def shallow2D(use_petsc=False,iplot=0,htmlplot=False,outdir='./_output',solver_type='classic'):
#===========================================================================
# Import libraries
#===========================================================================
import numpy as np
if use_petsc:
import petclaw as pyclaw
else:
import pyclaw
#===========================================================================
# Setup solver and solver parameters
#===========================================================================
if solver_type == 'classic':
solver = pyclaw.ClawSolver2D()
solver.limiters = pyclaw.limiters.tvd.MC
solver.dimensional_split=1
elif solver_type == 'sharpclaw':
solver = pyclaw.SharpClawSolver2D()
import riemann
solver.rp = riemann.rp2_shallow_roe_with_efix
solver.num_waves = 3
solver.bc_lower[0] = pyclaw.BC.extrap
solver.bc_upper[0] = pyclaw.BC.wall
solver.bc_lower[1] = pyclaw.BC.extrap
solver.bc_upper[1] = pyclaw.BC.wall
#===========================================================================
# Initialize domain and state, then initialize the solution associated to the
# state and finally initialize aux array
#===========================================================================
# Domain:
xlower = -2.5
xupper = 2.5
mx = 150
ylower = -2.5
yupper = 2.5
my = 150
x = pyclaw.Dimension('x',xlower,xupper,mx)
y = pyclaw.Dimension('y',ylower,yupper,my)
domain = pyclaw.Domain([x,y])
num_eqn = 3 # Number of equations
state = pyclaw.State(domain,num_eqn)
grav = 1.0 # Parameter (global auxiliary variable)
state.problem_data['grav'] = grav
# Initial solution
# ================
# Riemann states of the dam break problem
damRadius = 0.5
hl = 2.
ul = 0.
vl = 0.
hr = 1.
ur = 0.
vr = 0.
qinit(state,hl,ul,vl,hr,ur,vl,damRadius) # This function is defined above
#===========================================================================
# Set up controller and controller parameters
#===========================================================================
claw = pyclaw.Controller()
claw.tfinal = 2.5
claw.solution = pyclaw.Solution(state,domain)
claw.solver = solver
claw.outdir = outdir
claw.num_output_times = 10
#===========================================================================
# Solve the problem
#===========================================================================
status = claw.run()
#===========================================================================
# Plot results
#===========================================================================
if iplot: plot.interactive_plot(outdir=outdir,format=claw.output_format)
if htmlplot: plot.html_plot(outdir=outdir,format=claw.output_format)
def acoustics(kernel_language='Python',
petscPlot=False,
iplot=False,
htmlplot=False,
myplot=False,
outdir='./_output',
sclaw=False,
**kwargs):
import numpy as np
"""
1D acoustics example.
"""
from petclaw.patch import Patch
from petclaw.patch import Dimension
from pyclaw.solution import Solution
from pyclaw.controller import Controller
from petclaw import plot
petscts = kwargs.get('petscts', False)
# Initialize patch and solution
xmin = 0.0
xmax = 1.0
ncells = kwargs.get('ncells', 100)
x = Dimension('x', xmin, xmax, ncells, bc_lower=2, bc_upper=2)
patch = Patch(x)
rho = 1.0
bulk = 1.0
patch.problem_data['rho'] = rho
patch.problem_data['bulk'] = bulk
patch.problem_data['zz'] = np.sqrt(rho * bulk)
patch.problem_data['cc'] = np.sqrt(rho / bulk)
from classic1 import cparam
for key, value in patch.problem_data.iteritems():
setattr(cparam, key, value)
patch.num_eqn = 2
if sclaw:
patch.num_ghost = 3
patch.t = 0.0
# init_q_petsc_structures must be called
# before patch.x.centers and such can be accessed.
patch.init_q_petsc_structures()
xc = patch.x.centers
q = np.zeros([patch.num_eqn, len(xc)], order='F')
beta = 100
gamma = 0
x0 = 0.75
q[0, :] = np.exp(-beta * (xc - x0)**2) * np.cos(gamma * (xc - x0))
q[1, :] = 0.
