Learning Forum

[H_E]

Hi, I am using the Eulerian Granular Multiphase Model to simulate the melt flow behaviour of a metal powder filled polymer composite in a nozzle. I managed to get the contour of the pressure, velocity and temperature of the melt similar to that I found from literatures but when I tried to increase the volume fraction of the metal powder with others remain constant, it shows decreasing mixture viscosity and pressure drop, which contradicts with what I learned from the literatures.nDo anyone has any idea of what could be the possible cause of this or any solution to this?nThanks.n

[DrAmine]

Perhaps you add some more description about what you are doing, how are you doing it and what you have learned from the literature.n

[H_E]

I am modelling the melt flow behaviour of a metal powder filled composite FDM 3D printer filament in a 3D printer nozzle. I model it as a laminar flow with 2 phases. The primary phase is the polymer and the secondary phase is the metal powder. I use non-Newtonian-power-law for the viscosity of the polymer. For the boundary conditions, I use velocity-inlet, pressure-outlet and no-slip condition at wall. The initial temperature of the polymer and metal powder is 333K and the wall temperature is 543K. For the solution methods, I use Phase Coupled SIMPLE for pressure-velocity coupling scheme ad second order implicit transient formulation, least square cell based for spatial discretization of gradient, second order for pressure, second order upwind for momentum, and QUICK for volume fraction and energy. From the literature, the viscosity should be increasing with increase in powder volume fraction.nThanks.n

[Aitor]

Hello H_E,nCould you add some attachments with the solutions that you are obtaining? Powder viscosity is greater than polymer viscosity, and viscosity is a volume averaged parameter in multiphase models, so viscosity should increase if powder volume fraction increases.n

[H_E]

Hello Aitor,nBelow are the contour results I obtained with powder volume fraction of 0.6.nBelow are the contour results I obtained with powder volume fraction of 0.5.nThanks.n

[Aitor]

Hello H_E,nI do not understand how do you fix the volume fraction. Do you mean powder volume fraction at the inlet? Could you show the contours of powder volume fraction?n

[H_E]

Hello Aitor,nI fixed it in packing limit in the granular phase setup in multiphase model and at the inlet boundary condition.nn

[Aitor]

Hello H_E,nI have not used the Eulerian Granular Multiphase Model, but I understand that the packing limit is a limitation of the volume fraction of the granular phase based on the validity of the model.nI imagine that the contour of volume fraction in the case of 0.6 (inlet) will be similar than the one of 0.5 (inlet). If this is the case, I do not understand the contour of molecular viscosity. Assuming constant powder viscosity, which is the expression for the viscosity of the polymer? Have you checked if it is correct?n

[H_E]

Hello Aitor,nYes, the contour of volume fraction in the case of 0.6 is similar to that of 0.5.nI'm using non-newtonian power law for the polymer viscosity.nn

[Aitor]

Hello H_E,nI have no experience with the modeling of viscosity for non-newtonian fluids, so I cannot help you with that. However, it may be interesting to start with a constant polymer viscosity, to then switch to advanced models. n

[H_E]

Hello Aitor,nI changed the viscosity to be constant (10000 kg/ms) but the viscosity of the melt with higher powder volume fraction (0.6) still lower than that of 0.5.nFor powder volume fraction of 0.5,nFor powder volume fraction of 0.6,nn

[Aitor]

Hello H_E,nPolymer viscosity is equal to 10.000 Pa*s, and which is the value of powder viscosity?n

[H_E]

Hello Aitor,nThese are what I setup for the 'solids' in multiphase model window.nWhile for the viscosity of 'solids' under material properties, it is not used for the solutions as written in Ansys User Guide. I had tried to modify it and it really shows no difference in the solutions.n

[Aitor]

I see that powder viscosity is not constant. I would recommend you to start with simple cases as constant viscosities, so that the results can be intuitively understood. This would allow you to detect possible errors in the setup.n

[Rob]

Adding material at the packing limit is usually a bad idea as the solver must separate the particles in the model. If you have changed the inlet fraction of solids, have you also changed the polymer mass?

[H_E]

Hello Rob,nI don't really understand.nWhere should I go to change the polymer mass?nThanks.n

[H_E]

Hello Aitor,nI changed the granular viscosity to 1e-5 kg/ms and granular bulk viscosity to 0 but there is no difference in the mixture viscosity result.n

[DrAmine]

Granular run: why looking in the mixture viscosity it does not make a sense any more. You are trying to model an incompressible particle regime with pp collisions via extended KTGF. The viscosity and pressure of the granular phase Atmore just means for stress closure. n