esys.escript.pdetools Package¶
Classes¶
- class esys.escript.pdetools.ArithmeticTuple(*args)¶
Tuple supporting inplace update x+=y and scaling x=a*y where
x,y
is an ArithmeticTuple anda
is a float.Example of usage:
from esys.escript import Data from numpy import array a=eData(...) b=array([1.,4.]) x=ArithmeticTuple(a,b) y=5.*x
- __init__(*args)¶
Initializes object with elements
args
.- Parameters:
args – tuple of objects that support inplace add (x+=y) and scaling (x=a*y)
- class esys.escript.pdetools.CorrectionFailed¶
Exception thrown if no convergence has been achieved in the solution correction scheme.
- __init__(*args, **kwargs)¶
- class esys.escript.pdetools.Defect¶
Defines a non-linear defect F(x) of a variable x. This class includes two functions (bilinearform and eval) that must be overridden by subclassing before use.
- __init__()¶
Initializes defect.
- bilinearform(x0, x1)¶
Returns the inner product of x0 and x1
NOTE: MUST BE OVERRIDDEN BY A SUBCLASS
- Parameters:
x0 – value for x0
x1 – value for x1
- Returns:
the inner product of x0 and x1
- Return type:
float
- derivative(F0, x0, v, v_is_normalised=True)¶
Returns the directional derivative at
x0
in the direction ofv
.- Parameters:
F0 – value of this defect at x0
x0 – value at which derivative is calculated
v – direction
v_is_normalised – True to indicate that
v
is nomalized (self.norm(v)=0)
- Returns:
derivative of this defect at x0 in the direction of
v
- Note:
by default numerical evaluation (self.eval(x0+eps*v)-F0)/eps is used but this method maybe overwritten to use exact evaluation.
- eval(x)¶
Returns the value F of a given
x
.NOTE: MUST BE OVERRIDDEN BY A SUBCLASS
- Parameters:
x – value for which the defect
F
is evaluated- Returns:
value of the defect at
x
- getDerivativeIncrementLength()¶
Returns the relative increment length used to approximate the derivative of the defect. :return: value of the defect at
x
:rtype: positivefloat
- norm(x)¶
Returns the norm of argument
x
.- Parameters:
x – a value
- Returns:
norm of argument x
- Return type:
float
- Note:
by default
sqrt(self.bilinearform(x,x)
is returned.
- setDerivativeIncrementLength(inc=1.4901161193847656e-05)¶
Sets the relative length of the increment used to approximate the derivative of the defect. The increment is inc*norm(x)/norm(v)*v in the direction of v with x as a starting point.
- Parameters:
inc (positive
float
) – relative increment length
- class esys.escript.pdetools.HomogeneousSaddlePointProblem(**kwargs)¶
This class provides a framework for solving linear homogeneous saddle point problems of the form:
*Av+B^*p=f* *Bv =0*
for the unknowns v and p and given operators A and B and given right hand side f. B^* is the adjoint operator of B. A may depend weakly on v and p.
- __init__(**kwargs)¶
initializes the saddle point problem
- Bv(v, tol)¶
Returns Bv with accuracy
tol
(overwrite)- Return type:
equal to the type of p
- Note:
boundary conditions on p should be zero!
- getAbsoluteTolerance()¶
Returns the absolute tolerance.
- Returns:
absolute tolerance
- Return type:
float
- getDV(p, v, tol)¶
return a correction to the value for a given v and a given p with accuracy
tol
(overwrite)- Parameters:
p – pressure
v – pressure
- Returns:
dv given as dv= A^{-1} (f-A v-B^*p)
- Note:
Only A may depend on v and p
- getTolerance()¶
Returns the relative tolerance.
- Returns:
relative tolerance
- Return type:
float
- inner_p(p0, p1)¶
Returns inner product of p0 and p1 (overwrite).
- Parameters:
p0 – a pressure
p1 – a pressure
- Returns:
inner product of p0 and p1
- Return type:
float
- inner_pBv(p, Bv)¶
Returns inner product of element p and Bv (overwrite).
- Parameters:
p – a pressure increment
Bv – a residual
- Returns:
inner product of element p and Bv
- Return type:
float
- Note:
used if PCG is applied.
