Model dimentsions when using cyclic boundary conditions

RobertHB · August 28, 2017, 07:18

Hello together,

According to de Villiers (2006), when using cyclic boundary condition in a plane channel flow the model dimensions should be restricted to a minimum of 4h*2h*2h length units in stream-wise and span-wise directions. This is supposed to ensure that the turbulence structures of the flow fit into the domain.

Now, my domain resembles a plane channel flow, but is only bounded by one plane (the lower boundary in an implementation of a solid topography) and my top boundary holds a slip condition. Furthermore, due to the scale of my domain (4*2*2 µm) i do not observe any turbulence of the flow.

Question: When using cyclic boundary conditions for my case, do i need to stick to the above given restrictions to achieve a meaningful result, or do the dimension restrictions only apply to simulation which are supposed to simulate turbulence?

With regards,
Robert

FMDenaro · August 28, 2017, 08:41

yes, the dimensions are evaluated for turbulent structures..in your case I think that such a low scale drive to a laminar flow

CeesH · August 28, 2017, 08:54

your flow is certainly going to be laminar at those scales, so there should be no such restriction when solving a periodic domain. In fact, why do you want to CFD this? You could just take a 2-D (or 1-D, if you a have a wide channel) model from what I read, perhaps even an analytic solution.

RobertHB · August 28, 2017, 09:33

Dear CeesH,

the simulation is based on a 3D topography of a ceramic material after a dissolution experiment. The surface itself has a resolution of a few nanometers. We are investigating the near wall flow to determine if, on this scale, the surface roughness has an influence onto the flow. Trapping of the fluid within the pore space due to the formation of eddies is of great interest. At a later stage we aim to calculate the residence of our fluid and combine the information with a chemical model to improve the estimations of our lab experiments.

/edit: And yes, the flow is definitely laminar.

August 28, 2017, 07:18	Model dimentsions when using cyclic boundary conditions	#1
RobertHB Senior Member Robert Join Date: May 2015 Location: Bremen, GER Posts: 292 Rep Power: 12	Hello together, According to de Villiers (2006), when using cyclic boundary condition in a plane channel flow the model dimensions should be restricted to a minimum of 4h2h2h length units in stream-wise and span-wise directions. This is supposed to ensure that the turbulence structures of the flow fit into the domain. Now, my domain resembles a plane channel flow, but is only bounded by one plane (the lower boundary in an implementation of a solid topography) and my top boundary holds a slip condition. Furthermore, due to the scale of my domain (422 µm) i do not observe any turbulence of the flow. Question: When using cyclic boundary conditions for my case, do i need to stick to the above given restrictions to achieve a meaningful result, or do the dimension restrictions only apply to simulation which are supposed to simulate turbulence? With regards, Robert

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Overflow Error in Multiphase Modelling with Two Continuous Fluids	ashtonJ	CFX	6	August 11, 2014 15:32
Water subcooled boiling	Attesz	CFX	7	January 5, 2013 04:32
RPM in Wind Turbine	Pankaj	CFX	9	November 23, 2009 05:05
Water vapour condensation in CFX-5.7.1	hdj	CFX	1	November 27, 2005 08:15
Convective Heat Transfer - Heat Exchanger	Mark	CFX	6	November 15, 2004 16:55

August 28, 2017, 08:41		#2
FMDenaro Senior Member Filippo Maria Denaro Join Date: Jul 2010 Posts: 6,896 Rep Power: 73	yes, the dimensions are evaluated for turbulent structures..in your case I think that such a low scale drive to a laminar flow

August 28, 2017, 08:54		#3
CeesH Senior Member Cees Haringa Join Date: May 2013 Location: Delft Posts: 607 Rep Power: 0	your flow is certainly going to be laminar at those scales, so there should be no such restriction when solving a periodic domain. In fact, why do you want to CFD this? You could just take a 2-D (or 1-D, if you a have a wide channel) model from what I read, perhaps even an analytic solution.

August 28, 2017, 09:33		#4
RobertHB Senior Member Robert Join Date: May 2015 Location: Bremen, GER Posts: 292 Rep Power: 12	Dear CeesH, the simulation is based on a 3D topography of a ceramic material after a dissolution experiment. The surface itself has a resolution of a few nanometers. We are investigating the near wall flow to determine if, on this scale, the surface roughness has an influence onto the flow. Trapping of the fluid within the pore space due to the formation of eddies is of great interest. At a later stage we aim to calculate the residence of our fluid and combine the information with a chemical model to improve the estimations of our lab experiments. /edit: And yes, the flow is definitely laminar.