] \ |
| 228 |
+whereZero(x[2]) * [1.,1.,0.] \ |
| 229 |
+whereZero(x[2]-1) * [1.,1.,1.] |
| 230 |
|
| 231 |
|
| 232 |
sp=StokesProblemCartesian(self.domain) |
| 233 |
|
| 234 |
sp.initialize(f=F,fixed_u_mask=mask,eta=ETA) |
| 235 |
u0=(1-x[0])*x[0]*(1-x[1])*x[1]*[0.,0.,1.] |
| 236 |
p0=Scalar(P1,ReducedSolution(self.domain)) |
| 237 |
sp.setTolerance(self.TOL) |
| 238 |
u,p=sp.solve(u0,p0,show_details=DETAIL_VERBOSE, verbose=VERBOSE ,max_iter=100,usePCG=True) |
| 239 |
|
| 240 |
error_v0=Lsup(u[0]-u0[0]) |
| 241 |
error_v1=Lsup(u[1]-u0[1]) |
| 242 |
error_v2=Lsup(u[2]-u0[2])/0.25**2 |
| 243 |
error_p=Lsup(P1*x[0]*x[1]*x[2]+p) |
| 244 |
self.failUnless(error_p<10*self.TOL, "pressure error too large.") |
| 245 |
self.failUnless(error_v0<10*self.TOL, "0-velocity error too large.") |
| 246 |
self.failUnless(error_v1<10*self.TOL, "1-velocity error too large.") |
| 247 |
self.failUnless(error_v2<10*self.TOL, "2-velocity error too large.") |
| 248 |
|
| 249 |
def test_PCG_P_small(self): |
| 250 |
ETA=1. |
| 251 |
P1=1. |
| 252 |
|
| 253 |
x=self.domain.getX() |
| 254 |
F=-P1*x[1]*x[2]*[1.,0.,0.]-P1*x[0]*x[2]*[0.,1.,0.]+(2*ETA*((1-x[0])*x[0]+(1-x[1])*x[1])-P1*x[0]*x[1])*[0.,0.,1.] |
| 255 |
mask=whereZero(x[0]) * [1.,1.,1.] \ |
| 256 |
+whereZero(x[0]-1) * [1.,1.,1.] \ |
| 257 |
+whereZero(x[1]) * [1.,0.,1.] \ |
| 258 |
+whereZero(x[1]-1) * [1.,1.,1.] \ |
| 259 |
+whereZero(x[2]) * [1.,1.,0.] \ |
| 260 |
+whereZero(x[2]-1) * [1.,1.,1.] |
| 261 |
|
| 262 |
|
| 263 |
sp=StokesProblemCartesian(self.domain) |
| 264 |
|
| 265 |
sp.initialize(f=F,fixed_u_mask=mask,eta=ETA) |
| 266 |
u0=(1-x[0])*x[0]*(1-x[1])*x[1]*[0.,0.,1.] |
| 267 |
p0=Scalar(P1,ReducedSolution(self.domain)) |
| 268 |
sp.setTolerance(self.TOL*0.1) |
| 269 |
u,p=sp.solve(u0,p0,show_details=DETAIL_VERBOSE, verbose=VERBOSE ,max_iter=100,usePCG=True) |
| 270 |
error_v0=Lsup(u[0]-u0[0]) |
| 271 |
error_v1=Lsup(u[1]-u0[1]) |
| 272 |
error_v2=Lsup(u[2]-u0[2])/0.25**2 |
| 273 |
error_p=Lsup(P1*x[0]*x[1]*x[2]+p) |
| 274 |
self.failUnless(error_p<10*self.TOL, "pressure error too large.") |
| 275 |
self.failUnless(error_v0<10*self.TOL, "0-velocity error too large.") |
| 276 |
self.failUnless(error_v1<10*self.TOL, "1-velocity error too large.") |
| 277 |
self.failUnless(error_v2<10*self.TOL, "2-velocity error too large.") |
| 278 |
def test_PCG_P_large(self): |
| 279 |
ETA=1. |
| 280 |
P1=1000. |
| 281 |
|
| 282 |
x=self.domain.getX() |
| 283 |
F=-P1*x[1]*x[2]*[1.,0.,0.]-P1*x[0]*x[2]*[0.,1.,0.]+(2*ETA*((1-x[0])*x[0]+(1-x[1])*x[1])-P1*x[0]*x[1])*[0.,0.,1.] |
| 284 |
mask=whereZero(x[0]) * [1.,1.,1.] \ |
| 285 |
+whereZero(x[0]-1) * [1.,1.,1.] \ |
| 286 |
+whereZero(x[1]) * [1.,0.,1.] \ |
| 287 |
+whereZero(x[1]-1) * [1.,1.,1.] \ |
| 288 |
+whereZero(x[2]) * [1.,1.,0.] \ |
| 289 |
+whereZero(x[2]-1) * [1.,1.,1.] |
| 290 |
|
| 291 |
|
| 292 |
sp=StokesProblemCartesian(self.domain) |
| 293 |
|
| 294 |
sp.initialize(f=F,fixed_u_mask=mask,eta=ETA) |
| 295 |
u0=(1-x[0])*x[0]*(1-x[1])*x[1]*[0.,0.,1.] |
| 296 |
p0=Scalar(P1,ReducedSolution(self.domain)) |
| 297 |
sp.setTolerance(self.TOL) |
| 298 |
u,p=sp.solve(u0,-p0,show_details=DETAIL_VERBOSE, verbose=VERBOSE ,max_iter=100,usePCG=True) |
| 299 |
|
| 300 |
error_v0=Lsup(u[0]-u0[0]) |
| 301 |
error_v1=Lsup(u[1]-u0[1]) |
| 302 |
error_v2=Lsup(u[2]-u0[2])/0.25**2 |
| 303 |
error_p=Lsup(P1*x[0]*x[1]*x[2]+p)/P1 |
| 304 |
self.failUnless(error_p<10*self.TOL, "pressure error too large.") |
| 305 |
self.failUnless(error_v0<10*self.TOL, "0-velocity error too large.") |
| 306 |
self.failUnless(error_v1<10*self.TOL, "1-velocity error too large.") |
| 307 |
self.failUnless(error_v2<10*self.TOL, "2-velocity error too large.") |
| 308 |
|
| 309 |
def test_GMRES_P_0(self): |
| 310 |
ETA=1. |
| 311 |
P1=0. |
| 312 |
|
| 313 |
x=self.domain.getX() |
| 314 |
F=-P1*x[1]*x[2]*[1.,0.,0.]-P1*x[0]*x[2]*[0.,1.,0.]+(2*ETA*((1-x[0])*x[0]+(1-x[1])*x[1])-P1*x[0]*x[1])*[0.,0.,1.] |
| 315 |
x=self.domain.getX() |
| 316 |
mask=whereZero(x[0]) * [1.,1.,1.] \ |
| 317 |
+whereZero(x[0]-1) * [1.,1.,1.] \ |
| 318 |
+whereZero(x[1]) * [1.,1.,1.] \ |
| 319 |
+whereZero(x[1]-1) * [1.,1.,1.] \ |
| 320 |
+whereZero(x[2]) * [1.,1.,0.] \ |
| 321 |
+whereZero(x[2]-1) * [1.,1.,1.] |
| 322 |
|
| 323 |
|
| 324 |
sp=StokesProblemCartesian(self.domain) |
| 325 |
|
| 326 |
sp.initialize(f=F,fixed_u_mask=mask,eta=ETA) |
| 327 |
u0=(1-x[0])*x[0]*(1-x[1])*x[1]*[0.,0.,1.] |
| 328 |
p0=Scalar(P1,ReducedSolution(self.domain)) |
| 329 |