- norm_Bv(Bv)¶
Returns the norm of Bv (overwrite).
- Return type:
equal to the type of p
- Note:
boundary conditions on p should be zero!
- norm_p(p)¶
calculates the norm of
p
- Parameters:
p – a pressure
- Returns:
the norm of
p
using the inner product for pressure- Return type:
float
- norm_v(v)¶
Returns the norm of v (overwrite).
- Parameters:
v – a velovity
- Returns:
norm of v
- Return type:
non-negative
float
- resetControlParameters(K_p=1.0, K_v=1.0, rtol_max=0.01, rtol_min=1e-07, chi_max=0.5, reduction_factor=0.3, theta=0.1)¶
sets a control parameter
- Parameters:
K_p (
float
) – initial value for constant to adjust pressure toleranceK_v (
float
) – initial value for constant to adjust velocity tolerancertol_max (
float
) – maximuim relative tolerance used to calculate presssure and velocity increment.chi_max (
float
) – maximum tolerable converegence rate.reduction_factor (
float
) – reduction factor for adjustment factors.
- setAbsoluteTolerance(tolerance=0.0)¶
Sets the absolute tolerance.
- Parameters:
tolerance (non-negative
float
) – tolerance to be used
- setControlParameter(K_p=None, K_v=None, rtol_max=None, rtol_min=None, chi_max=None, reduction_factor=None, theta=None)¶
sets a control parameter
- Parameters:
K_p (
float
) – initial value for constant to adjust pressure toleranceK_v (
float
) – initial value for constant to adjust velocity tolerancertol_max (
float
) – maximuim relative tolerance used to calculate presssure and velocity increment.chi_max (
float
) – maximum tolerable converegence rate.
- setTolerance(tolerance=0.0001)¶
Sets the relative tolerance for (v,p).
- Parameters:
tolerance (non-negative
float
) – tolerance to be used
- solve(v, p, max_iter=20, verbose=False, usePCG=True, iter_restart=20, max_correction_steps=10)¶
Solves the saddle point problem using initial guesses v and p.
- Parameters:
v (
Data
) – initial guess for velocityp (
Data
) – initial guess for pressureusePCG (
bool
) – indicates the usage of the PCG rather than GMRES scheme.max_iter (
int
) – maximum number of iteration steps per correction attemptverbose (
bool
) – if True, shows information on the progress of the saddlepoint problem solver.iter_restart (
int
) – restart the iteration afteriter_restart
steps (only used if useUzaw=False)
- Return type:
tuple
ofData
objects- Note:
typically this method is overwritten by a subclass. It provides a wrapper for the
_solve
method.
- solve_AinvBt(dp, tol)¶
Solves A dv=B^*dp with accuracy
tol
- Parameters:
dp – a pressure increment
- Returns:
the solution of A dv=B^*dp
- Note:
boundary conditions on dv should be zero! A is the operator used in
getDV
and must not be altered.
- solve_prec(Bv, tol)¶
Provides a preconditioner for (BA^{-1}B^ * ) applied to Bv with accuracy
tol
- Return type:
equal to the type of p
- Note:
boundary conditions on p should be zero!
- class esys.escript.pdetools.IndefinitePreconditioner¶
Exception thrown if the preconditioner is not positive definite.
- __init__(*args, **kwargs)¶
- class esys.escript.pdetools.IterationBreakDown¶
Exception thrown if the iteration scheme encountered an incurable breakdown.
- __init__(*args, **kwargs)¶
- class esys.escript.pdetools.Locator(where, x=array([0., 0., 0.]))¶
Locator provides access to the values of data objects at a given spatial coordinate x.
In fact, a Locator object finds the sample in the set of samples of a given function space or domain which is closest to the given point x.
- __init__(where, x=array([0., 0., 0.]))¶
Initializes a Locator to access values in Data objects on the Doamin or FunctionSpace for the sample point which is closest to the given point x.
- Parameters:
where (
escript.FunctionSpace
) – function spacex (
numpy.ndarray
orlist
ofnumpy.ndarray
) – location(s) of the Locator
- getFunctionSpace()¶
Returns the function space of the Locator.
- getId(item=None)¶
Returns the identifier of the location.
- getValue(data)¶
Returns the value of
data
at the Locator ifdata
is aData
object otherwise the object is returned.
