How to extract the average unbalanced force ratio of blocks belonging to a certain group in the 3DEC 7.0

3DEC
1

Hi,
I want to extract the average unbalanced force ratio of blocks belonging to a certain group in the 3DEC 7.0, not for all blocks. How can I design the FISH command flow for this? I know that the command block.mech.ratio.avg can be used to extract the average unbalanced force ratio of all blocks in the model.
Thanks in advance!

2

Hi @JLUJYP !

I provide a Fish suggestion in PS of this message.
This solution is based on ratios at model gridpoints.
Notice that this solution works for block groups, you will have to adapt the script if you want to restrict to zone groups.

Warning : the final solution does not match the `block.mech.ratio.avg’, I am missing some details.
For example on a model of mine:

3dec>[block.mech.ratio.avg]
0.000150823
3dec>[ratio_all()]
0.000304065

@jhazzard, any hint on this?

Thanks!

Théophile

PS: the script

def ratio_in_group(group, slot)
    local ratio_avg = 0
    local ngps_in_group = 0
    loop foreach local bpnt block.list
        if block.isgroup(bpnt, group, slot)
            loop foreach local gppnt block.gplist(bpnt)
                ratio_avg += block.gp.ratio(gppnt)
                ngps_in_group += 1
            end_loop
        end_if
    end_loop
    if ngps_in_group > 0
        ratio_avg /= ngps_in_group
    end_if
    ratio_in_group = ratio_avg
end

def ratio_all()
    local ratio_avg = 0
    local ngps_in_group = 0
    loop foreach local bpnt block.list
        loop foreach local gppnt block.gplist(bpnt)
            ratio_avg += block.gp.ratio(gppnt)
            ngps_in_group += 1
        end_loop
    end_loop
    if ngps_in_group > 0
        ratio_avg /= ngps_in_group
    end_if
    ratio_all= ratio_avg
end
3

Hi, @Theophile,
Thank you very much! The script you provided should be able to solve the problem I encountered.
JLUJYP

4

This is almost correct. The problem is that the block.gp.ratio function gives the ratio of unbalanced force to total force for a single gridpoint. The global ratio is the sum of all unbalanced forces divided by the sum of all forces - essentially the sum of all numerators divided by the sum of all denominators, which is not the same as the sum of all the fractions.

You would have to do this calculation by hand using the block.go.force.unbal and block.gp.force.mag functions. There is one extra wrinkle if you have gravity. See how to do it below:

fish def get_ratio
total_unbal = 0
total = 0
loop foreach bp block.list

  ; here you can check for block group or whatever

  loop foreach gp block.gplist(bp)
    total_unbal += math.abs(block.gp.force.unbal.x(gp))+math.abs(block.gp.force.unbal.y(gp))
    ; need to include gravity for z component
    total_unbal += math.abs(block.gp.force.unbal.z(gp)+block.gp.mass(gp)*global.gravity.z)
    total += block.gp.force.mag(gp)
  end_loop
end_loop
get_ratio = total_unbal / total

end

5

PS @Theophile - how do you get your script to appear in grey with the correct spacing?

6

Hi!

Thanks for the clarification!

The block.gp.force.mag does not appear in 3DEC7’s documentation, nor in 3DEC9’s.

I wonder if the Fish connection with actual gp forces is done since I systematically get 0 for all gps in all models I tested.

I only use the “Preformatted test” option proposed on top of comment editing window (or “ctrl+e” on a new line). The tab key then inserts 2 spaces. As far as I’m concerned, I edit my whole scripts in Notepad++ to test them (where the tab key inserts 4 spaces), then copy/paste the script!

7

Holy moly, you are correct. I’ll add this to the docs for the next update.

It seems that 4 spaces are the key to the formatting …

Here’s a super-simple example that seems to work (in v9)

model new
model random 10000

block create brick -10 10
block zone gen edge 2.5
block zone cmodel assign elastic
block zone prop dens 2500 bulk 1e9 shear 0.6e9 

block gridpoint apply vel-z 0 range pos-z 0
;block face apply stress-zz -1e6 range pos-z 10

model grav 10

model large-strain on

fish def get_ratio
  total_unbal = 0
  total = 0
  loop foreach bp block.list

    ; here you can check for block group or whatever

    loop foreach gp block.gplist(bp)
      total_unbal += math.abs(block.gp.force.unbal.x(gp))+math.abs(block.gp.force.unbal.y(gp))
      ; need to include gravity for z component
      total_unbal += math.abs(block.gp.force.unbal.z(gp)+block.gp.mass(gp)*global.gravity.z)
      total += block.gp.force.mag(gp)
    end_loop
  end_loop
  get_ratio = total_unbal / total
end

model his mech ratio
fish his get_ratio
fish his total

model solve
8

I confirm that in 3DEC7, the script you gave crashes since the total block.gp.force.mag is 0. I am not up to date with version 9 of 3DEC, but I could reload a .sav from 3DEC7 then run the get_ratio function and I still get the same error.

I guess something happens in 3DEC7 resulting in a broken link from block.gp.force.mag to actual forces.

Anyway, thanks for having tested!

Théophile