Hi,

I am simulating the air injected horizontally into the bottomn of a water bath. The bath is like a cone which is about 0.35 m high and the bottom face radius is about 0.14m and upper is 0.25m. This is a obvious multiphase problem.(air and water). Computing domain also include a air cylindar which is 0.5 m in the height attached on the water surface. The diameter of nozzle is 0.002 m and total 7 nozzles (18 degree between each)are installed near the bottom of bath injecting horizontally. OUtlet is at the top of the air cylindar(selected as pressure outlet and gauge pressure is 0). The nozzles are velocity inlet. Mixture model is selcted for the current simulation.(water is primary phase and air is secondary phase). The velocity of air from nozzle outlet is about 10m/s-100m/s depending on differnt conditions. gravity is -9.8m/s along z direction.(height of the system in along z direction). mesh sizes ranges from 0.002 m to 0.02 m. About 250000 grids is obtained. I think the mesh is fine enough for current problem. PRESTO! is used for pressure discretization. Pressure-velocity couping is SIMPLE and others are first-order Upwind.Specified operating density is air's density. For inlet, the water velocity is zero and air is 10m/s and air volume fraction is 1. At outlet, the backflow volumn fraction of air is 1. Under-relaxation factors are reduced as half of default values given by FLUENT.

At first, steady simulation was carried out. The problem is that during the itteration, the reverse flow at the pressure outlet keeps rising and also the water level in the bath also keeps rising almost to half height of the air cylinder, which is totally mad!!.

Postprocessing indicate that the water comes from the reverse flow at the outlet. You know, it is impossible for that.(only blow air from bottom nozzles). It confused me very much. Does anybody have such kind of problems in the similar CFD problems? Thanks! Kim

HI!

I would use VoF multyphase model for that case. You probabilly get wierd results (rising of the water in air cylinder) because you don't use surface tracking with mixture model.

Regards MATEUS

Thanks! Mateus. As a matter of fact, I also tried VOF with IMPLICIT under steady conditions. Same problem took place. Why should water climb so high? We have carried out the water modelling, which indicates that splash will occur while air velocity is very high. I do not know if the problem is connected with too high air injection. Did you simulate such situations when very high velocity gas(>100m/s) is injected into water? Regards, Kim

Obviously, this is an unsteady problem. Change the setting and use VOF. It will work! The nozzle size is 0.002 m, but your mimimum mesh size is also 0.002m. Try to use a smaller mesh size for the domian close to nozzle.

Thanks! Boris. I am trying the unsteady simulation now. The problem is that the time step size is too small.(0.0005s) if I want to struggle for a convergence. As a matter of fact, I am concerning more on the gas plume distribution inside water bath. Accurate prediction surface shape is not so important. I would like to inlove the interaction between gas bubbles and liquid phase. It seems that VOF can not solve such problem. By the way, how about use Mixture model or Euler-euler method to solve the unsdeady problem?