sp.setTolerance(self.TOL) |
| 330 |
u,p=sp.solve(u0,p0,show_details=DETAIL_VERBOSE, verbose=VERBOSE,max_iter=100,usePCG=False,iter_restart=20) |
| 331 |
|
| 332 |
error_v0=Lsup(u[0]-u0[0]) |
| 333 |
error_v1=Lsup(u[1]-u0[1]) |
| 334 |
error_v2=Lsup(u[2]-u0[2])/0.25**2 |
| 335 |
error_p=Lsup(P1*x[0]*x[1]*x[2]+p) |
| 336 |
self.failUnless(error_p<10*self.TOL, "pressure error too large.") |
| 337 |
self.failUnless(error_v0<10*self.TOL, "0-velocity error too large.") |
| 338 |
self.failUnless(error_v1<10*self.TOL, "1-velocity error too large.") |
| 339 |
self.failUnless(error_v2<10*self.TOL, "2-velocity error too large.") |
| 340 |
def test_GMRES_P_small(self): |
| 341 |
ETA=1. |
| 342 |
P1=1. |
| 343 |
|
| 344 |
x=self.domain.getX() |
| 345 |
F=-P1*x[1]*x[2]*[1.,0.,0.]-P1*x[0]*x[2]*[0.,1.,0.]+(2*ETA*((1-x[0])*x[0]+(1-x[1])*x[1])-P1*x[0]*x[1])*[0.,0.,1.] |
| 346 |
mask=whereZero(x[0]) * [1.,1.,1.] \ |
| 347 |
+whereZero(x[0]-1) * [1.,1.,1.] \ |
| 348 |
+whereZero(x[1]) * [1.,1.,1.] \ |
| 349 |
+whereZero(x[1]-1) * [1.,1.,1.] \ |
| 350 |
+whereZero(x[2]) * [1.,1.,0.] \ |
| 351 |
+whereZero(x[2]-1) * [1.,1.,1.] |
| 352 |
|
| 353 |
|
| 354 |
sp=StokesProblemCartesian(self.domain) |
| 355 |
|
| 356 |
sp.initialize(f=F,fixed_u_mask=mask,eta=ETA) |
| 357 |
u0=(1-x[0])*x[0]*(1-x[1])*x[1]*[0.,0.,1.] |
| 358 |
p0=Scalar(P1,ReducedSolution(self.domain)) |
| 359 |
sp.setTolerance(self.TOL*0.1) |
| 360 |
u,p=sp.solve(u0,p0,show_details=DETAIL_VERBOSE and False, verbose=VERBOSE,max_iter=100,usePCG=False) |
| 361 |
|
| 362 |
error_v0=Lsup(u[0]-u0[0]) |
| 363 |
error_v1=Lsup(u[1]-u0[1]) |
| 364 |
error_v2=Lsup(u[2]-u0[2])/0.25**2 |
| 365 |
error_p=Lsup(P1*x[0]*x[1]*x[2]+p)/P1 |
| 366 |
self.failUnless(error_v0<10*self.TOL, "0-velocity error too large.") |
| 367 |
self.failUnless(error_v1<10*self.TOL, "1-velocity error too large.") |
| 368 |
self.failUnless(error_v2<10*self.TOL, "2-velocity error too large.") |
| 369 |
self.failUnless(error_p<10*self.TOL, "pressure error too large.") |
| 370 |
def test_GMRES_P_large(self): |
| 371 |
ETA=1. |
| 372 |
P1=1000. |
| 373 |
|
| 374 |
x=self.domain.getX() |
| 375 |
F=-P1*x[1]*x[2]*[1.,0.,0.]-P1*x[0]*x[2]*[0.,1.,0.]+(2*ETA*((1-x[0])*x[0]+(1-x[1])*x[1])-P1*x[0]*x[1])*[0.,0.,1.] |
| 376 |
mask=whereZero(x[0]) * [1.,1.,1.] \ |
| 377 |
+whereZero(x[0]-1) * [1.,1.,1.] \ |
| 378 |
+whereZero(x[1]) * [1.,0.,1.] \ |
| 379 |
+whereZero(x[1]-1) * [1.,1.,1.] \ |
| 380 |
+whereZero(x[2]) * [1.,1.,0.] \ |
| 381 |
+whereZero(x[2]-1) * [1.,1.,1.] |
| 382 |
|
| 383 |
|
| 384 |
sp=StokesProblemCartesian(self.domain) |
| 385 |
|
| 386 |
sp.initialize(f=F,fixed_u_mask=mask,eta=ETA) |
| 387 |
u0=(1-x[0])*x[0]*(1-x[1])*x[1]*[0.,0.,1.] |
| 388 |
p0=Scalar(P1,ReducedSolution(self.domain)) |
| 389 |
sp.setTolerance(self.TOL) |
| 390 |
u,p=sp.solve(u0,p0,show_details=DETAIL_VERBOSE, verbose=VERBOSE ,max_iter=100,usePCG=False) |
| 391 |
|
| 392 |
error_v0=Lsup(u[0]-u0[0]) |
| 393 |
error_v1=Lsup(u[1]-u0[1]) |
| 394 |
error_v2=Lsup(u[2]-u0[2])/0.25**2 |
| 395 |
error_p=Lsup(P1*x[0]*x[1]*x[2]+p)/P1 |
| 396 |
self.failUnless(error_p<10*self.TOL, "pressure error too large.") |
| 397 |
self.failUnless(error_v0<10*self.TOL, "0-velocity error too large.") |
| 398 |
self.failUnless(error_v1<10*self.TOL, "1-velocity error too large.") |
| 399 |
self.failUnless(error_v2<10*self.TOL, "2-velocity error too large.") |
| 400 |
#==================================================================================================================== |
| 401 |
class Test_Darcy(unittest.TestCase): |
| 402 |
# this is a simple test for the darcy flux problem |
| 403 |
# |
| 404 |
# |
| 405 |
# p = 1/k * ( 1/2* (fb-f0)/xb* x **2 + f0 * x - ub*x ) + p0 |
| 406 |
# |
| 407 |
# with f = (fb-f0)/xb* x + f0 |
| 408 |
# |
| 409 |
# u = f - k * p,x = ub |
| 410 |
# |
| 411 |
# we prescribe pressure at x=x0=0 to p0 |
| 412 |
# |
| 413 |
# if we prescribe the pressure on the bottom x=xb we set |
| 414 |
# |
| 415 |
# pb= 1/k * ( 1/2* (fb-f0)* xb + f0 * xb - ub*xb ) + p0 = 1/k * ((fb+f0)/2 - ub ) * xb + p0 |
| 416 |
# |
| 417 |
# which leads to ub = (fb+f0)/2-k*(pb-p0)/xb |
| 418 |
# |
| 419 |
def rescaleDomain(self): |
| 420 |
x=self.dom.getX().copy() |
| 421 |
for i in xrange(self.dom.getDim()): |
| 422 |
x_inf=inf(x[i]) |
| 423 |
x_sup=sup(x[i]) |
| 424 |
if i == self.dom.getDim()-1: |
| 425 |
x[i]=-self.WIDTH*(x[i]-x_sup)/(x_inf-x_sup) |
| 426 |
else: |
| 427 |
x[i]=self.WIDTH*(x[i]-x_inf)/(x_sup-x_inf) |
