Great video! I really liked how you explained making calculations faster with the timestep factor and showed the memory needed for 1 million cells in different methods.
so in short, increasing pseudo time step will increase convergence, but I have seen that for high gradients, the initial time step size should be kept very very low llike courant number
Not pseudo time step but time scale factor (this is a new addition in Fluent solver) Yes, as I said in the end of video, that for cases, when you have convergence issues with higher flow gradients in flow, decrease time scale factor was initial run and they try to increase it. But if you dont see any benefit, then it is better to use SIMPLE method in that case. As coupled solver is expensive on per iteration basis and also it required double RAM as compared to SIMPLE family of algorithms.
I'm working on hydro turbine with zero head and the only confusion is if we not make an animation of the rotation of blades by using mesh interface giving the RPM to blades and we use the method that you have used so the answers of the moment are correct? or we should animate and then find the moment values
Your explanation is so good. We are a startup looking forward to tie up with you as an industry expert. Can you provide your email so I can send details.
![[CFD] Residuals in CFD (Part 1) - Understanding Residuals](http://i.ytimg.com/vi/v9OnNeYH4Ok/mqdefault.jpg)
![[CFD] Residuals in CFD (Part 1) - Understanding Residuals](/img/tr.png)
Great video! I really liked how you explained making calculations faster with the timestep factor and showed the memory needed for 1 million cells in different methods.
Videos like this are very valuable. Thanks a lot for sharing your experience.
so in short, increasing pseudo time step will increase convergence, but I have seen that for high gradients, the initial time step size should be kept very very low llike courant number
Not pseudo time step but time scale factor (this is a new addition in Fluent solver)
Yes, as I said in the end of video, that for cases, when you have convergence issues with higher flow gradients in flow, decrease time scale factor was initial run and they try to increase it.
But if you dont see any benefit, then it is better to use SIMPLE method in that case. As coupled solver is expensive on per iteration basis and also it required double RAM as compared to SIMPLE family of algorithms.
Sir make a video..CFD Analysis on Solar Flat Plate Collector and Evacuated Tube Collector
I just need a geometry or sketch or dimensions along with required boundary conditions and I could a make video on it.
Okay Sir..give me your mail id..I will directly mail you...with all the details required.
I'm working on hydro turbine with zero head and the only confusion is if we not make an animation of the rotation of blades by using mesh interface giving the RPM to blades and we use the method that you have used so the answers of the moment are correct? or we should animate and then find the moment values
I have a problem exporting the residual to plot in originlab
Does the computer use ram according to the number of nodes or according to the number of cells?
Number of cells. In Finite volume method we refer to cells and in Finite differnce method (FDM) we use nodes.
Your explanation is so good. We are a startup looking forward to tie up with you as an industry expert. Can you provide your email so I can send details.
sijal2005@gmail.com