9 Replies Latest reply on Oct 11, 2013 5:40 PM by Jared Conway

    The "Take Previous Results" option

    Ramachandran Kasiraman

      Hi all,

       

      I am trying to figure out what the "take previous results" option in 'Flow Simulation' does. Since I had altered the properties of the porous media and my outlet boundary conditions, I was not able to continue the calculation. But, I was able to check the "take previous results" option and I was hoping that it will save me a lot of time during calculation. However, I was not able to see any difference. The entire meshing was done from scratch and when I looked at the different values, they seemed to have been intialized from the beginning. So, now I am confused as to what this option does. Thanks a lot.

       

      Ram.

        • Re: The "Take Previous Results" option
          Jared Conway

          "Take previous" uses your previous results as the initial conditions for the analysis.

          If you change your analysis significantly enough that a remesh is required, you cannot "Take previous". You could try the transferred boundary conditions option in the intial conditions, it will do its best to map the initial conditions from that previous analysis. This is more like post mesh initial conditions i believe.

           

          it may be helpful to show some screenshots of your screens when solving and to have a bit more information about what exactly you are changing that is forcing a mesh change.

           

          you may also want to try this on a very simple part just to get a hang of how the functions work.

            • Re: The "Take Previous Results" option
              Ramachandran Kasiraman

              Hi..

               

              Thanks for the reply.. That was exactly what I thought about that option. But, I guess the issue is that the results are not getting saved completely.. Is there any document on how the results are stored and organized in the flow simulation? I see "Saving results" and "Finishig.." in the status of the analysis window, but it goes back to calculation again at which point i close the window. I am not sure if that is causing the error.. Because, I get the "Take previous results" option when i try to do the analysis again, but I see that it is starting afresh. I will try to add some screenshots soon.. I am currently running a simulation..

               

              I have another quick question regarding the "Parametric Study" on a similar note.

              Parametric Study.jpg

              As you could see, I am trying to run the simulation for different boundary cases to determine the necessary parameters. My question is if the meshing and calculations will generally be started afresh everytime or will they be continued to attain quick convergence. As you can see, the parameters change gradually and hence continuing calculations through the different scenarios would make sense.. I would appreciate it if you could tell me as to what is the default way in which this is carried out and if there are options to change it if necessary..

               

              Additionally, i wanted to know if we can do a parametric study on the "humidity".. I don't see humidity under the "add a parameter" in the study.. I see temp, pressure and even turbulence parameters, but not humidity.. you can see that i have added humidity in the flow, but I am not able to add it as a parameter.

              gen sett.jpghumity and add a para.jpg

              To begin with, I was trying to do this same thing manually and thats when the "take previous results" didn't work.. I only changed the Outlet vol flow rate and the temp at the inlet.. there was no change made to the mesh settings and other stuff.. But it started the calculations afresh.. Hence, I believe that it was my mistake and I closed the window prematurely.. I will surely look into it and get back if the problem persists.. But any info regarding my parametric study question would be greatly appreciated.. Thank you very much for your time..

               

              Ram.

                • Re: The "Take Previous Results" option
                  Jared Conway

                  i would recommend creating new posts for each of your questions

                   

                  on the first issue, i would suggest trying on a simpler part and posting it if it shows the problem. if it doesn't, you may want to work with your reseller on the question so they can look at your files directly. i'm not really following your workflow and for what you're asking about, i don't think there is any documentation but i'm not really following what your question is either.

                   

                  parametric study, think about it this way, it automates creating the configs/projects, so they are independent and start anew. if you want to change their starting point, use the method i described with transferred boundary conditions. this will require a little bit of effort and you won't necssarily get the fancy plots from parametric study but with batch run, you can at least get all the runs in quickly.

                   

                  humidty in parametric study, not all parameters are editable. you'll need to find a way to workaround it or use the "old way". i'd have to dig in further to see if this is a limitation or a bug. again, maybe something to work with your reseller on so they can look at your specific case.

                  • Re: The "Take Previous Results" option
                    Paul Beierle

                    Humidity can be found in the "Fluid Subdomain" section. By selecting humidity through here it will appear in the study tree as a parameter.

                     

                    I find I get better 'paramaterization' in general using this feature; even if I'm only working working with a single fluid.

                     

                    This is a common thread throughout parametric studies. For example, If you were to assign a local mesh to your model, then that feature will show up as an editable parameter. Same goes for many others.

                     

                    As far as the inital question, I would only see using 'previous results' in part of a time-step analysis. Or, maybe when solving an internal study then using those results to influence a following one. However, my knowledge is more limited in that area and I'm pretty much speculating.

                     

                    HumidityCheck.png

                      • Re: The "Take Previous Results" option
                        Jared Conway

                        nice tip paul, (didn't try it but assume that you wouldn't post it if it didn't work). do you have an eSR for that already? to get access to other parameters that aren't available?

                          • Re: The "Take Previous Results" option
                            Paul Beierle

                            Yessir, I had just run a para-study using humidity as a variable; to see how it affected thrust output on a propulsor I'm working on.

                            Although, correct me if I'm wrong but I think the preset "air" fluid defaults at 50% when you don't have humidity selected.

                             

                            Oh, what is this esr you speak of?  Equivalent Series Resistance? Engineering Service Request? Erythrocyte Sedimentation Rate?

                             

                            Or were you asking the original poster?

                             

                             

                            Regardless, I :heart: parametric study. It's really intuitive and a great adjunct to the clone n' fig -> batch run dance. I'm sure it, like all the other modules, will only get better with time.