| 428 |
self.dom.setX(x) |
| 429 |
def getScalarMask(self,include_bottom=True): |
| 430 |
x=self.dom.getX().copy() |
| 431 |
x_inf=inf(x[self.dom.getDim()-1]) |
| 432 |
x_sup=sup(x[self.dom.getDim()-1]) |
| 433 |
out=whereZero(x[self.dom.getDim()-1]-x_sup) |
| 434 |
if include_bottom: out+=whereZero(x[self.dom.getDim()-1]-x_inf) |
| 435 |
return wherePositive(out) |
| 436 |
def getVectorMask(self,include_bottom=True): |
| 437 |
x=self.dom.getX().copy() |
| 438 |
out=Vector(0.,Solution(self.dom)) |
| 439 |
for i in xrange(self.dom.getDim()): |
| 440 |
x_inf=inf(x[i]) |
| 441 |
x_sup=sup(x[i]) |
| 442 |
if i != self.dom.getDim()-1: out[i]+=whereZero(x[i]-x_sup) |
| 443 |
if i != self.dom.getDim()-1 or include_bottom: out[i]+=whereZero(x[i]-x_inf) |
| 444 |
return wherePositive(out) |
| 445 |
|
| 446 |
def setSolutionFixedBottom(self, p0, pb, f0, fb, k): |
| 447 |
d=self.dom.getDim() |
| 448 |
x=self.dom.getX()[d-1] |
| 449 |
xb=inf(x) |
| 450 |
u=Vector(0.,Solution(self.dom))+kronecker(d)[d-1]*((f0+fb)/2.-k*(pb-p0)/xb) |
| 451 |
p=1./k*((fb-f0)/(xb*2.)* x**2 - (fb-f0)/2.*x)+(pb-p0)/xb*x + p0 |
| 452 |
f= ((fb-f0)/xb* x + f0)*kronecker(Function(self.dom))[d-1] |
| 453 |
return u,p,f |
| 454 |
|
| 455 |
def testConstF_FixedBottom_smallK(self): |
| 456 |
k=1.e-10 |
| 457 |
mp=self.getScalarMask(include_bottom=True) |
| 458 |
mv=self.getVectorMask(include_bottom=False) |
| 459 |
u_ref,p_ref,f=self.setSolutionFixedBottom(p0=2500,pb=4000.,f0=10.,fb=10.,k=k) |
| 460 |
p=p_ref*mp |
| 461 |
u=u_ref*mv |
| 462 |
df=DarcyFlow(self.dom) |
| 463 |
df.setValue(g=f, |
| 464 |
location_of_fixed_pressure=mp, |
| 465 |
location_of_fixed_flux=mv, |
| 466 |
permeability=Scalar(k,Function(self.dom))) |
| 467 |
df.setTolerance(rtol=self.TOL) |
| 468 |
df.setSubProblemTolerance() |
| 469 |
v,p=df.solve(u_ref,p, max_iter=100, verbose=VERBOSE) |
| 470 |
self.failUnless(Lsup(v-u_ref)<self.TOL*10.*Lsup(u_ref), "flux error too big.") |
| 471 |
self.failUnless(Lsup(p-p_ref)<self.TOL*10.*Lsup(p_ref), "pressure error too big.") |
| 472 |
def testConstF_FixedBottom_mediumK(self): |
| 473 |
k=1. |
| 474 |
mp=self.getScalarMask(include_bottom=True) |
| 475 |
mv=self.getVectorMask(include_bottom=False) |
| 476 |
u_ref,p_ref,f=self.setSolutionFixedBottom(p0=2500,pb=4000.,f0=10.,fb=10.,k=k) |
| 477 |
p=p_ref*mp |
| 478 |
u=u_ref*mv |
| 479 |
df=DarcyFlow(self.dom) |
| 480 |
df.setValue(g=f, |
| 481 |
location_of_fixed_pressure=mp, |
| 482 |
location_of_fixed_flux=mv, |
| 483 |
permeability=Scalar(k,Function(self.dom))) |
| 484 |
df.setTolerance(rtol=self.TOL) |
| 485 |
df.setSubProblemTolerance() |
| 486 |
v,p=df.solve(u,p,max_iter=100, verbose=VERBOSE ) |
| 487 |
self.failUnless(Lsup(p-p_ref)<self.TOL*10.*Lsup(p_ref), "pressure error too big.") |
| 488 |
self.failUnless(Lsup(v-u_ref)<self.TOL*10.*Lsup(u_ref), "flux error too big.") |
| 489 |
|
| 490 |
def testConstF_FixedBottom_largeK(self): |
| 491 |
k=1.e10 |
| 492 |
mp=self.getScalarMask(include_bottom=True) |
| 493 |
mv=self.getVectorMask(include_bottom=False) |
| 494 |
u_ref,p_ref,f=self.setSolutionFixedBottom(p0=2500,pb=4000.,f0=10.,fb=10.,k=k) |
| 495 |
p=p_ref*mp |
| 496 |
u=u_ref*mv |
| 497 |
df=DarcyFlow(self.dom) |
| 498 |
df.setValue(g=f, |
| 499 |
location_of_fixed_pressure=mp, |
| 500 |
location_of_fixed_flux=mv, |
| 501 |
permeability=Scalar(k,Function(self.dom))) |
| 502 |
df.setTolerance(rtol=self.TOL) |
| 503 |
df.setSubProblemTolerance() |
| 504 |
v,p=df.solve(u,p, max_iter=100, verbose=VERBOSE) |
| 505 |
self.failUnless(Lsup(v-u_ref)<self.TOL*10.*Lsup(u_ref), "flux error too big.") |
| 506 |
self.failUnless(Lsup(p-p_ref)<self.TOL*10.*Lsup(p_ref), "pressure error too big.") |
| 507 |
|
| 508 |
def testVarioF_FixedBottom_smallK(self): |
| 509 |
k=1.e-10 |
| 510 |
mp=self.getScalarMask(include_bottom=True) |
| 511 |
mv=self.getVectorMask(include_bottom=False) |
| 512 |
u_ref,p_ref,f=self.setSolutionFixedBottom(p0=2500,pb=4000.,f0=10.,fb=30.,k=k) |
| 513 |
p=p_ref*mp |
| 514 |
u=u_ref*mv |
| 515 |
df=DarcyFlow(self.dom) |
| 516 |
df.setValue(g=f, |
| 517 |
location_of_fixed_pressure=mp, |
| 518 |
location_of_fixed_flux=mv, |
| 519 |
permeability=Scalar(k,Function(self.dom))) |
| 520 |
df.setTolerance(rtol=self.TOL) |
| 521 |
df.setSubProblemTolerance() |
| 522 |
v,p=df.solve(u,p, max_iter=100, verbose=VERBOSE) |
| 523 |
self.failUnless(Lsup(v-u_ref)<self.TOL*10.*Lsup(u_ref), "flux error too big.") |
| 524 |
