Get the top 10% of the results from an analysis

How to get the average value for the top 10% ranked output value for a predefined Named Selection? In post processing, it would take only the 10% top ranked output value (e.g. NLPLWK) in a predefined Named Selection (Mesh,Nodes,Face) and plot it. Then it would calculate the average of those 10% shorted values with their corresponding node without including the max value as it may be an outlier.

Find more posts tagged with

AnsysACT

Mechanical

Comments

Kleanthis

name = 'Udres'
udrExpressionName = 'NLELWK'

try:
    DataModel.GetObjectsByName(name)[0].Delete()
except:
    pass


ns ='Selection'
# store link to solution
solu=ExtAPI.DataModel.Project.Model.Analyses[0].Solution
# add UDR
newUDR = solu.AddUserDefinedResult() 
newUDR.Name = name
newUDR.ScopingMethod = GeometryDefineByType.Component
newUDR.Location = DataModel.GetObjectsByName(ns)[0]
newUDR.Expression= udrExpressionName

newUDR.EvaluateAllResults()

udr = ExtAPI.DataModel.GetObjectsByName(name)[0] # get the result 
nodes = udr.PlotData['Node'] # get the nodal scope for the result
results = [value for value in udr.PlotData['Values']] # get the nodal results 


for result,node in zip(results,nodes):
        nomax.append([result,node])


shortList = [[i[0],i[1]] for i  in (nomax)]
#plotNodes = [str(i[0]) for i in shortList][1:-1]

res=sorted(shortList)

myres=res[(len(res)-int(0.1*len(res))):(len(res)-1)]

a = [i[0] for i in myres ]   
b=sum(a)
c= len(a)
average = b/c
print (average)

plotNodes = ''
for i in myres:
    plotNodes = plotNodes + str( i[1]) + ','
plotNodes = plotNodes[:-1] # remove last comma

# reset plot
newUDR.ClearGeneratedData()
newUDR.ScopingMethod = GeometryDefineByType.ResultFileItem
newUDR.ItemType = ResultFileItemType.NodeIDs
# create string of plot nodes

newUDR.SolverComponentIDs = plotNodes
newUDR.EvaluateAllResults()

Kleanthis