I have had a customer ask if we could model the flow of activated carbon through a duct in a coal fired power plant. They want to verify the particle distribution as well as optimize the injection grid. From expereince, I know that when I have tried to insert massed particles into a model, the results have been subject to question. This material is very fine with a bulk density of 32 lbs/ft^3.
Has anyone had any experience modeling a fluid solid system and then injecting massed particles and end up with good, substantial results? If so, any tips would be appreciated.
The injection rates are relatively low, so I think I can make it work. The customer also asked if we could quantify the dispersion of the particles across the outlet. Other than image capture, I am not aware of any other way to measure the output of massed particle injection. Any feedback on this?
From what I can guess, you'd be interested in how the particles are dispersed in the domain, and how the distribution is on your outlet.
From the documentation, one can tell that for the Langrangian equations of position are solved along with the particle velocity (using Newton's second law). The gravity and drag forces are also covered in momentum equations. But other forces that make model more comprehensive - like turbulent dispersion, lift, virtual mass force etc don't find any mention anywhere in the documentation. So there seem to be limitation. Most notably, it seems to be a one way coupling, such that only fluid imparts momentum to the solids, not the other way around.
But ths is just an informed guess. Autodesk personnel may add more here.
It will really depend on what exactly want to see from the simulation.
The injection rate into the duct stream is relatively low. I am not worried about the fluid flow being based on the air only.
I am primarily concerned with penetration depths, reactions around elbows/turning vanes etc.
I have had to deal with low density 2 phase flows more recently. I like the CFDesign/Simulation interface and would rather not have to shop for a new software package, but I have had a hard time believing the massed particle streams from CFDesign/Simulation.
The above comments are correct.
When you are modeling mass particles traces you do need to more careful on what your timestep is when you setup the massed particles. If you make that timestep too large then you particles travel will not be believed.
We have had a lot of success with user using the mass particles traces to simulate various forms of dust and other small particles moving through a system.
In terms of getting a distribution of your massed particles at the outlet, there isn't a great workflow to output the location of the particle at the exit plane. Using a screen shot showing the distribution at the exit plane and comparing your results it probably the workflow you will need to use as you mentioned.