#
##
## This file is part of pyFormex 2.0 (Mon Sep 14 12:29:05 CEST 2020)
## pyFormex is a tool for generating, manipulating and transforming 3D
## geometrical models by sequences of mathematical operations.
## Home page: http://pyformex.org
## Project page: http://savannah.nongnu.org/projects/pyformex/
## Copyright 2004-2020 (C) Benedict Verhegghe (benedict.verhegghe@ugent.be)
## Distributed under the GNU General Public License version 3 or later.
##
## This program is free software: you can redistribute it and/or modify
## it under the terms of the GNU General Public License as published by
## the Free Software Foundation, either version 3 of the License, or
## (at your option) any later version.
##
## This program is distributed in the hope that it will be useful,
## but WITHOUT ANY WARRANTY; without even the implied warranty of
## MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
## GNU General Public License for more details.
##
## You should have received a copy of the GNU General Public License
## along with this program. If not, see http://www.gnu.org/licenses/.
##
"""A class for holding results from Finite Element simulations.
"""
import re
from collections import OrderedDict
import numpy as np
import pyformex as pf
import pyformex.arraytools as at
from pyformex.script import export
[docs]class FeResult(object):
"""Finite Element Results Database.
This class can hold a collection of results from a Finite Element
simulation. While the class was designed for the post-processing
of Abaqus (tm) results, it can be used more generally to store
results from any program performing simulations over a mesh.
pyFormex comes with an included program `postabq` that scans an
Abaqus .fil output file and translates it into a pyFormex script.
Use it as follows::
postabq job.fil > job.py
Then execute the created script `job.py` from inside pyFormex. This
will create an FeResult instance with all the recognized results.
The structure of the FeResult class very closely follows that
of the Abaqus results database. There are some attributes
with general info and with the geometry (mesh) of the domain.
The simulation results are divided in 'steps' and inside each step
in 'increments'. Increments are usually connected to incremental time
and so are often the steps, though it is up to the user to interprete
the time. Steps could just as well be different unrelated simulations
performed over the same geometry.
In each step/increment result block, individual values can be accessed
by result codes. The naming mostly follows the result codes in Abaqus,
but components of vector/tensor values are number starting from 0, as
in Python and pyFormex.
Result codes:
- `U`: displacement vector
- `U0`, `U1`, `U2` : x, y, resp. z-component of displacement
- `S`: stress tensor
- `S0` .. `S5`: components of the (symmetric) stress tensor:
0..2 : x, y, z normal stress
3..5 : xy, yz, zx shear stress
"""
_name_ = '__FePost__'
re_Skey = re.compile("S[0-5]")
re_Ukey = re.compile("U[0-2]")
def __init__(self, name=_name_, datasize={'U': 3, 'S': 6, 'COORD': 3}):
self.name = name
self.datasize = datasize.copy()
self.about = {'creator': pf.Version(),
'created': pf.StartTime,
}
self.modeldone = False
self.labels = {}
self.nelems = 0
self.nnodes = 0
self.dofs = None
self.displ = None
self.nodid = None
self.nodes = None
self.elems = None
self.nset = None
self.nsetkey = None
self.eset = None
self.res = None
self.hdr = None
self.nodnr = 0
self.elnr = 0
def dataSize(self, key, data):
if key in self.datasize:
return self.datasize[key]
else:
return len(data)
def Abqver(self, version):
self.about.update({'abqver': version})
def Date(self, date, time):
self.about.update({'abqdate': date, 'abqtime': time})
def Size(self, nelems, nnodes, length):
self.nelems = nelems
self.nnodes = nnodes
self.length = length
self.nodid = -np.ones((nnodes,), dtype=np.int32)
self.nodes = np.zeros((nnodes, 3), dtype=np.float32)
self.elems = {}
self.nset = {}
self.eset = {}
def Dofs(self, data):
self.dofs = np.array(data)
self.displ = self.dofs[self.dofs[:6] > 0]
if self.displ.max() > 3:
self.datasize['U'] = 6
def Heading(self, head):
self.about.update({'heading': head})
def Node(self, nr, coords, normal=None):
self.nodid[self.nodnr] = nr
nn = len(coords)
self.nodes[self.nodnr][:nn] = coords
self.nodnr += 1
def Element(self, nr, typ, conn):
if typ not in self.elems:
self.elems[typ] = []
self.elems[typ].append(conn)
def Nodeset(self, key, data):
self.nsetkey = key
self.nset[key] = np.asarray(data)
def NodesetAdd(self, data):
self.nset[self.nsetkey] = np.union1d(self.nset[self.nsetkey], np.asarray(data))
def Elemset(self, key, data):
self.esetkey = key
self.eset[key] = np.asarray(data)
def ElemsetAdd(self, data):
self.eset[self.esetkey] = np.union1d(self.eset[self.esetkey], np.asarray(data))
def Finalize(self):
self.nid = at.inverseUniqueIndex(self.nodid)
for k in self.elems:
v = np.asarray(self.elems[k])
self.elems[k] = np.asarray(self.nid[v])
self.modeldone = True
# we use lists, to keep the cases in order
self.res = OrderedDict()
self.step = None
self.inc = None
[docs] def Increment(self, step, inc, **kargs):
"""Add a new step/increment to the database.