self.failUnless(Lsup(p-p_ref)<self.TOL*10.*Lsup(p_ref), "pressure error too big.") |
| 525 |
|
| 526 |
def testVarioF_FixedBottom_mediumK(self): |
| 527 |
k=1. |
| 528 |
mp=self.getScalarMask(include_bottom=True) |
| 529 |
mv=self.getVectorMask(include_bottom=False) |
| 530 |
u_ref,p_ref,f=self.setSolutionFixedBottom(p0=2500,pb=4000.,f0=10.,fb=30.,k=k) |
| 531 |
p=p_ref*mp |
| 532 |
u=u_ref*mv |
| 533 |
df=DarcyFlow(self.dom) |
| 534 |
df.setValue(g=f, |
| 535 |
location_of_fixed_pressure=mp, |
| 536 |
location_of_fixed_flux=mv, |
| 537 |
permeability=Scalar(k,Function(self.dom))) |
| 538 |
df.setTolerance(rtol=self.TOL) |
| 539 |
df.setSubProblemTolerance() |
| 540 |
v,p=df.solve(u,p, max_iter=100, verbose=VERBOSE) |
| 541 |
self.failUnless(Lsup(v-u_ref)<self.TOL*10.*Lsup(u_ref), "flux error too big.") |
| 542 |
self.failUnless(Lsup(p-p_ref)<self.TOL*10.*Lsup(p_ref), "pressure error too big.") |
| 543 |
|
| 544 |
def testVarioF_FixedBottom_largeK(self): |
| 545 |
k=1.e10 |
| 546 |
mp=self.getScalarMask(include_bottom=True) |
| 547 |
mv=self.getVectorMask(include_bottom=False) |
| 548 |
u_ref,p_ref,f=self.setSolutionFixedBottom(p0=2500,pb=4000.,f0=10.,fb=30.,k=k) |
| 549 |
p=p_ref*mp |
| 550 |
u=u_ref*mv |
| 551 |
df=DarcyFlow(self.dom) |
| 552 |
df.setValue(g=f, |
| 553 |
location_of_fixed_pressure=mp, |
| 554 |
location_of_fixed_flux=mv, |
| 555 |
permeability=Scalar(k,Function(self.dom))) |
| 556 |
df.setTolerance(rtol=self.TOL) |
| 557 |
df.setSubProblemTolerance() |
| 558 |
v,p=df.solve(u,p, max_iter=100, verbose=VERBOSE) |
| 559 |
self.failUnless(Lsup(v-u_ref)<self.TOL*10.*Lsup(u_ref), "flux error too big.") |
| 560 |
self.failUnless(Lsup(p-p_ref)<self.TOL*10.*Lsup(p_ref), "pressure error too big.") |
| 561 |
|
| 562 |
def testConstF_FreeBottom_smallK(self): |
| 563 |
k=1.e-10 |
| 564 |
mp=self.getScalarMask(include_bottom=False) |
| 565 |
mv=self.getVectorMask(include_bottom=True) |
| 566 |
u_ref,p_ref,f=self.setSolutionFixedBottom(p0=2500,pb=4000.,f0=10.,fb=10.,k=k) |
| 567 |
p=p_ref*mp |
| 568 |
u=u_ref*mv |
| 569 |
df=DarcyFlow(self.dom) |
| 570 |
df.setValue(g=f, |
| 571 |
location_of_fixed_pressure=mp, |
| 572 |
location_of_fixed_flux=mv, |
| 573 |
permeability=Scalar(k,Function(self.dom))) |
| 574 |
df.setTolerance(rtol=self.TOL) |
| 575 |
df.setSubProblemTolerance() |
| 576 |
v,p=df.solve(u,p, max_iter=100, verbose=VERBOSE) |
| 577 |
self.failUnless(Lsup(v-u_ref)<self.TOL*10.*Lsup(u_ref), "flux error too big.") |
| 578 |
self.failUnless(Lsup(p-p_ref)<self.TOL*10.*Lsup(p_ref), "pressure error too big.") |
| 579 |
|
| 580 |
def testConstF_FreeBottom_mediumK(self): |
| 581 |
k=1. |
| 582 |
mp=self.getScalarMask(include_bottom=False) |
| 583 |
mv=self.getVectorMask(include_bottom=True) |
| 584 |
u_ref,p_ref,f=self.setSolutionFixedBottom(p0=2500,pb=4000.,f0=10.,fb=10.,k=k) |
| 585 |
p=p_ref*mp |
| 586 |
u=u_ref*mv |
| 587 |
df=DarcyFlow(self.dom) |
| 588 |
df.setValue(g=f, |
| 589 |
location_of_fixed_pressure=mp, |
| 590 |
location_of_fixed_flux=mv, |
| 591 |
permeability=Scalar(k,Function(self.dom))) |
| 592 |
df.setTolerance(rtol=self.TOL) |
| 593 |
df.setSubProblemTolerance() |
| 594 |
v,p=df.solve(u,p, max_iter=100, verbose=VERBOSE) |
| 595 |
self.failUnless(Lsup(v-u_ref)<self.TOL*10.*Lsup(u_ref), "flux error too big.") |
| 596 |
self.failUnless(Lsup(p-p_ref)<self.TOL*10.*Lsup(p_ref), "pressure error too big.") |
| 597 |
|
| 598 |
def testConstF_FreeBottom_largeK(self): |
| 599 |
k=1.e10 |
| 600 |
mp=self.getScalarMask(include_bottom=False) |
| 601 |
mv=self.getVectorMask(include_bottom=True) |
| 602 |
u_ref,p_ref,f=self.setSolutionFixedBottom(p0=2500,pb=4000.,f0=10.,fb=10.,k=k) |
| 603 |
p=p_ref*mp |
| 604 |
u=u_ref*mv |
| 605 |
df=DarcyFlow(self.dom) |
| 606 |
df.setValue(g=f, |
| 607 |
location_of_fixed_pressure=mp, |
| 608 |
location_of_fixed_flux=mv, |
| 609 |
permeability=Scalar(k,Function(self.dom))) |
| 610 |
df.setTolerance(rtol=self.TOL) |
| 611 |
df.setSubProblemTolerance() |
| 612 |
v,p=df.solve(u,p, max_iter=100, verbose=VERBOSE) |
| 613 |
self.failUnless(Lsup(v-u_ref)<self.TOL*10.*Lsup(u_ref), "flux error too big.") |
| 614 |
self.failUnless(Lsup(p-p_ref)<self.TOL*10.*Lsup(p_ref), "pressure error too big.") |
| 615 |
|
| 616 |
def testVarioF_FreeBottom_smallK(self): |
| 617 |
k=1.e-10 |
| 618 |
mp=self.getScalarMask(include_bottom=False) |
| 619 |
mv=self.getVectorMask(include_bottom=True) |
| 620 |
u_ref,p_ref,f=self.setSolutionFixedBottom(p0=2500,pb=4000.,f0=10.,fb=30.,k=k) |
