Firstly, in my simulation, the inlet solid mass loading is 0.0533,and the solid volume fraction is 2.36e-5. I think in this condition the particles should feed back the gas field, so I use the two- way coupling Discrete Particle Model in Fluent. I know that the influence of the particles on the gas flow field is considered by the momentum transfer of each particle while conveyed through each control volume in Fluent model. In order to enhance this influence, I use the numerical particles, i.e. I multiply the drag coefficience CD with a factor much bigger than 1(for example 100000) by UDF using DEFINE_DPM_DRAG, and change the inlet mass loading by changing this factor. This implies that I view the individual numerical particles as parcels. However, after doing this, when the calculation is conveyed, the velocity of gas phase is not reduced greatly. I want to know whether this method is right or not, and what leads to this? Secondly, I know that Fluent provides us the Discrete Phase Model that don't take inter-particle collision into account, because Fluent considers the inlet solid volume fraction is too low to care for it. However, in my model, I want to extend the range of the solid volume fraction, so I have to consider the inter-particle collision. Concerning particle-particle collision, two models are known in general. One is the hard sphere model and the other is the soft sphere model. In my calculation, I will use the hard sphere model. The hard sphere model is based on the integrated forms of the equations of motion which are called impulse equations£¬and thus instantaneous deformation of the particle does not appear explicitly in the formulation The relation between the pre- and post-collision velocities is obtained using the coefficient of restitution e and the friction coefficient f. When the particulate phase is disperse£¬it is sufficient to consider only binary collisions and not multiple collisions. In such cases the hard sphere model is usually used because it is easy to use. In this model, we have to deal with the collision probability and so on. I know that Fluent can do this at least can add a force which base on the collision to the Fluent model, but I think it is difficult to take these into account in UDF, meanwhile, I am not familiar with the Fluent UDF. So I hope you can give me some advice about this UDF. Lastly, after the simulation is over, I want to display particles number / unit volume about particle concentration. When I display \contours\ discrete phase model\dpm-concentration, however it will give me the results in kg/m3. I want to know which macro can change this in Fluent. Thanks a lot for your considering my questions. Best regards.

If your particles all have the same mass, then the number concentration is proportional to the mass concentration, and you could write a simple custom field function.

As for the whole collision thing, I think it's probably possible, but it's gonna be a serious piece of work. The trick will be locating when collisions occur, particularly if the size of the particle is small compared to the size it can move in any given time step, and/or the frequency of collisions is high.

Good luck

Rob

"inlet solid mass loading is 0.0533,and the solid volume fraction is 2.36e-5" You dont neccesary have a two-way coupling at those conditions. The mass loading is actually not very high, it is in the upper range of one way coupling or lower range of two-way. The coupling also depends on the Stokes number, i.e. the the particle response time (depends on the density of your particles, diameter, viscosity, etc.) and a global time scale. I do not understand why you multiply the drag with a factor ...., why is it neccesary? Regards Jan

Thanks for your answering my questions. To Rob Unfortunately in my simulation, particles have Rosin-Rommler distribution, so I don¡¯t know how to deal with this. In addition, as for collision, I only consider particles collision between the current cell and its adjacent ones, and each cell has one or two particles, because I use the numerical particles (i.e. one numerical particle stands for many particles). My question is that how I can get the particles¡¯ information which is in the adjacent current cell. Hope you can give me some advice, thanks very much. To Jan I know as for my question, sometimes one-way coupling is enough, but I want to extend the range of my question, so I have to use the two-way coupling. Elghobashi (1994) has mapped the interaction between particles and turbulence by means of two characteristic dimensionless quantities, viz. the volume fraction of particles, and the ratio of the particle response or relaxation time, and the Kolmogorov time scale. When the volume fraction of particles is 2.36e-5, the momentum transfer from the particles is large enough to alter the turbulence structure, so I have to consider the two-way coupling. After considering the two-way coupling, the velocity of gas phase should be damped, although the turbulence intensity may be not change. I don¡¯t know why my simulation has not this result. Why do I multiply the drag with a factor much bigger than 1? I think I can view the individual numerical particles as parcels and enhance the influence of the particles on the gas flow field by means of this method. Maybe not right, welcome you give me some advice. Thanks in advance. Best regards to you, fpingqian