5 Replies Latest reply on Sep 3, 2013 2:29 PM by Jared Conway

    Question about going from steady-state to transient study

    Dan Hofstetter

      I'm confused about running transient airflow studies with contaminant gases.  I have been attempting to set up a study where a fan is used to evacuate contaminant gas from a tank, but I noticed while solving that the velocity field is not necessarily well-resolved during the beginning of the transient study.  Wouldn't this cause inaccurate results when looking at contaminant gas concentrations in the tank versus time?

       

      To overcome this, I attempted to set up a steady-state study without contaminant gas and 1) run it until the automatic convergence criteria were all met, then 2) set up a time-dependent study with transferred boundary conditions and then add an initial condition where the tank has a concentration of 100 ppm of contaminant gas.  I can't seem to get it set up correctly though.  Some problems I have noticed:

       

      1.   I tried to run the transient study with transferred initial conditions from the steady-state one (From the General Settings>Initial Conditions>Parameter Definition dialog).  However, I don't think the initial conditions are getting transferred correctly, because the transient calculation always starts re-solving the velocity field from scratch, i.e. the flow vectors are not the same as from the end of the steady-state solution.

       

      2.  I attempted to continue the calculation for the steady-state solution, but with the new initial condition in the tank unsuppressed and as a time-dependent study, but the solver always starts the calculation from the beginning.

       

      3.  For a transient-only study, I tried setting up an initial condition for a solid body inside the tank  with 100 ppm of contaminant gas being turned on after a number of time steps.  So at t=0 the concentration in the tank was 0 ppm, at t=20 the concentration was still 0 ppm, then at t = 21 the concentration should have jumped to 100 ppm.  I don't see this happening in the transient study though, the contaminant gas concentration inside the tank never rises above 0 ppm.

       

      Can anyone offer any suggestions?  I am going to take a look at the Tracer Gas analysis that the HVAC module offers, but I would rather do this without additional expense for add-ins if possible.  It seems to me that the contaminant gas will act as a passive scalar, which means once the flow field is resolved, I shouldn't have to continue to calculate the flow parameters for the simulation, but only continue the calculation with the final values for velocity in all mesh cells with the new values of contaminant gases until the final time step is reached.

       

      Thanks,

       

      Dan

        • Re: Question about going from steady-state to transient study
          Jared Conway

          the tracer study seems like the best fit for what you are trying to do. like you said, solve the flow field, this will remain the same. inject the tracer and see where the contaminant goes.

           

          that being said, i'm not sure i understand how you want to handle the problem with a transient analysis. are you using a flow volume flow rate to inject it? or are you dispersing the contaminant evenly or through a local initial condition. maybe taking a step back to that point would help me.

           

          overall though, your general methods for starting a transient with a steady state..etc seem correct.

            • Re: Question about going from steady-state to transient study
              Dan Hofstetter

              Jared,

               

              Thanks for the reply.  I'm using the transient analysis because I need to determine how long it will take to remove the contaminant gases.  The entire tank will be filled with a uniform gas concentration through a local initial condition.  Additionally, there will be a small mass flow rate of contaminant gas generated at the tank floor surface during the transient part of the simulation.

               

              Is there a way to make #3 above work?  With a local initial condition becoming filled with contaminant gas at a certain time in the transient simulation?  If I could do this, at least I could let the flow field resolve during the first 100 or so seconds, then turn on the gas at t = 101.  This also assumes I can find a way to disable automatic refinement after t = 100, so that the solver doesn't change the flow field again.

               

              Thanks,

               

              Dan

                • Re: Question about going from steady-state to transient study
                  Jared Conway

                  how are you trying to turn the initial condition on/off? i have to think about it a bit but i can't think of any way to do this.

                   

                  does the initial conditions have the contaminant in it and spread out among the computational domain?

                   

                  regarding the automatic refinement, i'm not sure what you mean.

                    • Re: Question about going from steady-state to transient study
                      Dan Hofstetter

                      By using an f(time) table.  To run this as a transient simulation, the contaminant gas shouldn't be introduced until the flow field is resolved.  For the initial conditions, the contaminant gas concentration would be 0 ppm everywhere.  Assuming the flow field resolves within the first 100 seconds of simulation time, I was trying to set a local initial condition for a solid body inside the tank (disabled in component control) to have a concentration of 0 ppm from 0-100 seconds, then 100 ppm starting at 101 seconds.

                       

                      Regarding the automatic refinement, inserting a goal plot displays the values for each goal variable during the calculation.  These eventually reach a semi-steady value as the criteria are met.  I noticed that at each refinement interval these values seem to reset somewhat, so the goal plot shows spikes in the values.  Once the criteria are met, it would be nice to have automatic refinement stop so that this doesn't happen again during the time when the contaminant gas is being used.