| 621 |
p=p_ref*mp |
| 622 |
u=u_ref*mv |
| 623 |
df=DarcyFlow(self.dom) |
| 624 |
df.setValue(g=f, |
| 625 |
location_of_fixed_pressure=mp, |
| 626 |
location_of_fixed_flux=mv, |
| 627 |
permeability=Scalar(k,Function(self.dom))) |
| 628 |
df.setTolerance(rtol=self.TOL) |
| 629 |
df.setSubProblemTolerance() |
| 630 |
v,p=df.solve(u,p, max_iter=100, verbose=VERBOSE) |
| 631 |
self.failUnless(Lsup(v-u_ref)<self.TOL*25.*Lsup(u_ref), "flux error too big.") # 25 because of disc. error. |
| 632 |
self.failUnless(Lsup(p-p_ref)<self.TOL*25.*Lsup(p_ref), "pressure error too big.") |
| 633 |
|
| 634 |
def testVarioF_FreeBottom_mediumK(self): |
| 635 |
k=1. |
| 636 |
mp=self.getScalarMask(include_bottom=False) |
| 637 |
mv=self.getVectorMask(include_bottom=True) |
| 638 |
u_ref,p_ref,f=self.setSolutionFixedBottom(p0=2500,pb=4000.,f0=10.,fb=30.,k=k) |
| 639 |
p=p_ref*mp |
| 640 |
u=u_ref*mv |
| 641 |
df=DarcyFlow(self.dom) |
| 642 |
df.setValue(g=f, |
| 643 |
location_of_fixed_pressure=mp, |
| 644 |
location_of_fixed_flux=mv, |
| 645 |
permeability=Scalar(k,Function(self.dom))) |
| 646 |
df.setTolerance(rtol=self.TOL) |
| 647 |
df.setSubProblemTolerance() |
| 648 |
v,p=df.solve(u,p, max_iter=100, verbose=VERBOSE) |
| 649 |
self.failUnless(Lsup(v-u_ref)<self.TOL*10.*Lsup(u_ref), "flux error too big.") |
| 650 |
self.failUnless(Lsup(p-p_ref)<self.TOL*10.*Lsup(p_ref), "pressure error too big.") |
| 651 |
|
| 652 |
def testVarioF_FreeBottom_largeK(self): |
| 653 |
k=1.e10 |
| 654 |
mp=self.getScalarMask(include_bottom=False) |
| 655 |
mv=self.getVectorMask(include_bottom=True) |
| 656 |
u_ref,p_ref,f=self.setSolutionFixedBottom(p0=2500,pb=4000.,f0=10.,fb=30.,k=k) |
| 657 |
p=p_ref*mp |
| 658 |
u=u_ref*mv |
| 659 |
df=DarcyFlow(self.dom) |
| 660 |
df.setValue(g=f, |
| 661 |
location_of_fixed_pressure=mp, |
| 662 |
location_of_fixed_flux=mv, |
| 663 |
permeability=Scalar(k,Function(self.dom))) |
| 664 |
df.setTolerance(rtol=self.TOL) |
| 665 |
df.setSubProblemTolerance() |
| 666 |
v,p=df.solve(u,p, max_iter=100, verbose=VERBOSE) |
| 667 |
self.failUnless(Lsup(v-u_ref)<self.TOL*10.*Lsup(u_ref), "flux error too big.") |
| 668 |
self.failUnless(Lsup(p-p_ref)<self.TOL*10.*Lsup(p_ref), "pressure error too big.") |
| 669 |
|
| 670 |
class Test_Darcy2D(Test_Darcy): |
| 671 |
TOL=1e-4 |
| 672 |
WIDTH=1. |
| 673 |
def setUp(self): |
| 674 |
NE=40 # wrning smaller NE may case a failure for VarioF tests due to discretization errors. |
| 675 |
self.dom = Rectangle(NE,NE) |
| 676 |
self.rescaleDomain() |
| 677 |
def tearDown(self): |
| 678 |
del self.dom |
| 679 |
class Test_Darcy3D(Test_Darcy): |
| 680 |
TOL=1e-4 |
| 681 |
WIDTH=1. |
| 682 |
def setUp(self): |
| 683 |
NE=25 # wrning smaller NE may case a failure for VarioF tests due to discretization errors. |
| 684 |
self.dom = Brick(NE,NE,NE) |
| 685 |
self.rescaleDomain() |
| 686 |
def tearDown(self): |
| 687 |
del self.dom |
| 688 |
|
| 689 |
|
| 690 |
class Test_Mountains3D(unittest.TestCase): |
| 691 |
def setUp(self): |
| 692 |
NE=16 |
| 693 |
self.TOL=1e-4 |
| 694 |
self.domain=Brick(NE,NE,NE,order=1,useFullElementOrder=True) |
| 695 |
def tearDown(self): |
| 696 |
del self.domain |
| 697 |
def test_periodic(self): |
| 698 |
EPS=0.01 |
| 699 |
|
| 700 |
x=self.domain.getX() |
| 701 |
v = Vector(0.0, Solution(self.domain)) |
| 702 |
a0=1 |
| 703 |
a1=1 |
| 704 |
n0=2 |
| 705 |
n1=2 |
| 706 |
n2=0.5 |
| 707 |
a2=-(a0*n0+a1*n1)/n2 |
| 708 |
v[0]=a0*sin(pi*n0*x[0])* cos(pi*n1*x[1])* cos(pi*n2*x[2]) |
| 709 |
v[1]=a1*cos(pi*n0*x[0])* sin(pi*n1*x[1])* cos(pi*n2*x[2]) |
| 710 |
v[2]=a2*cos(pi*n0*x[0])* cos(pi*n1*x[1])* sin(pi*n2*x[2]) |
| 711 |
|
| 712 |
H_t=Scalar(0.0, Solution(self.domain)) |
| 713 |
mts=Mountains(self.domain,v,eps=EPS,z=1) |
| 714 |
u,Z=mts.update(u=v,H_t=H_t) |
| 715 |
|
| 716 |
error_int=integrate(Z) |
| 717 |
self.failUnless(error_int<self.TOL, "Boundary intergral is too large.") |
| 718 |
|
| 719 |
class Test_Mountains2D(unittest.TestCase): |
| 720 |
def setUp(self): |
| 721 |
NE=16 |
| 722 |
self.TOL=1e-4 |
| 723 |
self.domain=Rectangle(NE,NE,order=1,useFullElementOrder=True) |
| 724 |
def tearDown(self): |
| 725 |
del self.domain |
| 726 |
def test_periodic(self): |
| 727 |
EPS=0.01 |
| 728 |
|
| 729 |
x=self.domain.getX() |
| 730 |
v = Vector(0.0, Solution(self.domain)) |
| 731 |
a0=1 |
| 732 |
n0=1 |
| 733 |
n1=0.5 |
| 734 |
a1=-(a0*n0)/n1 |
| 735 |
v[0]=a0*sin(pi*n0*x[0])* cos(pi*n1*x[1]) |
| 736 |
