如何通过Python脚本告诉Abaqus在错误后继续？

Question

At the moment I´m programming a Python-script that I start in Abaqus CAE. It´s purpose is to generate a dataset for a neural network. Therefore the script picks a random node out of a before defined nodeset and applies a displacement and a rotation on that node. Then it tells abaqus to solve that simulation and reads the odb-file after the simulation is complete. After that the script is continuing with the next random chosen node for a before chosen amount of iterations. The script works fine code-wise, but now and then the random chosen inputs for the displacement and the rotation are too difficult for abaqus to solve, so it crashes with a "Increment-Error" ( Too many increments made for this increment....). And because of that there is a .lck file in the directory that prevents the Python-script to get access to read the odb-File and therefore also the Python-script crashes. I already tried to fix this error by editing the Increment size and the maximum number of increments, but both ideas have not worked. My latest idea was to implement a "if"-condition that tells abaqus to delete all "Job_1.*"-files and to continue with the next node and other random chosen input variables, if there exists a .lck file directly after the job is done. But also this idea has not worked. Because of that i ask you guys, if any of you has an idea how i can solve this problem. Is there any way to tell Abaqus it should continue with another node after such an error accured? Here is the Code:

import assembly  
import step  
import interaction  
import load  
import mesh
import optimization
import job
import sketch
import visualization
import xyPlot
import displayGroupOdbToolset as dgo
import connectorBehavior
from time import *
import datetime
import string
import odbAccess
from abaqus import getInput
from random import choice
from random import random
from math import pi
from math import sqrt
import os
import glob

def create_dataset(dataset_length):
    print "Start"
    global dataset
    a = mdb.models['Model-1'].rootAssembly
    n1 = a.instances['PART-1-1'].nodes 
    session.journalOptions.setValues(replayGeometry=COORDINATE, recoverGeometry=COORDINATE) 
    nodes1 =n1[157:159]+n1[190:194]+n1[198:200]+n1[201:209]+n1[239:247]+.... #Abaqus Node-Set for random pick
    nodesetlen=len(nodes1)
    i=0
    while i < dataset_length:  # while-Loop with the length of the requested dataset-length
        print "%i. Loop begins" %(i+1)
        k = choice(range(nodesetlen)) #random-index
        x = nodes1[k:k+1]  # random node out of the nodeset
        region = a.Set(nodes=x, name='BC_RH')  # Abaqus region definition (because the new node-position )
        datum = mdb.models['Model-1'].rootAssembly.datums[49] 
        mdb.models['Model-1'].DisplacementBC(name='BC_RH', createStepName='Step-1',
                                             region=region, u1=(random()*30),
                                             u2=(random()*30), u3=(random()*30),
                                             ur1=(random()*pi/4),
                                             ur2=(random()*pi/4), ur3=(random()*pi/4),
                                             amplitude=UNSET, fixed=OFF, distributionType=UNIFORM, fieldName='',
                                             localCsys=datum)  # Sets the inputs for the abaqus-Job_1
        mdb.jobs['Job_1'].submit(consistencyChecking=OFF) # submits the Job
        print "Job ist submitted"
        mdb.jobs['Job_1'].waitForCompletion() #Python waits for Abaqus until the Job is done
        print "Job is done"
        my_file = "U:/Job_1.lck"
        with open("Job_1.sta") as rfile:
            line = rfile.readlines()[-1]
        if line == " THE ANALYSIS HAS NOT BEEN COMPLETED\n":
            print "Increment Error next Loop will start"
            sleep(30) #just to be sure that Abaqus closed all files
        else:
             odb = openOdb(path='Job_1.odb') # Opens ODB-File
             # HERE ARE A LOT OF OUTPUT-CALCULATIONS IN THE REAL FILE
             # DOESNT INFLUENCE THE PROBLEM
             odb.close()
             i += 1
    print "Erfolgreich beendet"
    return dataset

dataset_length = 500
create_dataset(dataset_length)

Here is the errror in Abaqus CAE:

1. Loop begins
Recent Node-Label:  22432
Job ist submitted
Job Job_1: Analysis Input File Processor completed successfully.
Error in job Job_1: Too many attempts made for this increment
Error in job Job_1: THE ANALYSIS HAS BEEN TERMINATED DUE TO PREVIOUS ERRORS. ALL OUTPUT REQUESTS HAVE BEEN WRITTEN FOR THE LAST CONVERGED INCREMENT.
Job Job_1: Abaqus/Standard aborted due to errors.
Job is done
Increment Error, next Loop will start
1. Loop begins
Recent Node-Label:  20996
Job ist submitted
Error in job Job_1: Abaqus/Standard Analysis exited with an error - Please see the  message file for possible error messages if the file exists.
Error in job Job_1: Abaqus/Standard Analysis exited with an error - Please see the  message file for possible error messages if the file exists.
Job Job_1 aborted due to errors.
Job Job_1 aborted due to errors.
Job is done

And after that outprinting the Python-error-Window tells me that there is no .sta file.

Answer 1

Use try and except clause:

try:
   code goes here
except IncrementError:
   log error and continue

Answer 2

I'd suggest using try-except from https://docs.python.org/2/tutorial/errors.html

try: 
   # do what needs to be done here
except Increment-Error:
    print "Oh no, and Incremental Error has happened!"

Answer 3

in your edited version you still open and attempt to process the odb after detecting a failure.

You should have (note the affirmative)

   if line == " THE ANALYSIS HAS COMPLETED\n":
     odb = openOdb(path='Job_1.odb') # Opens ODB-File
      # HERE ARE A LOT OF OUTPUT-CALCULATIONS IN THE REAL FILE
      # DOESNT INFLUENCE THE PROBLEM
     odb.close()
   else:
     print "some error message"

i don't see why you need to do anything besides print an error message, the main loop will just continue and create a new case will it not?

I would also for good measure do os.remove("Job-1.sta") before submitting the job.