- getX()¶
Returns the exact coordinates of the Locator.
- setValue(data, v)¶
Sets the value of the
data
at the Locator.
- class esys.escript.pdetools.MaxIterReached¶
Exception thrown if the maximum number of iteration steps is reached.
- __init__(*args, **kwargs)¶
- class esys.escript.pdetools.NegativeNorm¶
Exception thrown if a norm calculation returns a negative norm.
- __init__(*args, **kwargs)¶
- class esys.escript.pdetools.NoPDE(domain, D=None, Y=None, q=None, r=None)¶
Solves the following problem for u:
kronecker[i,j]*D[j]*u[j]=Y[i]
with constraint
u[j]=r[j] where q[j]>0
where D, Y, r and q are given functions of rank 1.
In the case of scalars this takes the form
D*u=Y
with constraint
u=r where q>0
where D, Y, r and q are given scalar functions.
The constraint overwrites any other condition.
- Note:
This class is similar to the
linearPDEs.LinearPDE
class with A=B=C=X=0 but has the intention that all input parameters are given inSolution
orReducedSolution
.
- __init__(domain, D=None, Y=None, q=None, r=None)¶
Initializes the problem.
- Parameters:
domain (
Domain
) – domain of the PDED (
float
,int
,numpy.ndarray
,Data
) – coefficient of the solutionY (
float
,int
,numpy.ndarray
,Data
) – right hand sideq (
float
,int
,numpy.ndarray
,Data
) – location of constraintsr (
float
,int
,numpy.ndarray
,Data
) – value of solution at locations of constraints
- getSolution()¶
Returns the solution.
- Returns:
the solution of the problem
- Return type:
Data
object in theFunctionSpace
Solution
orReducedSolution
- setReducedOff()¶
Sets the
FunctionSpace
of the solution toSolution
.
- setReducedOn()¶
Sets the
FunctionSpace
of the solution toReducedSolution
.
- setValue(D=None, Y=None, q=None, r=None)¶
Assigns values to the parameters.
- Parameters:
D (
float
,int
,numpy.ndarray
,Data
) – coefficient of the solutionY (
float
,int
,numpy.ndarray
,Data
) – right hand sideq (
float
,int
,numpy.ndarray
,Data
) – location of constraintsr (
float
,int
,numpy.ndarray
,Data
) – value of solution at locations of constraints
- class esys.escript.pdetools.Projector(domain, reduce=True, fast=True)¶
The Projector is a factory which projects a discontinuous function onto a continuous function on a given domain.
- __init__(domain, reduce=True, fast=True)¶
Creates a continuous function space projector for a domain.
- Parameters:
domain – Domain of the projection.
reduce – Flag to reduce projection order
fast – Flag to use a fast method based on matrix lumping
- getSolverOptions()¶
Returns the solver options of the PDE solver.
- Return type:
linearPDEs.SolverOptions
- getValue(input_data)¶
Projects
input_data
onto a continuous function.- Parameters:
input_data – the data to be projected
- class esys.escript.pdetools.SolverSchemeException¶
This is a generic exception thrown by solvers.
- __init__(*args, **kwargs)¶
- class esys.escript.pdetools.TimeIntegrationManager(*inital_values, **kwargs)¶
A simple mechanism to manage time dependend values.
Typical usage is:
dt=0.1 # time increment tm=TimeIntegrationManager(inital_value,p=1) while t<1. v_guess=tm.extrapolate(dt) # extrapolate to t+dt v=... tm.checkin(dt,v) t+=dt
- Note:
currently only p=1 is supported.
- __init__(*inital_values, **kwargs)¶
Sets up the value manager where
inital_values
are the initial values and p is the order used for extrapolation.
- checkin(dt, *values)¶
Adds new values to the manager. The p+1 last values are lost.
- extrapolate(dt)¶
Extrapolates to
dt
forward in time.
- getTime()¶
- getValue()¶
Functions¶
- esys.escript.pdetools.BoundaryValuesFromVolumeTag(domain, **values)¶
Creates a mask on the Solution(domain) function space where the value is one for samples that touch regions tagged by tags.