v[1]=a1*cos(pi*n0*x[0])* sin(pi*n1*x[1]) |
| 737 |
|
| 738 |
H_t=Scalar(0.0, Solution(self.domain)) |
| 739 |
mts=Mountains(self.domain,v,eps=EPS,z=1) |
| 740 |
u,Z=mts.update(u=v,H_t=H_t) |
| 741 |
|
| 742 |
error_int=integrate(Z) |
| 743 |
self.failUnless(error_int<self.TOL, "Boundary intergral is too large.") |
| 744 |
|
| 745 |
|
| 746 |
|
| 747 |
class Test_Rheologies(unittest.TestCase): |
| 748 |
|
| 749 |
""" |
| 750 |
this is the program used to generate the powerlaw tests: |
| 751 |
|
| 752 |
TAU_Y=100. |
| 753 |
N=10 |
| 754 |
M=5 |
| 755 |
|
| 756 |
def getE(tau): |
| 757 |
if tau<=TAU_Y: |
| 758 |
return 1./(0.5+20*sqrt(tau)) |
| 759 |
else: |
| 760 |
raise ValueError,"out of range." |
| 761 |
tau=[ i* (TAU_Y/N) for i in range(N+1)] + [TAU_Y for i in range(M)] |
| 762 |
e=[ tau[i]/getE(tau[i]) for i in range(N+1)] |
| 763 |
e+= [ (i+1+j)* (max(e)/N) for j in range(M) ] |
| 764 |
|
| 765 |
print tau |
| 766 |
print e |
| 767 |
""" |
| 768 |
TOL=1.e-8 |
| 769 |
def test_PowerLaw_Init(self): |
| 770 |
pl=PowerLaw(numMaterials=3,verbose=VERBOSE) |
| 771 |
|
| 772 |
self.failUnlessEqual(pl.getNumMaterials(),3,"num materials is wrong") |
| 773 |
self.failUnlessEqual(pl.validMaterialId(0),True,"material id 0 not found") |
| 774 |
self.failUnlessEqual(pl.validMaterialId(1),True,"material id 1 not found") |
| 775 |
self.failUnlessEqual(pl.validMaterialId(2),True,"material id 2 not found") |
| 776 |
self.failUnlessEqual(pl.validMaterialId(3),False,"material id 3 not found") |
| 777 |
|
| 778 |
self.failUnlessEqual(pl.getFriction(),None,"initial friction wrong.") |
| 779 |
self.failUnlessEqual(pl.getTauY(),None,"initial tau y wrong.") |
| 780 |
pl.setDruckerPragerLaw(tau_Y=10,friction=3) |
| 781 |
self.failUnlessEqual(pl.getFriction(),3,"friction wrong.") |
| 782 |
self.failUnlessEqual(pl.getTauY(),10,"tau y wrong.") |
| 783 |
|
| 784 |
self.failUnlessEqual(pl.getElasticShearModulus(),None,"initial shear modulus is wrong") |
| 785 |
pl.setElasticShearModulus(1000) |
| 786 |
self.failUnlessEqual(pl.getElasticShearModulus(),1000.,"shear modulus is wrong") |
| 787 |
|
| 788 |
e=pl.getEtaN() |
| 789 |
self.failUnlessEqual(len(e),3,"initial length of etaN is wrong.") |
| 790 |
self.failUnlessEqual(e,[None, None, None],"initial etaN are wrong.") |
| 791 |
self.failUnlessEqual(pl.getEtaN(0),None,"initial material id 0 is wrong") |
| 792 |
self.failUnlessEqual(pl.getEtaN(1),None,"initial material id 1 is wrong") |
| 793 |
self.failUnlessEqual(pl.getEtaN(2),None,"initial material id 2 is wrong") |
| 794 |
self.failUnlessRaises(ValueError, pl.getEtaN, 3) |
| 795 |
|
| 796 |
self.failUnlessRaises(ValueError,pl.setPowerLaws,eta_N=[10,20,30,40],tau_t=[2], power=[5,6,7]) |
| 797 |
self.failUnlessRaises(ValueError,pl.setPowerLaws,eta_N=[10,20,30],tau_t=[2,4,5,5], power=[5,6,7]) |
| 798 |
self.failUnlessRaises(ValueError,pl.setPowerLaws,eta_N=[10,20,30],tau_t=[2,1,2], power=[5,6,7,7]) |
| 799 |
|
| 800 |
pl.setPowerLaws(eta_N=[10,20,30],tau_t=[100,200,300], power=[1,2,3]) |
| 801 |
self.failUnlessEqual(pl.getPower(),[1,2,3],"powers are wrong.") |
| 802 |
self.failUnlessEqual(pl.getTauT(),[100,200,300],"tau t are wrong.") |
| 803 |
self.failUnlessEqual(pl.getEtaN(),[10,20,30],"etaN are wrong.") |
| 804 |
|
| 805 |
pl.setPowerLaw(id=0,eta_N=40,tau_t=400, power=4) |
| 806 |
self.failUnlessEqual(pl.getPower(),[4,2,3],"powers are wrong.") |
| 807 |
self.failUnlessEqual(pl.getTauT(),[400,200,300],"tau t are wrong.") |
| 808 |
self.failUnlessEqual(pl.getEtaN(),[40,20,30],"etaN are wrong.") |
| 809 |
|
| 810 |
self.failUnlessRaises(ValueError,pl.getPower,-1) |
| 811 |
self.failUnlessEqual(pl.getPower(0),4,"power 0 is wrong.") |
| 812 |
self.failUnlessEqual(pl.getPower(1),2,"power 1 is wrong.") |
| 813 |
self.failUnlessEqual(pl.getPower(2),3,"power 2 is wrong.") |
| 814 |
self.failUnlessRaises(ValueError,pl.getPower,3) |
| 815 |
|
| 816 |
self.failUnlessRaises(ValueError,pl.getTauT,-1) |
| 817 |
self.failUnlessEqual(pl.getTauT(0),400,"tau t 0 is wrong.") |
| 818 |
self.failUnlessEqual(pl.getTauT(1),200,"tau t 1 is wrong.") |
| 819 |
self.failUnlessEqual(pl.getTauT(2),300,"tau t 2 is wrong.") |
| 820 |
self.failUnlessRaises(ValueError,pl.getTauT,3) |
| 821 |
|
| 822 |
self.failUnlessRaises(ValueError,pl.getEtaN,-1) |
| 823 |
self.failUnlessEqual(pl.getEtaN(0),40,"eta n 0 is wrong.") |
| 824 |
self.failUnlessEqual(pl.getEtaN(1),20,"eta n 1 is wrong.") |
| 825 |
self.failUnlessEqual(pl.getEtaN(2),30,"eta n 2 is wrong.") |
