[songkai]

Hello everyone,

now i am working about the flow inside the cylinder and my domain is very simple just a pipe. Now i have known the velocity, temperature and pressure of inlet and the density of gas is also known. Because the case is in the internal combustion engine, these physical values according to crank angle have relative huge changes. So i have some doubts:

1.for the choice of inlet typ. if i choose velocity inlet, i can not set the BC of pressure. But in the internal combustion engine, i think, the pressure inlet is a very important BC.

2.for the choice of EOS. During the process of calculation of these BC values, i used ideal gas law. So i think i should choose ideal gas, but if there is not data of pressure BC input, the ideal gas law does not have any significance. But if i choose the density rho(T), that does not match with my calculation.

So anyone can help me!
Thanks AD!

Songkai

[ashokac7]

Originally Posted by songkai

Hello everyone,

now i am working about the flow inside the cylinder and my domain is very simple just a pipe. Now i have known the velocity, temperature and pressure of inlet and the density of gas is also known. Because the case is in the internal combustion engine, these physical values according to crank angle have relative huge changes. So i have some doubts:

1.for the choice of inlet typ. if i choose velocity inlet, i can not set the BC of pressure. But in the internal combustion engine, i think, the pressure inlet is a very important BC.

2.for the choice of EOS. During the process of calculation of these BC values, i used ideal gas law. So i think i should choose ideal gas, but if there is not data of pressure BC input, the ideal gas law does not have any significance. But if i choose the density rho(T), that does not match with my calculation.

So anyone can help me!
Thanks AD!

Songkai

First thing first, you are trying to do many things at once. Anyone would suggest that it is better to have first few approximations deviating a lot from actual results than to stuck in between choosing the physics.


So take velocity inlet. analytically calculate the pressure drop across tube (as you have mentioned) and that's how you get pressure at outlet. There is also an another way to do it, first take pressure at outlet as given pressure at inlet. Do a steady analysis and calculate the pressure drop. Then change pressure at outlet.



You can certainly choose pressure as inlet but I think it is better to start with velocity inlet, that way we learn about the convergence issues first.


I am unable to understand your second point. Please elaborate. Choose whatever suggested in research paper, and see what is percentage deviation you are getting from the experimental results. (Let us know)



And if it is is simple pipe, what are the factors which will cause density variation? Like how much is a temperature difference. And see if constant density is a too bad approximation?


Hope this helps.

[songkai]

Originally Posted by ashokac7

First thing first, you are trying to do many things at once. Anyone would suggest that it is better to have first few approximations deviating a lot from actual results than to stuck in between choosing the physics.


So take velocity inlet. analytically calculate the pressure drop across tube (as you have mentioned) and that's how you get pressure at outlet. There is also an another way to do it, first take pressure at outlet as given pressure at inlet. Do a steady analysis and calculate the pressure drop. Then change pressure at outlet.



You can certainly choose pressure as inlet but I think it is better to start with velocity inlet, that way we learn about the convergence issues first.


I am unable to understand your second point. Please elaborate. Choose whatever suggested in research paper, and see what is percentage deviation you are getting from the experimental results. (Let us know)



And if it is is simple pipe, what are the factors which will cause density variation? Like how much is a temperature difference. And see if constant density is a too bad approximation?


Hope this helps.

Firstly thank u a lot, ashokac7.

The simulated environment is the cylinder of the internal combustion engine, so the temperature and pressure density change with time reciprocating. The loction,that i want to focus on, is the cylinder wall, the boundary layer between fluid and solid. I have tried from beginning of steady to unsteady with v(t) and T(t) step to step and good convergence. Inlet is velocity inlet and outlet is pressure outlet. During these steps i set the pressure in outlet with default value (0.0 Pa) and constant density. But i think the density is also changing according to the time, constant density will be not correct and the pressure should be also putted in. Now can u understand?

Songkai

[ashokac7]

Originally Posted by songkai

Firstly thank u a lot, ashokac7.

The simulated environment is the cylinder of the internal combustion engine, so the temperature and pressure density change with time reciprocating. The loction,that i want to focus on, is the cylinder wall, the boundary layer between fluid and solid. I have tried from beginning of steady to unsteady with v(t) and T(t) step to step and good convergence. Inlet is velocity inlet and outlet is pressure outlet. During these steps i set the pressure in outlet with default value (0.0 Pa) and constant density. But i think the density is also changing according to the time, constant density will be not correct and the pressure should be also putted in. Now can u understand?

Songkai

Okay. So you are saying you want to analyze combustion chamber as pipe. And what about the combustion modelling? I thought it is exhaust pipe or something.

I understand the problem now. But what is your objective. Can you describe.

[songkai]

Originally Posted by ashokac7

Okay. So you are saying you want to analyze combustion chamber as pipe. And what about the combustion modelling? I thought it is exhaust pipe or something.

I understand the problem now. But what is your objective. Can you describe.

just choose a small location on the chamber wall near the valve seat. The 2D domain just 5 mm*10 mm for flow model and 10*10 mm for solid model. The one side of solid model contacts with one side of flow model. The BC is cycle according to the reciprocating piston without reaction. In fact there is not a detailed modelling. Just simulate as a pipe with v(t), p(t),T(t) and rho(t) as BC in inlet and focus on the heat transfer on the interface between flow model and solid model. Now can u understand?

Thank u a lot

[ashokac7]

Originally Posted by songkai

just choose a small location on the chamber wall near the valve seat. The 2D domain just 5 mm*10 mm for flow model and 10*10 mm for solid model. The one side of solid model contacts with one side of flow model. The BC is cycle according to the reciprocating piston without reaction. In fact there is not a detailed modelling. Just simulate as a pipe with v(t), p(t),T(t) and rho(t) as BC in inlet and focus on the heat transfer on the interface between flow model and solid model. Now can u understand?

Thank u a lot

Yes. I get your problem now. I have very limited knowledge in this area though. What I understand is density changes with temperature and so. Then is it that necessary to include time dependency of it.? (as you said rho(t))

You have cyclic BC's, right? And I am assuming you are running 4-5 cycles and then taking the results for last cycle. And these results are cycle averaged, as considering results at some point of time may be misleading.
And time step is such that single time step will cause approx 1 deg crank angle rotation.

For the pressure inlet boundary condition, put inlet as pressure outlet and apply table, and at outlet give negative mass flow BC's. For such a small domain I don't think mass flow at inlet and outlet will vary by large quantity. (I may be wrong).

Hope this helps.