Usage: m=BoundaryValuesFromVolumeTag(domain, ham=1, f=6)
- Parameters:
domain (
escript.Domain
) – domain to be used- Returns:
a mask which marks samples that are touching the boundary tagged by any of the given tags
- Return type:
escript.Data
of rank 0
- esys.escript.pdetools.GMRES(r, Aprod, x, bilinearform, atol=0, rtol=1e-08, iter_max=100, iter_restart=20, verbose=False, P_R=None)¶
Solver for
Ax=b
with a general operator A (more details required!). It uses the generalized minimum residual method (GMRES).
The iteration is terminated if
|r| <= atol+rtol*|r0|
where r0 is the initial residual and |.| is the energy norm. In fact
|r| = sqrt( bilinearform(r,r))
- Parameters:
r (any object supporting inplace add (x+=y) and scaling (x=scalar*y)) – initial residual r=b-Ax.
r
is altered.x (same like
r
) – an initial guess for the solutionAprod (function
Aprod(x)
wherex
is of the same object like argumentx
. The returned object needs to be of the same type like argumentr
.) – returns the value Axbilinearform (function
bilinearform(x,r)
wherex
is of the same type like argumentx
andr
. The returned value is afloat
.) – inner product<x,r>
atol (non-negative
float
) – absolute tolerancertol (non-negative
float
) – relative toleranceiter_max (
int
) – maximum number of iteration stepsiter_restart (
int
) – in order to save memory the orthogonalization process is terminated afteriter_restart
steps and the iteration is restarted.
- Returns:
the solution approximation and the corresponding residual
- Return type:
tuple
- Warning:
r
andx
are altered.
- esys.escript.pdetools.MINRES(r, Aprod, x, Msolve, bilinearform, atol=0, rtol=1e-08, iter_max=100)¶
Solver for
Ax=b
with a symmetric and positive definite operator A (more details required!). It uses the minimum residual method (MINRES) with preconditioner M providing an approximation of A.
The iteration is terminated if
|r| <= atol+rtol*|r0|
where r0 is the initial residual and |.| is the energy norm. In fact
|r| = sqrt( bilinearform(Msolve(r),r))
For details on the preconditioned conjugate gradient method see the book:
“Templates for the Solution of Linear Systems by R. Barrett, M. Berry, T.F. Chan, J. Demmel, J. Donato, J. Dongarra, V. Eijkhout, R. Pozo, C. Romine, and H. van der Vorst”.
- Parameters:
r (any object supporting inplace add (x+=y) and scaling (x=scalar*y)) – initial residual r=b-Ax.
r
is altered.x (any object supporting inplace add (x+=y) and scaling (x=scalar*y)) – an initial guess for the solution
Aprod (function
Aprod(x)
wherex
is of the same object like argumentx
. The returned object needs to be of the same type like argumentr
.) – returns the value AxMsolve (function
Msolve(r)
wherer
is of the same type like argumentr
. The returned object needs to be of the same type like argumentx
.) – solves Mx=rbilinearform (function
bilinearform(x,r)
wherex
is of the same type like argumentx
andr
is. The returned value is afloat
.) – inner product<x,r>
atol (non-negative
float
) – absolute tolerancertol (non-negative
float
) – relative toleranceiter_max (
int
) – maximum number of iteration steps
- Returns:
the solution approximation and the corresponding residual
- Return type:
tuple
- Warning:
r
andx
are altered.
- esys.escript.pdetools.MaskFromBoundaryTag(domain, *tags)¶
Creates a mask on the Solution(domain) function space where the value is one for samples that touch the boundary tagged by tags.
Usage: m=MaskFromBoundaryTag(domain, “left”, “right”)
- Parameters:
domain (
escript.Domain
) – domain to be usedtags (
str
) – boundary tags
- Returns:
a mask which marks samples that are touching the boundary tagged by any of the given tags
- Return type:
escript.Data
of rank 0
- esys.escript.pdetools.MaskFromTag(domain, *tags)¶
Creates a mask on the Solution(domain) function space where the value is one for samples that touch regions tagged by tags.
