def f_CG(t):
timing = t * sqrt(g * h0) / L # dimensionless
w, z, u = analytical_sol(array([-0.5]), timing) # dimensionless
wB = w * h0 # dimensional
uB = u * sqrt(g * h0) # dimensional
zB = z * h0 # dimensional
hB = wB - zB # dimensional
pB = uB * hB # dimensional
#[ 'stage', 'Xmomentum', 'Ymomentum']
return hstack((wB, pB, array([0.0]))) # dimensional
def f_CG(t):
timing = t*sqrt(g*h0)/L # dimensionless
w, z, u = analytical_sol(array([-0.5]),timing) # dimensionless
wB = w*h0 # dimensional
uB = u*sqrt(g*h0) # dimensional
zB = z*h0 # dimensional
hB = wB - zB # dimensional
pB = uB * hB # dimensional
#[ 'stage', 'Xmomentum', 'Ymomentum']
return hstack( (wB, pB, array([0.0])) ) # dimensional
def test_carrier_greenspan_transient(self):
if verbose:
print
print indent + 'Running simulation script'
s = 'numerical_cg_transient.py'
res = anuga.run_anuga_script(s, args=args)
# Test that script runs ok
assert res == 0
if verbose:
print indent + 'Testing accuracy'
import anuga.utilities.plot_utils as util
from analytical_cg_transient import analytical_sol
p_st = util.get_output('carrier_greenspan.sww')
p2_st = util.get_centroids(p_st)
v = p2_st.y[10]
v2 = (p2_st.y == v)
x_n = p2_st.x[v2]
ids = [1, 15, 30]
use_absolute = [False, True, True]
n = len(ids)
W, UH, Z, H, U = [], [], [], [], []
W_n, U_n, UH_n = [], [], []
#Dimensional parameters
L = 5e4 # Length of channel (m)
h0 = 5e2 # Height at origin when the water is still
g = 9.81 # Gravity
for i, id in enumerate(ids):
w, z, u = analytical_sol(x_n / L,
p2_st.time[id] * sqrt(g * h0) / L)
w = w * h0
z = z * h0
u = u * sqrt(g * h0)
h = w - z
uh = u * h
W.append(w)
Z.append(z)
H.append(h)
U.append(u)
UH.append(uh)
W_n.append(p2_st.stage[id, v2])
UH_n.append(p2_st.xmom[id, v2])
U_n.append(p2_st.xvel[id, v2])
#print id, numpy.max(H[i]), numpy.min(H[i])
#print id, numpy.max(U[i]), numpy.min(U[i])
#print id, numpy.max(U_n[i]), numpy.min(U_n[i])
#Test stages
# Calculate L^1 error at times
ew = numpy.zeros(n)
for i, id in enumerate(ids):
ew[i] = numpy.sum(numpy.abs(W_n[i] - W[i])) / numpy.sum(
numpy.abs(W[i]))
print
print indent + 'L^1 Errors in stage: ', ew
#Test xmomenta
# Calculate L^1 error at times
euh = numpy.zeros(n)
for i, id in enumerate(ids):
euh[i] = numpy.sum(numpy.abs(UH_n[i] - U[i] * H[i])) / numpy.sum(
numpy.abs(UH_n[i]))
print indent + 'L^1 Errors in xmomentum: ', euh
#Test xvelocity
# Calculate L^1 error at times
eu = numpy.zeros(n)
for i, id in enumerate(ids):
if use_absolute[i]:
eu[i] = numpy.sum(numpy.abs(U_n[i] - U[i])) / len(U[i])
else:
eu[i] = numpy.sum(numpy.abs(U_n[i] - U[i])) / numpy.sum(
numpy.abs(U[i]))
print indent + 'L^1 Errors in xvelocity: ', eu
for i, id in enumerate(ids):
assert ew[i] < 0.01, 'L^1 error %g greater than 1 percent' % ew[i]
for i, id in enumerate(ids):
assert euh[
i] < 0.025, 'L^1 error %g greater than 2.5 percent' % euh[i]
for i, id in enumerate(ids):
assert eu[
i] < 0.025, 'L^1 error %g greater than 2.5 percent' % eu[i]
def stage(x, y):
w, z, u = analytical_sol(x / L, 0.0)
return h0 * w
def test_carrier_greenspan_transient(self):
if verbose:
print
print indent+'Running simulation script'
s = 'numerical_cg_transient.py'
res = anuga.run_anuga_script(s,args=args)
# Test that script runs ok
assert res == 0
if verbose:
print indent+'Testing accuracy'
import anuga.utilities.plot_utils as util
from analytical_cg_transient import analytical_sol
p_st = util.get_output('carrier_greenspan.sww')
p2_st=util.get_centroids(p_st)
v = p2_st.y[10]
v2=(p2_st.y==v)
x_n = p2_st.x[v2]
ids = [1, 15, 30]
use_absolute = [False, True, True]
n = len(ids)
W, UH, Z, H, U = [], [], [], [], []
W_n, U_n, UH_n = [], [], []
#Dimensional parameters
L = 5e4 # Length of channel (m)
h0 = 5e2 # Height at origin when the water is still
g = 9.81 # Gravity
for i, id in enumerate(ids):
w, z, u = analytical_sol(x_n/L, p2_st.time[id]*sqrt(g*h0)/L)
w = w*h0
z = z*h0
u = u*sqrt(g*h0)
h = w-z
uh = u*h
W.append(w)
Z.append(z)
H.append(h)
U.append(u)
UH.append(uh)
W_n.append(p2_st.stage[id,v2])
UH_n.append(p2_st.xmom[id,v2])
U_n.append(p2_st.xvel[id,v2])
#print id, numpy.max(H[i]), numpy.min(H[i])
#print id, numpy.max(U[i]), numpy.min(U[i])
#print id, numpy.max(U_n[i]), numpy.min(U_n[i])
#Test stages
# Calculate L^1 error at times
ew = numpy.zeros(n)
for i, id in enumerate(ids):
ew[i] = numpy.sum(numpy.abs(W_n[i]-W[i]))/numpy.sum(numpy.abs(W[i]))
print
print indent+'L^1 Errors in stage: ', ew
#Test xmomenta
# Calculate L^1 error at times
euh = numpy.zeros(n)
for i, id in enumerate(ids):
euh[i] = numpy.sum(numpy.abs(UH_n[i]-U[i]*H[i]))/numpy.sum(numpy.abs(UH_n[i]))
print indent+'L^1 Errors in xmomentum: ',euh
#Test xvelocity
# Calculate L^1 error at times
eu = numpy.zeros(n)
for i, id in enumerate(ids):
if use_absolute[i]:
eu[i] = numpy.sum(numpy.abs(U_n[i]-U[i]))/len(U[i])
else:
eu[i] = numpy.sum(numpy.abs(U_n[i]-U[i]))/numpy.sum(numpy.abs(U[i]))
print indent+'L^1 Errors in xvelocity: ', eu
for i, id in enumerate(ids):
assert ew[i] < 0.01, 'L^1 error %g greater than 1 percent'% ew[i]
for i, id in enumerate(ids):
assert euh[i] < 0.025, 'L^1 error %g greater than 2.5 percent'% euh[i]
for i, id in enumerate(ids):
assert eu[i] < 0.025, 'L^1 error %g greater than 2.5 percent'% eu[i]
def stage(x,y):
w,z,u = analytical_sol(x/L,0.0)
return h0*w