| 826 |
self.failUnlessRaises(ValueError,pl.getEtaN,3) |
| 827 |
|
| 828 |
def checkResult(self,id,gamma_dot_, eta, tau_ref): |
| 829 |
self.failUnless(eta>=0,"eta needs to be positive (test %s)"%id) |
| 830 |
error=abs(gamma_dot_*eta-tau_ref) |
| 831 |
self.failUnless(error<=self.TOL*tau_ref,"eta is wrong: error = gamma_dot_*eta-tau_ref = %s * %s - %s = %s (test %s)"%(gamma_dot_,eta,tau_ref,error,id)) |
| 832 |
|
| 833 |
def test_PowerLaw_Linear(self): |
| 834 |
taus= [0.0, 10.0, 20.0, 30.0, 40.0, 50.0, 60.0, 70.0, 80.0, 90.0, 100.0, 100.0, 100.0, 100.0, 100.0, 100.0] |
| 835 |
gamma_dot_s=[0.0, 5.0, 10.0, 15.0, 20.0, 25.0, 30.0, 35.0, 40.0, 45.0, 50.0, 55.0, 60.0, 65.0, 70.0, 75.0] |
| 836 |
pl=PowerLaw(numMaterials=1,verbose=VERBOSE) |
| 837 |
pl.setDruckerPragerLaw(tau_Y=100.) |
| 838 |
pl.setPowerLaw(eta_N=2.) |
| 839 |
pl.setEtaTolerance(self.TOL) |
| 840 |
for i in xrange(len(taus)): self.checkResult(i,gamma_dot_s[i], pl.getEtaEff(gamma_dot_s[i]),taus[i]) |
| 841 |
|
| 842 |
def test_PowerLaw_QuadLarge(self): |
| 843 |
taus=[0.0, 10.0, 20.0, 30.0, 40.0, 50.0, 60.0, 70.0, 80.0, 90.0, 100.0, 100.0, 100.0, 100.0, 100.0, 100.0] |
| 844 |
gamma_dot_s=[0.0, 637.45553203367592, 1798.8543819998317, 3301.3353450309969, 5079.6442562694074, 7096.067811865476, 9325.1600308977995, 11748.240371477059, 14350.835055998656, 17121.29936490925, 20050.0, 22055.0, 24060.0, 26065.0, 28070.0, 30075.0] |
| 845 |
pl=PowerLaw(numMaterials=2,verbose=VERBOSE) |
| 846 |
pl.setDruckerPragerLaw(tau_Y=100.) |
| 847 |
pl.setPowerLaws(eta_N=[2.,0.01],tau_t=[1, 25.], power=[1,2]) |
| 848 |
pl.setEtaTolerance(self.TOL) |
| 849 |
for i in xrange(len(taus)): self.checkResult(i,gamma_dot_s[i], pl.getEtaEff(gamma_dot_s[i]),taus[i]) |
| 850 |
|
| 851 |
def test_PowerLaw_QuadSmall(self): |
| 852 |
taus=[0.0, 10.0, 20.0, 30.0, 40.0, 50.0, 60.0, 70.0, 80.0, 90.0, 100.0, 100.0, 100.0, 100.0, 100.0, 100.0] |
| 853 |
gamma_dot_s=[0.0, 5.632455532033676, 11.788854381999831, 18.286335345030995, 25.059644256269408, 32.071067811865476, 39.295160030897804, 46.713240371477056, 54.310835055998652, 62.076299364909254, 70.0, 77.0, 84.0, 91.0, 98.0, 105.0] |
| 854 |
pl=PowerLaw(numMaterials=2,verbose=VERBOSE) |
| 855 |
pl.setDruckerPragerLaw(tau_Y=100.) |
| 856 |
pl.setPowerLaws(eta_N=[2.,10.],tau_t=[1, 25.], power=[1,2]) |
| 857 |
pl.setEtaTolerance(self.TOL) |
| 858 |
for i in xrange(len(taus)): self.checkResult(i,gamma_dot_s[i], pl.getEtaEff(gamma_dot_s[i]),taus[i]) |
| 859 |
|
| 860 |
def test_PowerLaw_CubeLarge(self): |
| 861 |
taus=[0.0, 10.0, 20.0, 30.0, 40.0, 50.0, 60.0, 70.0, 80.0, 90.0, 100.0, 100.0, 100.0, 100.0, 100.0, 100.0] |
| 862 |
gamma_dot_s=[0.0, 51.415888336127786, 157.36125994561547, 304.6468153816889, 487.84283811405851, 703.60440414872664, 949.57131887226353, 1223.9494765692673, 1525.3084267560891, 1852.4689652218574, 2204.4346900318833, 2424.8781590350718, 2645.3216280382603, 2865.7650970414484, 3086.2085660446369, 3306.6520350478249] |
| 863 |
pl=PowerLaw(numMaterials=2,verbose=VERBOSE) |
| 864 |
pl.setDruckerPragerLaw(tau_Y=100.) |
| 865 |
pl.setPowerLaws(eta_N=[2.,1./16.],tau_t=[1, 64.], power=[1,3]) |
| 866 |
pl.setEtaTolerance(self.TOL) |
| 867 |
for i in xrange(len(taus)): self.checkResult(i,gamma_dot_s[i], pl.getEtaEff(gamma_dot_s[i]),taus[i]) |
| 868 |
|
| 869 |
def test_PowerLaw_CubeSmall(self): |
| 870 |
taus=[0.0, 10.0, 20.0, 30.0, 40.0, 50.0, 60.0, 70.0, 80.0, 90.0, 100.0, 100.0, 100.0, 100.0, 100.0, 100.0] |
| 871 |
gamma_dot_s=[0.0, 5.4641588833612778, 11.473612599456157, 17.89646815381689, 24.678428381140588, 31.786044041487269, 39.195713188722635, 46.889494765692675, 54.853084267560895, 63.074689652218574, 71.544346900318828, 78.698781590350706, 85.853216280382583, 93.007650970414474, 100.16208566044635, 107.316520350478] |
| 872 |
pl=PowerLaw(numMaterials=2,verbose=VERBOSE) |
| 873 |
pl.setDruckerPragerLaw(tau_Y=100.) |
| 874 |
pl.setPowerLaws(eta_N=[2.,25./4.],tau_t=[1, 64.], power=[1,3]) |
| 875 |
pl.setEtaTolerance(self.TOL) |
| 876 |
for i in xrange(len(taus)): self.checkResult(i,gamma_dot_s[i], pl.getEtaEff(gamma_dot_s[i]),taus[i]) |
| 877 |
|
| 878 |
def test_PowerLaw_QuadLarge_CubeLarge(self): |
| 879 |
taus=[0.0, 10.0, 20.0, 30.0, 40.0, 50.0, 60.0, 70.0, 80.0, 90.0, 100.0, 100.0, 100.0, 100.0, 100.0, 100.0] |
| 880 |
