|
[Sponsors] |
January 16, 2020, 12:32 |
Flow direction problem
|
#1 |
Senior Member
Join Date: Aug 2012
Posts: 268
Rep Power: 14 |
I know from Fluid Mechanic course that flow direction should be from high pressure to low pressure.
In the attached image, the right area has lower pressure than the left area. Why doesn't the flow direction in this image obey this rule? Absolute pressure and meridional velocity vectors are time-averaged. Flow direction is from left to right. |
|
January 16, 2020, 12:56 |
|
#2 |
Senior Member
Erik
Join Date: Feb 2011
Location: Earth (Land portion)
Posts: 1,167
Rep Power: 23 |
Flow doesn't always flow from simply just higher pressure to lower pressure. There is much more involved. Shear forces are included in the Navier stokes equations as well.
Think about a rotating vortex, or any recirculating flow pattern. It would be impossible for this to always flow from higher to lower pressure, ad the flow comes back to its starting point. |
|
January 16, 2020, 13:12 |
|
#3 |
Senior Member
Join Date: Jun 2009
Posts: 1,804
Rep Power: 32 |
Also, be careful interpreting time averaged flows. Context is very important.
It we setup two parallel plates with an inlet and outlet and impose a sinusoidal inlet condition with zero mean flow, the time average plot in the domain will show 0 velocity everywhere. Does it mean there is no flow throughout the simulation? |
|
January 17, 2020, 06:48 |
|
#4 | |
Senior Member
Join Date: Aug 2012
Posts: 268
Rep Power: 14 |
Quote:
How does CFX calculate time-averaged velocity vectors or pressure contours for a transient flow? Does CFX calculate time-averaged result as instantaneous value minus fluctuation or perturbation part or it is found by evaluating the integral over a dt? |
||
January 17, 2020, 08:39 |
|
#5 |
Senior Member
Join Date: Jun 2009
Posts: 1,804
Rep Power: 32 |
There is only one definition for time average
Time averaged value of < > = Integral over time interval of ( < > dt) / (time interval) The special case of equal time step for the simulation, the equation simplifies to Time averaged value of < > = sum of ( weights * time values) if the weights are the ones for the trapezoidal rule, Time averaged values of < > = Sum of time values / Number of time values. i.e. arithmetic average. |
|
January 17, 2020, 12:13 |
|
#6 |
Senior Member
Join Date: Aug 2012
Posts: 268
Rep Power: 14 |
Can I ask in what situations a time-averaged result should not be used?
Last edited by Julian121; January 17, 2020 at 23:43. |
|
January 19, 2020, 03:16 |
|
#7 |
Super Moderator
Glenn Horrocks
Join Date: Mar 2009
Location: Sydney, Australia
Posts: 17,703
Rep Power: 143 |
Also keep in mind the Bernoulli equation - high velocity flow has low pressure. So if you look at static pressure alone you can get a misleading picture about what is driving the flow. Dynamic pressure is a better one to look at for a pressure driven flow.
__________________
Note: I do not answer CFD questions by PM. CFD questions should be posted on the forum. |
|
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
Problem with FsiFOAM simulation of beams (2-4 beams) in a steady simple shear flow | Aliiiii | OpenFOAM Running, Solving & CFD | 1 | February 27, 2019 12:26 |
Review: Reversed flow | CRT | FLUENT | 1 | May 7, 2018 05:36 |
Back flow problem in gas cyclone | lakhi | FLUENT | 0 | August 31, 2012 04:27 |
Periodic flow boundary condition problem | sudha | FLUENT | 3 | April 28, 2004 08:40 |
fluid flow fundas | ram | Main CFD Forum | 5 | June 17, 2000 21:31 |