Usage: m=MaskFromTag(domain, “ham”)
- Parameters:
domain (
escript.Domain
) – domain to be usedtags (
str
) – boundary tags
- Returns:
a mask which marks samples that are touching the boundary tagged by any of the given tags
- Return type:
escript.Data
of rank 0
- esys.escript.pdetools.NewtonGMRES(defect, x, iter_max=100, sub_iter_max=20, atol=0, rtol=0.0001, subtol_max=0.5, gamma=0.9, verbose=False)¶
Solves a non-linear problem F(x)=0 for unknown x using the stopping criterion:
norm(F(x) <= atol + rtol * norm(F(x0)
where x0 is the initial guess.
- Parameters:
defect (
Defect
) – object defining the function F.defect.norm
defines the norm used in the stopping criterion.x (any object type allowing basic operations such as
numpy.ndarray
,Data
) – initial guess for the solution,x
is altered.iter_max (positive
int
) – maximum number of iteration stepssub_iter_max (positive
int
) – maximum number of inner iteration stepsatol (positive
float
) – absolute tolerance for the solutionrtol (positive
float
) – relative tolerance for the solutiongamma (positive
float
, less than 1) – tolerance safety factor for inner iterationsubtol_max (positive
float
, less than 1) – upper bound for inner tolerance
- Returns:
an approximation of the solution with the desired accuracy
- Return type:
same type as the initial guess
- esys.escript.pdetools.PCG(r, Aprod, x, Msolve, bilinearform, atol=0, rtol=1e-08, iter_max=100, initial_guess=True, verbose=False)¶
Solver for
Ax=b
with a symmetric and positive definite operator A (more details required!). It uses the conjugate gradient method with preconditioner M providing an approximation of A.
The iteration is terminated if
|r| <= atol+rtol*|r0|
where r0 is the initial residual and |.| is the energy norm. In fact
|r| = sqrt( bilinearform(Msolve(r),r))
For details on the preconditioned conjugate gradient method see the book:
“Templates for the Solution of Linear Systems by R. Barrett, M. Berry, T.F. Chan, J. Demmel, J. Donato, J. Dongarra, V. Eijkhout, R. Pozo, C. Romine, and H. van der Vorst”.
- Parameters:
r (any object supporting inplace add (x+=y) and scaling (x=scalar*y)) – initial residual r=b-Ax.
r
is altered.x (any object supporting inplace add (x+=y) and scaling (x=scalar*y)) – an initial guess for the solution
Aprod (function
Aprod(x)
wherex
is of the same object like argumentx
. The returned object needs to be of the same type like argumentr
.) – returns the value AxMsolve (function
Msolve(r)
wherer
is of the same type like argumentr
. The returned object needs to be of the same type like argumentx
.) – solves Mx=rbilinearform (function
bilinearform(x,r)
wherex
is of the same type like argumentx
andr
is. The returned value is afloat
.) – inner product<x,r>
atol (non-negative
float
) – absolute tolerancertol (non-negative
float
) – relative toleranceiter_max (
int
) – maximum number of iteration steps
- Returns:
the solution approximation and the corresponding residual
- Return type:
tuple
- Warning:
r
andx
are altered.
- esys.escript.pdetools.TFQMR(r, Aprod, x, bilinearform, atol=0, rtol=1e-08, iter_max=100)¶
Solver for
Ax=b
with a general operator A (more details required!). It uses the Transpose-Free Quasi-Minimal Residual method (TFQMR).
The iteration is terminated if
|r| <= atol+rtol*|r0|
where r0 is the initial residual and |.| is the energy norm. In fact
|r| = sqrt( bilinearform(r,r))
- Parameters:
r (any object supporting inplace add (x+=y) and scaling (x=scalar*y)) – initial residual r=b-Ax.
r
is altered.x (same like
r
) – an initial guess for the solutionAprod (function
Aprod(x)
wherex
is of the same object like argumentx
. The returned object needs to be of the same type like argumentr
.) – returns the value Axbilinearform (function
bilinearform(x,r)
wherex
is of the same type like argumentx
andr
. The returned value is afloat
.) – inner product<x,r>
atol (non-negative
float
) – absolute tolerancertol (non-negative
float
) – relative toleranceiter_max (
int
) – maximum number of iteration steps
- Return type:
tuple
- Warning:
r
andx
are altered.
- esys.escript.pdetools.getInfLocator(arg)¶
Return a Locator for a point with the inf value over all arg.
- esys.escript.pdetools.getSupLocator(arg)¶
Return a Locator for a point with the sup value over all arg.