patch.q = q
init_solution = Solution(patch)
if sclaw:
from petclaw.sharpclaw import SharpClawSolver1D
solver = SharpClawSolver1D()
solver.lim_type = kwargs.get('lim_type', 2)
solver.time_integrator = 'SSP33'
solver.num_waves = 2
solver.char_decomp = 0
else:
from petclaw.clawpack import ClawSolver1D
solver = ClawSolver1D()
solver.num_waves = 2
from pyclaw import limiters
solver.mthlim = limiters.tvd.MC
if kernel_language == 'Python': solver.set_riemann_solver('acoustics')
solver.dt = patch.delta[0] / patch.problem_data['cc'] * 0.1
solver.kernel_language = kernel_language
claw = Controller()
#claw.output_style = 3
claw.keep_copy = True
claw.num_output_times = 5
# The output format MUST be set to petsc!
claw.output_format = 'ascii'
claw.outdir = outdir
claw.tfinal = 1.0
claw.solution = init_solution
claw.solver = solver
# Solve
if petscts:
# Uses PETSc time integrator. Needs sclaw=1 to get semidiscrete form. Examples:
# ./acoustics.py sclaw=1 petscts=1 -ts_monitor_solution -ts_type theta -snes_mf_operator -ts_dt 0.02 -ts_max_steps 50 -ts_max_time 10 -draw_pause 0.05
# ./acoustics.py sclaw=1 petscts=1 -ts_monitor_solution -ts_type ssp -ts_ssp_type rk104
x = patch.q_da.createGlobalVector()
f = x.duplicate()
ts = PETSc.TS().create(PETSc.COMM_WORLD)
ts.setDM(patch.q_da)
ts.setProblemType(PETSc.TS.ProblemType.NONLINEAR)
ts.setType(ts.Type.THETA)
eqn = AcousticEquation(patch, solver)
x[:] = patch.q
ts.setRHSFunction(eqn.rhsfunction, f)
ts.setIFunction(eqn.ifunction, f)
ts.setIJacobian(eqn.ijacobian, patch.q_da.createMat())
ts.setTimeStep(solver.dt)
ts.setDuration(claw.tfinal, max_steps=1000)
ts.setFromOptions()
q0 = x[:].copy()
ts.solve(x)
qfinal = x[:].copy()
dx = patch.delta[0]
else:
status = claw.run()
#This test is set up so that the waves pass through the domain
#exactly once, and the final solution should be equal to the
#initial condition. Here we output the 1-norm of their difference.
q0 = claw.frames[0].patch.gqVec.getArray().reshape([-1])
qfinal = claw.frames[
claw.num_output_times].patch.gqVec.getArray().reshape([-1])
dx = claw.frames[0].patch.delta[0]
if htmlplot: plot.html_plot(outdir=outdir, format=output_format)
if iplot: plot.interactive_plot(outdir=outdir, format=output_format)
if petscPlot: plot.plotPetsc(output_object)
print 'Max error:', np.max(qfinal - q0)
return dx * np.sum(np.abs(qfinal - q0))
def acoustics(kernel_language='Python',petscPlot=False,iplot=False,htmlplot=False,myplot=False,outdir='./_output',sclaw=False,**kwargs):
import numpy as np
"""
1D acoustics example.
"""
from petclaw.patch import Patch
from petclaw.patch import Dimension
from pyclaw.solution import Solution
from pyclaw.controller import Controller
from petclaw import plot
petscts = kwargs.get('petscts', False)
# Initialize patch and solution
xmin = 0.0
xmax = 1.0
ncells = kwargs.get('ncells',100)
x = Dimension('x',xmin,xmax,ncells,bc_lower=2,bc_upper=2)
patch = Patch(x)
rho = 1.0
bulk = 1.0
patch.problem_data['rho']=rho
patch.problem_data['bulk']=bulk
patch.problem_data['zz']=np.sqrt(rho*bulk)
patch.problem_data['cc']=np.sqrt(rho/bulk)
from classic1 import cparam
for key,value in patch.problem_data.iteritems(): setattr(cparam,key,value)
patch.num_eqn=2
if sclaw:
patch.num_ghost=3
patch.t = 0.0
# init_q_petsc_structures must be called
# before patch.x.centers and such can be accessed.
patch.init_q_petsc_structures()
xc=patch.x.centers
q=np.zeros([patch.num_eqn,len(xc)], order = 'F')
beta=100; gamma=0; x0=0.75
q[0,:] = np.exp(-beta * (xc-x0)**2) * np.cos(gamma * (xc - x0))
q[1,:]=0.