gamma_dot_s=[0.0, 683.87142036980367, 1946.2156419454475, 3590.982160412686, 5547.4870943834667, 7774.6722160142008, 10244.731349770063, 12937.189848046326, 15836.143482754744, 18928.768330131104, 22204.434690031885, 24424.878159035074, 26645.321628038262, 28865.765097041451, 31086.208566044639, 33306.652035047824] |
| 881 |
pl=PowerLaw(numMaterials=3,verbose=VERBOSE) |
| 882 |
pl.setDruckerPragerLaw(tau_Y=100.) |
| 883 |
pl.setPowerLaws(eta_N=[2.,0.01,1./16.],tau_t=[1, 25.,64.], power=[1,2,3]) |
| 884 |
pl.setEtaTolerance(self.TOL) |
| 885 |
for i in xrange(len(taus)): self.checkResult(i,gamma_dot_s[i], pl.getEtaEff(gamma_dot_s[i]),taus[i]) |
| 886 |
|
| 887 |
def test_PowerLaw_QuadLarge_CubeSmall(self): |
| 888 |
taus=[0.0, 10.0, 20.0, 30.0, 40.0, 50.0, 60.0, 70.0, 80.0, 90.0, 100.0, 100.0, 100.0, 100.0, 100.0, 100.0] |
| 889 |
gamma_dot_s=[0.0, 637.9196909170372, 1800.3279945992881, 3304.2318131848137, 5084.3226846505486, 7102.853855906963, 9334.3557440865225, 11760.129866242751, 14365.688140266215, 17139.374054561471, 20071.544346900318, 22078.698781590349, 24085.853216280382, 26093.007650970416, 28100.162085660446, 30107.316520350476] |
| 890 |
pl=PowerLaw(numMaterials=3,verbose=VERBOSE) |
| 891 |
pl.setDruckerPragerLaw(tau_Y=100.) |
| 892 |
pl.setPowerLaws(eta_N=[2.,0.01,25./4.],tau_t=[1, 25.,64.], power=[1,2,3]) |
| 893 |
pl.setEtaTolerance(self.TOL) |
| 894 |
for i in xrange(len(taus)): self.checkResult(i,gamma_dot_s[i], pl.getEtaEff(gamma_dot_s[i]),taus[i]) |
| 895 |
|
| 896 |
def test_PowerLaw_QuadSmall_CubeLarge(self): |
| 897 |
taus=[0.0, 10.0, 20.0, 30.0, 40.0, 50.0, 60.0, 70.0, 80.0, 90.0, 100.0, 100.0, 100.0, 100.0, 100.0, 100.0] |
| 898 |
gamma_dot_s=[0.0, 52.04834386816146, 159.15011432761528, 307.93315072671987, 492.9024823703279, 710.67547196059206, 958.86647890316135, 1235.6627169407443, 1539.6192618120876, 1869.5452645867665, 2224.4346900318833, 2446.8781590350718, 2669.3216280382603, 2891.7650970414484, 3114.2085660446369, 3336.6520350478249] |
| 899 |
|
| 900 |
pl=PowerLaw(numMaterials=3,verbose=VERBOSE) |
| 901 |
pl.setDruckerPragerLaw(tau_Y=100.) |
| 902 |
pl.setPowerLaws(eta_N=[2.,10.,1./16.],tau_t=[1, 25.,64.], power=[1,2,3]) |
| 903 |
pl.setEtaTolerance(self.TOL) |
| 904 |
for i in xrange(len(taus)): self.checkResult(i,gamma_dot_s[i], pl.getEtaEff(gamma_dot_s[i]),taus[i]) |
| 905 |
|
| 906 |
def test_PowerLaw_QuadSmall_CubeSmall(self): |
| 907 |
taus=[0.0, 10.0, 20.0, 30.0, 40.0, 50.0, 60.0, 70.0, 80.0, 90.0, 100.0, 100.0, 100.0, 100.0, 100.0, 100.0] |
| 908 |
gamma_dot_s=[0.0, 6.0966144153949529, 13.262466981455987, 21.182803498847885, 29.738072637409996, 38.857111853352741, 48.49087321962044, 58.602735137169738, 69.16391932355954, 80.150989017127827, 91.544346900318828, 100.69878159035071, 109.85321628038258, 119.00765097041447, 128.16208566044637, 137.31652035047824] |
| 909 |
pl=PowerLaw(numMaterials=3,verbose=VERBOSE) |
| 910 |
pl.setDruckerPragerLaw(tau_Y=100.) |
| 911 |
pl.setPowerLaws(eta_N=[2.,10.,25./4.],tau_t=[1, 25.,64.], power=[1,2,3]) |
| 912 |
pl.setEtaTolerance(self.TOL) |
| 913 |
for i in xrange(len(taus)): self.checkResult(i,gamma_dot_s[i], pl.getEtaEff(gamma_dot_s[i]),taus[i]) |
| 914 |
|
| 915 |
def test_PowerLaw_withShear(self): |
| 916 |
taus=[0.0, 10.0, 20.0, 30.0, 40.0, 50.0, 60.0, 70.0, 80.0, 90.0, 100.0, 100.0, 100.0, 100.0, 100.0, 100.0] |
| 917 |
gamma_dot_s=[0.0, 15.0, 30.0, 45.0, 60.0, 75.0, 90.0, 105.0, 120.0, 135.0, 150.0, 165.0, 180.0, 195.0, 210.0, 225.0] |
| 918 |
pl=PowerLaw(numMaterials=1,verbose=VERBOSE) |
| 919 |
pl.setDruckerPragerLaw(tau_Y=100.) |
| 920 |
pl.setPowerLaw(eta_N=2.) |
| 921 |
pl.setElasticShearModulus(3.) |
| 922 |
dt=1./3. |
| 923 |
pl.setEtaTolerance(self.TOL) |
| 924 |
self.failUnlessRaises(ValueError, pl.getEtaEff,gamma_dot_s[0]) |
| 925 |
for i in xrange(len(taus)): self.checkResult(i,gamma_dot_s[i], pl.getEtaEff(gamma_dot_s[i],dt=dt),taus[i]) |
| 926 |
|
| 927 |
if __name__ == '__main__': |
| 928 |
suite = unittest.TestSuite() |
| 929 |
|
| 930 |
suite.addTest(unittest.makeSuite(Test_StokesProblemCartesian2D)) |
| 931 |
suite.addTest(unittest.makeSuite(Test_Darcy3D)) |
| 932 |
suite.addTest(unittest.makeSuite(Test_Darcy2D)) |
| 933 |
suite.addTest(unittest.makeSuite(Test_StokesProblemCartesian3D)) |
| 934 |
suite.addTest(unittest.makeSuite(Test_Mountains3D)) |
| 935 |
suite.addTest(unittest.makeSuite(Test_Mountains2D)) |
| 936 |
suite.addTest(unittest.makeSuite(Test_Rheologies)) |
| 937 |
s=unittest.TextTestRunner(verbosity=2).run(suite) |
| 938 |
if not s.wasSuccessful(): sys.exit(1) |
| 939 |
|
Parent Directory
| 
Revision Log