FNO Reader

[Anonymous]

@John_Holtz 

I have also attached FNO file for a very simple simulation along with a csv file created using FNO reader for principal stresses in A,B and C direction and Octahedral stress. I plan to replace the values of Octahedral stress with sum of principal stresses. However I keep stumbling upon this error (see image).

Strangely, the error appears even when I try to append the FNO file with the values of same csv file which was derived from FNO file. what am I missing here?

[Anonymous]

@John_Holtz 

I have also attached FNO file for a very simple simulation along with a csv file created using FNO reader for principal stresses in A,B and C direction and Octahedral stress. I plan to replace the values of Octahedral stress with sum of principal stresses. However I keep stumbling upon this error (see image).

Strangely, the error appears even when I try to append the FNO file with the values of same csv file which was derived from FNO file. what am I missing here?

[AstroJohnPE]

Hi @Anonymous 

Thanks for reporting this issue. Let me make sure that I understand what you are doing.

• On page 1 of FNO Reader, you selected the second option and chose the comma as the separator (item 1 in the image below).
• You selected the CSV filename (item 2).
• On page 3, you selected the existing FNO file (item 3).
• You indicated to append the results (item 4).
• When you click the "Create Output" button, that's when you get the error.

The error message indicates that some number is too large. Since those steps worked for me, I am not entirely sure how FNO Reader is finding a wrong number. The message does not indicate that the files were "in use", but just to eliminate one possibility, I assume that the CSV file was not being edited by any other program, and you were not viewing the results in Inventor at the time. Is that correct?

What characters does your computer use for the decimal symbol? Is it a period or a comma? (Or something else.) It could be misreading the commas that separate the columns as a decimal point if your computer is setup with the comma as the decimal symbol. Could you try writing out the CSV file using a semi-colon as the separator? (On page 3 when converting the FNO file to text, change the separator from Automatic or Comma to Semi-colon.)

Thanks. I will look at the code to see what math operation could be causing an issue.

John

[AstroJohnPE]

Hi @Anonymous 

Thanks for reporting this issue. Let me make sure that I understand what you are doing.

• On page 1 of FNO Reader, you selected the second option and chose the comma as the separator (item 1 in the image below).
• You selected the CSV filename (item 2).
• On page 3, you selected the existing FNO file (item 3).
• You indicated to append the results (item 4).
• When you click the "Create Output" button, that's when you get the error.

The error message indicates that some number is too large. Since those steps worked for me, I am not entirely sure how FNO Reader is finding a wrong number. The message does not indicate that the files were "in use", but just to eliminate one possibility, I assume that the CSV file was not being edited by any other program, and you were not viewing the results in Inventor at the time. Is that correct?

What characters does your computer use for the decimal symbol? Is it a period or a comma? (Or something else.) It could be misreading the commas that separate the columns as a decimal point if your computer is setup with the comma as the decimal symbol. Could you try writing out the CSV file using a semi-colon as the separator? (On page 3 when converting the FNO file to text, change the separator from Automatic or Comma to Semi-colon.)

Thanks. I will look at the code to see what math operation could be causing an issue.

John

[Anonymous]

Very strange that a long reply I just posted suddenly disappeared? I did see it appear below John's reply so it was successfully posted. Anyway I will summarise it below. 

Currently I am experimenting with a model with 600k elements and the csv file obtained from FNO reader for principal stresses A/B/C and octahedral stress is about 2.4 million long. My typical models will be over 1 million elements which means csv file will be even bigger. So I can't use Excel for obtaining sum of principal stresses as no. of csv files will be large due to a limit of 1 million lines that Excel can deal with. So I use MATLAB/Octave to replace the octahedral stress values with sum of principal stresses and it is quick. 

Now, what I have found is that if a csv file is created from FNO file and if I immediately append it to the FNO file, it works. However if I try appending with modified csv file (with sum of principal stresses), it fails. So I use following workflow:

1) Copy original FNO file and csv file to a folder - say Folder A

2) Modify csv file for sum of principal stresses and then copy the FNO file and modified csv file to a different folder, say Folder B

3) Copy contents of folder A to InCAD\FEA directory and append the original FNO file with original csv file - this generally always works and has an effect of "disciplining" FNO reader somehow!

4) Then I copy contents of folder B to InCAD\FEA directory and append with modified csv file, which I find works.

So basically, if step #4 is done without doing step #3, it fails! Which is bizarre.

EDIT: To answer John's queries, yes I use period for decimals so csv file shouldn't have any issue. 

[Anonymous]

Very strange that a long reply I just posted suddenly disappeared? I did see it appear below John's reply so it was successfully posted. Anyway I will summarise it below. 

Currently I am experimenting with a model with 600k elements and the csv file obtained from FNO reader for principal stresses A/B/C and octahedral stress is about 2.4 million long. My typical models will be over 1 million elements which means csv file will be even bigger. So I can't use Excel for obtaining sum of principal stresses as no. of csv files will be large due to a limit of 1 million lines that Excel can deal with. So I use MATLAB/Octave to replace the octahedral stress values with sum of principal stresses and it is quick. 

Now, what I have found is that if a csv file is created from FNO file and if I immediately append it to the FNO file, it works. However if I try appending with modified csv file (with sum of principal stresses), it fails. So I use following workflow:

1) Copy original FNO file and csv file to a folder - say Folder A

2) Modify csv file for sum of principal stresses and then copy the FNO file and modified csv file to a different folder, say Folder B

3) Copy contents of folder A to InCAD\FEA directory and append the original FNO file with original csv file - this generally always works and has an effect of "disciplining" FNO reader somehow!

4) Then I copy contents of folder B to InCAD\FEA directory and append with modified csv file, which I find works.

So basically, if step #4 is done without doing step #3, it fails! Which is bizarre.

EDIT: To answer John's queries, yes I use period for decimals so csv file shouldn't have any issue. 

[Anonymous · Accepted solution]

For anyone who may stumble upon this - the issue was how FNO reader reads the csv files you are trying to append. e.g. if you are trying to append data from file-RSLT.csv and there is another file such as file-RSLT - Copy.csv, then this second file is included by FNO reader as perhaps it searches with wildcard "file-RSLT*" and if this file doesn't conform to a specific structure, it fails to append. 

[Anonymous · Accepted solution]

For anyone who may stumble upon this - the issue was how FNO reader reads the csv files you are trying to append. e.g. if you are trying to append data from file-RSLT.csv and there is another file such as file-RSLT - Copy.csv, then this second file is included by FNO reader as perhaps it searches with wildcard "file-RSLT*" and if this file doesn't conform to a specific structure, it fails to append.