patch.q=q
init_solution = Solution(patch)
if sclaw:
from petclaw.sharpclaw import SharpClawSolver1D
solver = SharpClawSolver1D()
solver.lim_type = kwargs.get('lim_type',2)
solver.time_integrator = 'SSP33'
solver.num_waves=2
solver.char_decomp=0
else:
from petclaw.clawpack import ClawSolver1D
solver = ClawSolver1D()
solver.num_waves=2
from pyclaw import limiters
solver.mthlim = limiters.tvd.MC
if kernel_language=='Python': solver.set_riemann_solver('acoustics')
solver.dt=patch.delta[0]/patch.problem_data['cc']*0.1
solver.kernel_language=kernel_language
claw = Controller()
#claw.output_style = 3
claw.keep_copy = True
claw.num_output_times = 5
# The output format MUST be set to petsc!
claw.output_format = 'ascii'
claw.outdir = outdir
claw.tfinal = 1.0
claw.solution = init_solution
claw.solver = solver
# Solve
if petscts:
# Uses PETSc time integrator. Needs sclaw=1 to get semidiscrete form. Examples:
# ./acoustics.py sclaw=1 petscts=1 -ts_monitor_solution -ts_type theta -snes_mf_operator -ts_dt 0.02 -ts_max_steps 50 -ts_max_time 10 -draw_pause 0.05
# ./acoustics.py sclaw=1 petscts=1 -ts_monitor_solution -ts_type ssp -ts_ssp_type rk104
x = patch.q_da.createGlobalVector()
f = x.duplicate()
ts = PETSc.TS().create(PETSc.COMM_WORLD)
ts.setDM(patch.q_da)
ts.setProblemType(PETSc.TS.ProblemType.NONLINEAR)
ts.setType(ts.Type.THETA)
eqn = AcousticEquation(patch, solver)
x[:] = patch.q
ts.setRHSFunction(eqn.rhsfunction, f)
ts.setIFunction(eqn.ifunction, f)
ts.setIJacobian(eqn.ijacobian, patch.q_da.createMat())
ts.setTimeStep(solver.dt)
ts.setDuration(claw.tfinal, max_steps=1000)
ts.setFromOptions()
q0 = x[:].copy()
ts.solve(x)
qfinal = x[:].copy()
dx = patch.delta[0]
else:
status = claw.run()
#This test is set up so that the waves pass through the domain
#exactly once, and the final solution should be equal to the
#initial condition. Here we output the 1-norm of their difference.
q0=claw.frames[0].patch.gqVec.getArray().reshape([-1])
qfinal=claw.frames[claw.num_output_times].patch.gqVec.getArray().reshape([-1])
dx=claw.frames[0].patch.delta[0]
if htmlplot: plot.html_plot(outdir=outdir,format=output_format)
if iplot: plot.interactive_plot(outdir=outdir,format=output_format)
if petscPlot: plot.plotPetsc(output_object)
print 'Max error:', np.max(qfinal - q0)
return dx*np.sum(np.abs(qfinal-q0))
def shallow2D(use_petsc=False, iplot=0, htmlplot=False, outdir="./_output", solver_type="classic"):
# ===========================================================================
# Import libraries
# ===========================================================================
import numpy as np
if use_petsc:
import petclaw as pyclaw
else:
import pyclaw
# ===========================================================================
# Setup solver and solver parameters
# ===========================================================================
if solver_type == "classic":
solver = pyclaw.ClawSolver2D()
elif solver_type == "sharpclaw":
solver = pyclaw.SharpClawSolver2D()
solver.mwaves = 3
solver.limiters = pyclaw.limiters.tvd.MC
solver.bc_lower[0] = pyclaw.BC.outflow
solver.bc_upper[0] = pyclaw.BC.reflecting
solver.bc_lower[1] = pyclaw.BC.outflow
solver.bc_upper[1] = pyclaw.BC.reflecting
solver.dim_split = 1
# ===========================================================================
# Initialize grid and state, then initialize the solution associated to the
# state and finally initialize aux array
# ===========================================================================
# Grid:
xlower = -2.5
xupper = 2.5
mx = 150
ylower = -2.5
yupper = 2.5
my = 150
x = pyclaw.Dimension("x", xlower, xupper, mx)
y = pyclaw.Dimension("y", ylower, yupper, my)
grid = pyclaw.Grid([x, y])
meqn = 3 # Number of equations
state = pyclaw.State(grid, meqn)
