Dear Sohail Lone ,

What do you mean by the mesh quality is not good ?

For the simulation a stirred tank reactor using CFD, you can follow this steps and ask the right questions to improve the modelization of studied system :

1- What is the physical domain model ? Cylindrical with flat plate or rounded bottom, with or without baffles ? the mixing impeller type or anchor.

the geometrical characteristics help us to make a choice.

2- What is the software used to this study ? commercial or home made

For the commercial like Comsol Use you can choice different mesh and size like Ansys Fluent, by finite element or finite volume.

For home made software i use Fortan program with a grid generation subroutine to do this according to the geometry of the system.

3- Is the fluid stirred behavior newtonian or not newtonian ?

4- Are the limit layer playing a significant role in the physic phenomenon ? in this case there is special mashing.

welcome for other questions to find out how to help you.

best regards,

houda Baghli.

https://www.sciencedirect.com/search?qs=mesh%20quality%20improvement%20of%20a%20stirred%20tank%20reactor&show=25&sortBy=relevance

Dear Sohail Lone ,

What do you mean by the mesh quality is not good ?

For the simulation a stirred tank reactor using CFD, you can follow this steps and ask the right questions to improve the modelization of studied system :

1- What is the physical domain model ? Cylindrical with flat plate or rounded bottom, with or without baffles ? the mixing impeller type or anchor.

the geometrical characteristics help us to make a choice.

2- What is the software used to this study ? commercial or home made

For the commercial like Comsol Use you can choice different mesh and size like Ansys Fluent, by finite element or finite volume.

For home made software i use Fortan program with a grid generation subroutine to do this according to the geometry of the system.

3- Is the fluid stirred behavior newtonian or not newtonian ?

4- Are the limit layer playing a significant role in the physic phenomenon ? in this case there is special mashing.

welcome for other questions to find out how to help you.

best regards,

houda Baghli.

With the amount of information provided, it's difficult to give a complete answer - can you give some geometry details?

Anyway, if I am correct from previous posts, you use ansys. For a standard Rushton tank: extend the shaft all the way through the center, cut the remaining domain vertically into at least 2 zones, and you can use sweep meshing to make a high-quality, full hexahedral mesh. Do note that with this you cannot add inflation layers and such near the wall; it's great for bulk flow, but perhaps not the best if you are interested in wall details. Also, for a Rushton turbine, don't resolve wall thickness. (see e.g. Coroneo e.a. 2011, Gunyol e.a. 2009) . Using a 2D impeller gives much better predictions for energy dissipation throughout the domain.

For unstructured impellers: convert tetrahedral mesh to polyhedral. Much higher quality, generally lower computation time.

Good luck!