This method can be used to add a new increment to an existing step,
or to add a new step and set the initial increment, or to just select
an existing step/inc combination.
If the step/inc combination is new, a new empty result record is created.
The result record of the specified step/inc becomes the current result.
"""
if not self.modeldone:
self.Finalize()
if step != self.step:
if step not in self.res:
self.res[step] = OrderedDict()
self.step = step
self.inc = None
res = self.res[self.step]
if inc != self.inc:
if inc not in res:
res[inc] = OrderedDict()
self.inc = inc
self.R = self.res[self.step][self.inc]
def EndIncrement(self):
if not self.modeldone:
self.Finalize()
self.step = self.inc = -1
def Label(self, tag, value):
self.labels[tag] = value
def NodeOutput(self, key, nodid, data):
if key not in self.R:
self.R[key] = np.zeros((self.nnodes, self.dataSize(key, data)),
dtype=np.float32)
if key == 'U':
self.R[key][nodid-1][self.displ-1] = data
elif key == 'S':
n1 = self.hdr['ndi']
# n2 = self.hdr['nshr'] # TODO: unused ??
ind = np.arange(len(data))
ind[n1:] += (3-n1)
# print(ind)
self.R[key][nodid-1][ind] = data
else:
self.R[key][nodid-1][:len(data)] = data
def ElemHeader(self, **kargs):
self.hdr = dict(**kargs)
def ElemOutput(self, key, data):
if self.hdr['loc'] == 'na':
self.NodeOutput(key, self.hdr['i'], data)
[docs] def Export(self):
"""Align on the last increment and export results"""
try:
self.step = list(self.res.keys())[-1]
self.inc = list(self.res[self.step].keys())[-1]
self.R = self.res[self.step][self.inc]
except Exception:
self.step = None
self.inc = None
self.R = None
export({self.name: self, self._name_: self})
print("Read %d nodes, %d elements" % (self.nnodes, self.nelems))
if self.res is None:
print("No results")
else:
print("Steps: %s" % list(self.res.keys()))
[docs] def do_nothing(*arg, **kargs):
"""A do nothing function to stand in for as yet undefined functions."""
pass
TotalEnergies = do_nothing
OutputRequest = do_nothing
Coordinates = do_nothing
Displacements = do_nothing
Unknown = do_nothing
[docs] def setStepInc(self, step, inc=1):
"""Set the database pointer to a given step,inc pair.
This sets the step and inc attributes to the given values, and puts
the corresponding results in the R attribute. If the step.inc pair does
not exist, an empty results dict is set.
"""
try:
self.step = step
self.inc = inc
self.R = self.res[self.step][self.inc]
except Exception:
self.R = {}
[docs] def getSteps(self):
"""Return all the step keys."""
return list(self.res.keys())
[docs] def getIncs(self, step):
"""Return all the incs for given step."""
if step in self.res:
return list(self.res[step].keys())
else:
return []
[docs] def nextStep(self):
"""Skips to the start of the next step."""
if self.step < self.getSteps()[-1]:
self.setStepInc(self.step+1)
[docs] def nextInc(self):
"""Skips to the next increment.
The next increment is either the next increment of the current step,
or the first increment of the next step.
"""
if self.inc < self.getIncs(self.step)[-1]:
self.setStepInc(self.step, self.inc+1)
else:
self.nextStep()
[docs] def prevStep(self):
"""Skips to the start of the previous step."""
if self.step > 1:
self.setStepInc(self.step-1)
[docs] def prevInc(self):
"""Skips to the previous increment.
The previous increment is either the previous increment of the current
step, or the last increment of the previous step.
"""
if self.inc > 1:
self.setStepInc(self.step, self.inc-1)
else:
if self.step > 1:
step = self.step-1
inc = self.getIncs(step)[-1]
self.setStepInc(step, inc)
[docs] def getres(self, key, domain='nodes'):
"""Return the results of the current step/inc for given key.
The key may include a component to return only a single column
of a multicolumn value.
"""
components = '012'
if self.re_Skey.match(key):
if self.datasize['S']==3:
components = '013'
else:
components = '012345'
elif self.re_Ukey.match(key):
if self.datasize['U']==2:
components = '01'
else:
components = '012'
comp = components.find(key[-1])
if comp >= 0:
key = key[:-1]
if key in self.R:
val = self.R[key]
if comp in range(val.shape[1]):
return val[:, comp]
else:
return val
else:
return None
[docs] def printSteps(self):
"""Print the steps/increments/resultcodes for which we have results."""
if self.res is not None:
for i, step in self.res.items():
for j, inc in step.items():
for k, v in inc.items():
if isinstance(v, np.ndarray):
data = "%s %s" % (v.dtype.kind, str(v.shape))
else:
data = str(v)
print("Step %s, Inc %s, Res %s (%s)" % (i, j, k, data))
def save(self, filename):
from pyformex.plugins.saveload import savePZF
savePZF(filename, _type="FeResult", **self.__dict__)
# End
