[jins9158]

I am working about ventilation's system fan analysis

I want to get about mass flow data for CFD result because my senior want to know about performance difference by seeing mass flow data

So I am setting and Inlet is "Total Pressure", Outlet is "Static Pressure", rotation is applying.

When I see CFX guide book, I saw several boundary type and I couldn't understand

1. Most Robust
- Inlet [Velocity/Mass Flow] / Outlet [Static Pressure]
- Inlet total pressure is an implicit result of the prediction

2. Robust
- Inlet [Total Pressure] / Outlet [Velocity/Mass Flow]
- static pressure at Outlet and velocity at Inlet are part of the solution

3. Sensitive to Initial Guess
- Inlet [Total Pressure] / Outlet [Static Pressure]
- The system mass flow is part of the solution

4. Very Unreliable
- Inlet [Static Pressure] / Outlet [Static Pressure]
- Not recommend inlet total pressure level and mass flow are implicit result of the prediction

Why "Sensitive to Initial Guess"?
What difference does it make when calculated by an internal algorithm?

[Opaque]

I think you are interpreting the information listed backward.

If the solution is well converged, that solution should be independent of the initial guess.

HOWEVER, if the initial guess is poor, there is no guarantee you can converge the solution; therefore, there is nothing to compare against.

Hope the above helps,

[ghorrocks]

Another way to look at it is that the list is in order of numerical difficulty.

Option 1 is the most numerically easy, so you should get the best convergence with this option.

Through to Option 4 which is the most numerically difficult, and you are most likely to have problems with convergence if you have set it up this way.

[jins9158]

Originally Posted by ghorrocks

Another way to look at it is that the list is in order of numerical difficulty.

Option 1 is the most numerically easy, so you should get the best convergence with this option.

Through to Option 4 which is the most numerically difficult, and you are most likely to have problems with convergence if you have set it up this way.

thank you for replying my question

So you mean if the convergence is good, all option's is ok?

Is it right?

[ghorrocks]

Good convergence is one of many, many things which need to be right to have an accurate simulation. See FAQ: https://www.cfd-online.com/Wiki/Ansy..._inaccurate.3F

[jins9158]

Originally Posted by ghorrocks

Good convergence is one of many, many things which need to be right to have an accurate simulation. See FAQ: https://www.cfd-online.com/Wiki/Ansy..._inaccurate.3F

Thank you for answering my question

I saw link's information and I realized that I already knew information

So, I think you wrong my question

I want to know about the process's difference in condition of 4 option

The guide book just let me know that "1 option is very robust", "2 option is robust" briefly

I am wondering why in detail (for example, formulation's difference)

thank you for your concern

[ghorrocks]

Opaque's comment on another thread is relevant here: Convergence Issue for Unsteady Compressor Simulations

Again, it all comes back to numerical stability. The different boundary condition combinations result in different system unknown variables to be solved. Some of these system variables are numerically easier to solve than others. (By system variable, I mean total flow rate, overall pressure drop and system level parameters like that)