11 Replies Latest reply on Nov 25, 2014 11:21 AM by Jared Conway

    Time step

    Aibolat Dyuisenakhmetov

      I am new to SW and I am curios if there is any way to control time step in SW FlowSimulation.

        • Re: Time step
          Jared Conway

          assuming transient analysis?

          if you go in calculation control > there is a manual time step option

            • Re: Time step
              Aibolat Dyuisenakhmetov

              Thank you for your answer. You replied to my question about centrifugal pump. As far as I understand the simulation of centrifugal pump is semi-transient. This question is connected with it. Since the value of torque oscillates we thought that increasing time step will solve this problem.

              Kind Regards,

              Aibolat

                • Re: Time step
                  Jared Conway

                  increasing the time step will probably make things worse

                    • Re: Time step
                      Aibolat Dyuisenakhmetov

                      Dear Jared Conway,

                      Is there any possibility to change time step in a semi-transient(steady-state) analysis? We ran the simulation of centrifugal pump with casing and faced several different problems. Firstly, mass flow rate did not converge to specified value of 0.1% from total mass flow rate. Secondly, when we compare experimental and numerical values of torque, the experimental is 0.21Nm and mumerical is 0.12 Nm. Thirdly, friction focres exhibits fluctuations with amplitude of 50% of average value. However, the numerical pressure drop is in a good agreement with experimental one. We think that von Karman votices form right after the blades and this introduces such oscillations. In order to get rid of them we thought to increase time step. Thank you in advance. Any help will be appreciated.

                      Regards, Aibolat

                        • Re: Time step
                          Jared Conway

                          the controls are what you are see are what you get, as such, no there are no settings to change that

                           

                          best to report to your var and get a convo started with the developers

                           

                          my first thoughts are that you may want to make sure that you are following all the recommendations regarding rotating regions because the solutions will be very sensitive to that and make sure when you are outputting torque you are using best practices on face selection. if some parameters are good and some aren't, i'd guess outputting method to be the issue.

                            • Re: Time step
                              Aibolat Dyuisenakhmetov

                              Thank a lot for your answer. Now I have another questions. The first one is concerned with the outlet mass flow rate convergence criteria. What percentage of outlet mass flow rate convergence typically is acceptable? We put like 0.1% but mass flow rate do not converges. The second question is can we use transient analysis for the analysis of centrifugal pump performance?

                              Regards,

                              Aibolat

                                • Re: Time step
                                  Jared Conway

                                  "what convergence is acceptable" is up to you. for a large flow rate, 0.1% might be over kill. for very small, that might be too big.

                                   

                                  when you say it doesn't converge, that could mean a great number of things. for example maybe your time step is not small enough and it is impossible to get down to 0.1%, it is hard to say without actually working through the analysis with you.

                                   

                                  can transient simulation be used for a centrifugal pump? if you look at the help and tech references you will see that it is supported within certain guidelines. what exactly are you trying to solve and why can't it be done with steady state?

                                   

                                  in either case, I would recommend making sure your problems solves with steady state first.

                                    • Re: Time step
                                      Aibolat Dyuisenakhmetov

                                      The inlet flow rate is 0.5175 l/s.  The problem is as far as I understand the we cannot control time step in FlowSimulation it is rather determined by the size of the smallest

                                      cell. for_sw_forum.jpgIn addtion, we are forced to refine mesh thoroughly since in some place we have a distance of 2 mm between impeller and inner wall of the casing(the smallest gap in blue colored region). The diameter of the impeller is 120 mm and rotationg region's diameter is 2mm larger. It seems to me that the problem is pressure fluctuations created when one of the impeller blades goes through that region near to exit and where we have the smallest gap.