[OzanF.]

Hi there,

I'm modelling an intake manifold for my universities formula student car, the inlet has constant V,P for this model, but outlet conditions, P's are changing with time. I have 4 outlet (1 open and 3 closed) and the open one must change by the time. (EX: t:1; Outlet1pen, Outlet2,3,4:close/ t:2; Outlet2pen, Outlet1,3,4:close etc.)

Can i model this situation and observe backflow and mass flow values.

Thanks.

[ghorrocks]

What is constant V,P at the inlet?

I have done inlet manifold modelling successfully by putting an opening at atmospheric pressure at the inlet trumpet, then using mass flow boundaries at each valve. You can then define a mass flow versus time for each valve which includes no flow when they are shut.

[OzanF.]

My opening condition is also atmospheric pressure but i need a pressure versus time for each valve (0.85*P(atm), vacuum ), and also i need the same no flow conditions for the closed valves. Can i make mass flow versus time for each valve defining not the mass flow but the pressure for the opened one, and no flow for the closed valves?

[ghorrocks]

The pressure at the valves is generally not easy to know, but the mass flow is more easily known. That is why I recommend using a mass flow boundary, not a pressure boundary at the valves. It also easily takes care of the times when the valve is closed by defining zero flow.

If you insist on using pressure boundary then you will have to stop and restart the simulation to a new simulation with a wall for the closed valve. You might also be able to avoid this by putting a momentum source in front of the pressure boundary which sets the flow to zero when the valve is shut. But this is all pretty horrible, this is much better done by defining the mass flow.

[OzanF.]

The reason why i insist pressure boundary is actually i have the pressure versus time values of the valves, and i want to see the value of mass flow at each valves. I'm also thinking the stop and restart the simulation way but i'm not sure because by that way i wan't see the effects between valves (effects of the first flow on the second valve),can i?

[ghorrocks]

If you have the pressure then you can use a pressure boundary. You should at least try the massflow approach as you can usually make reasonable estimates of the massflow from the piston motion.

Stopping and restarting will still include the valve to valve effects. When you restart with a valve now open or shut then you use the old run as the initial condition so the flow variables are mapped over from the previous run.

[OzanF.]

Actually i made a similar thing but my main problem is i couldn't understand how to use the old run as the initial conditions. Here's what i do.

First i made flow analysis 1,(transient flow 0-100s, times step:1s(normally this period is too long but a made this to test the model) in which inlet is openning with atmospherical pressure, outlet 1 is 0.85 Pa,outlet2,3,4 are closed(wall), and initial conditions are 0 for pressure and velocity

Than i insert another flow analysis 2, (transient 100-200s, time steps 1s) in which openning is the same but outlet 2 is 0.85 pa and the rests are closed, for the initial conditions of this analysis my only option is to use the same conditions i could't find how to use old conditions.

And than i add 2 configurations, configuration 1 has the activation control 1(start of the simulation), configuration 2 has the activation control 2 ( end of the configuration- configuration 1), i tried to use this multiconfiguration system to model "use the old run as initial conditions of the second", but i'm not sure about that because cfd-post is not showing flow analysis 1, when i solve this. I observe just the second flow.

Thanks

[ghorrocks]

No, for analysis 2 you use the results from analysis 1 as initial conditions. Do not use the configuration.

If all you know about the pressure is that it is 0.85*atmospheric then can I suggest this is so wildly inaccurate that you should not use it. Any true IC engine manifold has all sorts of high frequency variations and for IC engines these are very important. If you do not have a pressure versus crank angle trace with at least 1°CA resolution then your pressure will be wrong.

So I recommend you to go to the massflow approach I have been talking about all along. You can define a massflow versus time curve from the piston motion. And you can use your 0.85*atmospheric pressure as a check that your results are correct.

[OzanF.]

Ok i will try the mass flow approach but i still dont know how to use the results of the first as the second's initial conditions without configurations, are you making this from the run definiton by creating an initial value which is equal to first step? Or is there another way, if there is another way is there any tutorial about it?

thanks

[ghorrocks]

You simply use the previous simulation as an initial condition for the next. This is covered by the tutorials.

But with the mass flow approach you do not need this. You can just define zero mass flow when the valves are shut and it works fine.

[Santoshnalanagula]

Originally Posted by ghorrocks

You simply use the previous simulation as an initial condition for the next. This is covered by the tutorials.

But with the mass flow approach you do not need this. You can just define zero mass flow when the valves are shut and it works fine.

Hi
You mean giving outlet condition as mass flow condition ? Is it posibile in Fluent ?
I have similar problem one inlet and two outlet ( twin cylinder intake manifold ) I want to impose outlet condition either mass vs crank or velocity vs crank or pressure vs crank. i have transient data from GT power. How can i create opening and closing condition in outlet wall ?

[ghorrocks]

If you are using Fluent then try the Fluent forum.