Working with large data sets (loads) 2

[wth]

I've received a printout from FEM-analyze that I'm trying to find a suitable way to organize and work with using mathcad and/or excel.

The data set contains basic loads for 14 rigid nodes, with 12 different load cases. So each line consists of: node/LC, Fx, Fy, Fz, Mx, My, Mz, "name of load case.
Example of the table in a smaller scale (random numbers):

http://i.imgur.com/KR2aL.png

What I'm trying to do, is finding an easy way to calculate relevant load combinations for each node, and then sort it in order to find the worst design loads in total. The loads are for 7 identical combined footings, 2 nodes pr footing. A load case would look something like 1.5*LC1+1.2*LC2...

So far I've got the table in excel, and I've messed around with different matrices in Mathcad, but I'm unable to find a suitable set up for what I'm trying to do.. So any tips for how to work with data sets like this is much appreciated!

(sorry if this is better placed under another forum, but figured structural engineers are most likely to have encountered similar problems)

 

[dik]

You have to filter your data output... most programs allow you to do this...

Consider a worst case scenario... there are an infinite (really!) number of points within a structure... and, being a finite individual, you can only appreciate a few of them in your relatively short lifetime...

Dik

 

[wth]

I can filter the orignial report in excel..
The problem here is that it could be multiple worst case scenarios depending on what I'm designing for (soil pressure, stability, structural..), and I have 14x13x6 different loads.. So I was hoping to somehow have the different load combinations calculated before filtering.

One way I suppose it can be done is separating each load case(ie snow load), then assign them to matrices in Mathcad, and calculate my load combinations.. but then I'll just have even more results that needs to be filtered..

Guess I'll just have take the worst in each direction from the basic loads, even if they don't belong to the same node, and use those for a worst than worst case scenario..

 

[dik]

Does your program allow you to create a stress/force envelope using various loading combinations... This would go a long way to sorting out the chaff...

Dik

 

[wth]

My FEM-software? I've received the loads from a 3rd party, so I don't have the actual model, just a print out of the results.

I could of course use FEM myself, import/enter the loads and various combinations.. have to check on Monday if any of our software have the ability to import loads from text.. thanks for the idea!

 

[IDS]

This should be fairly straightforward in Excel.

Set up a table of load factors for each combination.
Use Vlookup to enter the load factor for each loadcase in a column at the end of your data.
For the first node use sumproduct to find the factored value for each result.
Copy those formulas down for each of the other nodes.
Rather than sorting, use the Max and Min functions to find the critical value, and the Match function to locate which node has that value.

For the last step it would probably be best to use the Index or Offset functions to combine the results into a contiguous column, rather than results separated by spaces.

Doug Jenkins
Interactive Design Services

http://newtonexcelbach.wordpress.com/

 

[wth]

Thanks, sounds like what I'm after!
Did a quick test in open office here (saved in excel format) on a sample sheet.

Not very fluid so far, have to try some more at work with excel and a bigger screen..

 

[IDS]

I have added some suggestions to your sheet (attached).

The Hlookup forumulas in cells I8:K8 you can just copy down as far as you need. Note that the FALSE on the end of the formula is important.

The sumproduct formulas in Colums L to R you need to copy down in blocks of 4 rows:
Select L8:R11
Copy
Select from L12 4n rows, where n is the number of nodes.
Paste

Open Office should be fine for all this by the way (as far as I know).

Doug Jenkins
Interactive Design Services

http://newtonexcelbach.wordpress.com/

 

[wth]

Definitely an improvement, looks to be a lot less manual input, thanks again!

Seems to work fine in Open Office, lack of large monitors on the other hand..

 

[dik]

Thanks... useful for other stuff...

Dik