grav = 1.0 # Parameter (global auxiliary variable)
state.aux_global["grav"] = grav
# Initial solution
# ================
# Riemann states of the dam break problem
damRadius = 0.5
hl = 2.0
ul = 0.0
vl = 0.0
hr = 1.0
ur = 0.0
vr = 0.0
qinit(state, hl, ul, vl, hr, ur, vl, damRadius) # This function is defined above
# ===========================================================================
# Set up controller and controller parameters
# ===========================================================================
claw = pyclaw.Controller()
claw.tfinal = 2.5
claw.solution = pyclaw.Solution(state)
claw.solver = solver
claw.outdir = outdir
claw.nout = 10
# ===========================================================================
# Solve the problem
# ===========================================================================
status = claw.run()
# ===========================================================================
# Plot results
# ===========================================================================
if iplot:
plot.interactive_plot(outdir=outdir, format=claw.output_format)
if htmlplot:
plot.html_plot(outdir=outdir, format=claw.output_format)
def shallow2D(use_petsc=False,
iplot=0,
htmlplot=False,
outdir='./_output',
solver_type='classic'):
#===========================================================================
# Import libraries
#===========================================================================
import numpy as np
if use_petsc:
import petclaw as pyclaw
else:
import pyclaw
#===========================================================================
# Setup solver and solver parameters
#===========================================================================
if solver_type == 'classic':
solver = pyclaw.ClawSolver2D()
elif solver_type == 'sharpclaw':
solver = pyclaw.SharpClawSolver2D()
solver.num_waves = 3
solver.limiters = pyclaw.limiters.tvd.MC
solver.bc_lower[0] = pyclaw.BC.extrap
solver.bc_upper[0] = pyclaw.BC.wall
solver.bc_lower[1] = pyclaw.BC.extrap
solver.bc_upper[1] = pyclaw.BC.wall
solver.dimensional_split = 1
#===========================================================================
# Initialize domain and state, then initialize the solution associated to the
# state and finally initialize aux array
#===========================================================================
# Domain:
xlower = -2.5
xupper = 2.5
mx = 150
ylower = -2.5
yupper = 2.5
my = 150
x = pyclaw.Dimension('x', xlower, xupper, mx)
y = pyclaw.Dimension('y', ylower, yupper, my)
domain = pyclaw.Domain([x, y])
num_eqn = 3 # Number of equations
state = pyclaw.State(domain, num_eqn)
grav = 1.0 # Parameter (global auxiliary variable)
state.problem_data['grav'] = grav
# Initial solution
# ================
# Riemann states of the dam break problem
damRadius = 0.5
hl = 2.
ul = 0.
vl = 0.
hr = 1.
ur = 0.
vr = 0.
qinit(state, hl, ul, vl, hr, ur, vl,
damRadius) # This function is defined above
#===========================================================================
# Set up controller and controller parameters
#===========================================================================
claw = pyclaw.Controller()
claw.tfinal = 2.5
claw.solution = pyclaw.Solution(state, domain)
claw.solver = solver
claw.outdir = outdir
claw.num_output_times = 10
#===========================================================================
# Solve the problem
#===========================================================================
status = claw.run()
#===========================================================================
# Plot results
#===========================================================================
if iplot: plot.interactive_plot(outdir=outdir, format=claw.output_format)
if htmlplot: plot.html_plot(outdir=outdir, format=claw.output_format)
