Query regarding force calculation

Comments

• rubens

  May 2023

  Hi @Vijay,

  
  

  Have you tried gaugeForces ?

  It computes forces throughout the simulation, i.e., it cannot be used after simulation like ComputeForce.

  
  

  For more information, in the PDf file XML_GUIDE_vXX, look for the item XML file: Special-Gauges. Also, see the file _FmtXML_Gauges.xml in folder doc/xml_format

  Below is an example to compute resultant forces on boundary particles with mkbound=1 and mkbound=2 (no need to add 10 to mklbound!), and considering all time steps.

  Just add in the xml file the code below

     ...
     </casedef>
     <execution>
        <special>
           <gauges>
              <default>
                 <savevtkpart value="true" comment="Creates VTK files for each PART (default=false)" />
                 <computedt value="0" comment="Time between measurements. 0:all steps (default=TimeOut)" units_comment="s" />
                 <_computetime start="0.1" end="0.2" comment="Start and end of measures. (default=simulation time)" units_comment="s" />
                 <output value="true" comment="Creates CSV files of measurements (default=false)" />
                 <outputdt value="0" comment="Time between output measurements. 0:all steps (default=TimeOut)" units_comment="s" />
                 <_outputtime start="0" end="10" comment="Start and end of output measures. (default=simulation time)" units_comment="s" />
              </default>
              <!-- Calculates force sumation on selected fixed or moving particles (using only fluid particles). -->
              <force name="Force1">
                 <savevtkpart value="true" comment="Creates VTK files for each PART (default=default.savevtkpart)" />
                 <output value="true" comment="Creates CSV files of measurements (default=default.output)" />
                 <target mkbound="1" comment="Indicates the mkbound of selected particles to compute forces" />
              </force>
              <force name="Force2">
                 <savevtkpart value="true" comment="Creates VTK files for each PART (default=default.savevtkpart)" />
                 <output value="true" comment="Creates CSV files of measurements (default=default.output)" />
                 <target mkbound="2" comment="Indicates the mkbound of selected particles to compute forces" />
              </force>
           </gauges>
        </special>
        <parameters>
        ...
  

  
  

  Hope this help you.
• Vijay

  May 2023

  Hi @rubens ,

  Thank you so much for your reply. I have not tried this method yet. I will definitely give a try to implement this for the force calculations. Thanks again!

  
  

  However, I will keep rerunning the simulation as a last option as the it is computationally expensive. So, if @rubens or anyone has a idea about how to make this resolve the problem just with computeForce, please let me know.

  
  

  Thank you..
• rubens

  May 2023

  Hi @Vijay ,

  Maybe the problem is that you considered all the building solid walls.

  ComputeForces gets accelerations of all selected wall MKbound particles, see Eq. 48 in https://link.springer.com/article/10.1007/s40571-021-00404-2

  I think in your case the drag forces are like the image below :

  

  According to the file doc/guides/ComputeForces_Help.out , you can set filters, and the option below looks like what you need, i.e. set the region containing the face of the building hit by the flow.

     -onlypos:xmin:ymin:zmin:xmax:ymax:zmax Indicates limits of particles

  The classic 3D dam-break case in the main folder adopts a different material for the face of the column. It also avoid forces on the lateral and back walls.

  
  

  Hope this works now!
• Vijay

  May 2023

  Hi @rubens ,

  Yes. The picture you have shown is correctly depicted what I want! Also thanks for sharing the article.

  However, for my case, I want to consider all faces, instead of just a front face. So, I am taking a stl file for building and assigning mk number.

  Initially I tried computeForce with single building and I got desired results.

  

  Although, here mk assignment was as follows -

  					<!-- Piston -->			
  					<setdrawmode mode="full" />
  					<setmkbound mk="10" />
  					<drawbox>
  						<boxfill>solid</boxfill>
  						<point x="4" y="1.0" z="0.0" />    
  						<size x="0.05" y="4.0" z="4.0" />
  					</drawbox>
  
  
  					<!-- Flume -->
  
  
  					<!-- Basin -->
  					<setdrawmode mode="face" />
  					<setmkbound mk="0" />
  					<drawfilestl file="basin.stl" objname="basin" autofill="false">
  						<drawscale x="0.001" y="0.001" z="0.001" />
  						<drawmove x="17.05" y="1" z="0" />
  					</drawfilestl>
  					<!-- Building -->	
  					<setdrawmode mode="face" />
  					<setmkbound mk="1"/>
  					<drawfilestl file="building.stl" objname="building" autofill="false">
  						<drawscale x="0.001" y="0.001" z="0.001" />
  						<drawmove x="25.79" y="2.8" z="0.795" />
  					</drawfilestl>												
  

  So here the force on single building has been calculated using

  ComputeForces -dirin out -onlymk:11 -savecsv forces
  

  So, I am curious to know if this is working for single building, then why its not working for multiple building. I must have made an error in assigning mk values. What do you think?
• Alex

  May 2023

  You should use "-viscoauto" in order to use the same viscosity values you used during the simulation

  
  

  Regards
• Vijay

  May 2023

  Thanks @Alex , I will use following command using -viscoauto for multiple building case

  ComputeForces -dirin out -viscoauto -onlymk:14 -savecsv forces
  

  And will update about the results.

  Thanks!