I have a question about gravity. I am modeling fume diffusion from chimney to the outer space.I set the gravity(y=-9.81 in two dimention)which i think is necessary as well as the bouyant force.But I can not get a converged solution.Can anybody tell me the reason? Thank you

why you set the gravity in two dimention(-9.81)?

If I understand correctly, you have a 2-D model and have set gravity to 9.81 m/s in the negative Y direction. While failure to converge in a buoyancy problem can have many causes, I would first check the fluid properties. Specifically, are all densities specified a functions of temperature? No density gradient = no driving force = no convergence.

But if there is no temperature difference,that is ,the density is a constant,is it necessary too? Do gravity function too?

Normally, when you have two fluids with different densities, the buoyancy is taking in account even if you have no temperature difference. Do you use a mixing law for yours two fluids ?

HV

Sorry,I can not understand your meaning?I don't use mixing law,neither do know how to use it. hoping for your answer.thank you

If I understand, you want to simulate the mixing of two fluid : flume fluid and air ? How do you simulate fume ?

(1). It seems to me that only professionals should be allowed to "fool around" with the commercial codes. (2). New comers to CFD should only be allowed to ask questions consistent with their level of background training. (3). I always get the feeling that commercial codes are doing more damages to CFD field than providing more useful helps.

Yes,I am simulating the mixing of two fluid:fume and air.I set the multiple species,and in the mixture species,I set the species of the fume.In the boundary condition(velocity inlet),I define the mass fraction of each species. Is there anything wrong or anything unreasonable? Thank you for help.

It is the same way that I used to simulate the mixing of methane and air, so four species (CH4, O2, CO2 and N2).

Do you make a steady-state or a transient simulation ? Is it in an opened or closed space ? What kind of geometry do you use.

I make a steady state simulation,and the space is infinite space actually,but in the gambit,I set the rectangular space(two dimension)and the cube with the size much larger than that of the chimney. how do you simulate?Have you got the satisfied answer? thank you

Do you have a recirculation at your exit boudaries ? because when I worked in steady-state with a concentration of CH4 near to zero at the exit a non-physical recirculation is created which led to no-convergence. This thing appears both in Fluent and Star-CD.

HV

thank you,I'll see and try again. But I have another question:I think the diffusion is let from two reasons:transfer diffusion and molecular diffusion.But my conculsion seems only relative to transfer diffusion instead of molecular diffusion which can reach to the ground.Straightly,the concentration of the ground in my simulation is nealy zero,why? hope for your help.

I don't understand your question.

My question is that I can not get an reasonable solution to my simulation of smoke diffusion from the chimney to the outer space.I think the species of smoke can reach to the ground,but in my solution,the concentration of the species on the ground is nearly zero.If you are doing the same work, I want to know how you set the parameters in fluent. Thank you.

(1). "I think the species of smoke can reach to the ground,but in my solution,the concentration of the species on the ground is nearly zero". (2). Not long ago, I had seen the huge plume formed by the burning of a oil refinery plant near Long Beach on TV, the dark black plume moved up vertically and was carried away by the afternoon wind for a long distance. It also reached the ground level, because I could smell it. (3). The black smoke is normally particles, so obviously these particles will sink to the ground because of the gravitational force. (4). The convection alone will also bring the species from the exhaust to the ground level, because of the near-field flow behavior. (5). It is likely that the boundary conditions are not fine enough for the near field to capture the detailed flow field, or it could be because the model in the code is limited. Most code will give you a solution, but whether it is correct or not normally requires further validation.

thank you Mr.John C. You say that the reason is probably that the model in the code is limited,but in the "gambit",I have to set the limited model instead of unlimited one. Can you tell me in detail how to set boundary conditions more reasonably? thank you again.

(1). Currently, I am not using the code, so I can't advice you how to do it unless I can check it out myself first. (2). The best you can do is to look for a sample case from the vendor, to see whether they can show you that the code has the capability you want.

But Mr. John C,I am in urgent and I can not find the vendor because I get FLUENT from a friend of mine.If I send you the case file of my simulation, can you help me check the code?

(1). I would not have the time and the program to review your input. (2). But you still can talk to your friend and ask him to get related sample from the vendor, if they have it.