stopping mass flow of droplets in the midst of a simulation

alastormoody11 · June 22, 2010, 05:27

Hi,

I am modeling a spray injection and the injection lasts only for 20ms and the injected particles take 0.14s to traverse the flow domain, to make the simulation realistic as well as to reduce the computational cost(I will be tracking 5million particles instead of 1 lakh particles) I guess I would have to write a user subroutine for this since there is no readymade function for this sort of thing, if anyone has faced a similar sort of problem can they share their code or suggest any other method of achieving this sort of mass flow.

ghorrocks · June 22, 2010, 19:58

? What are you talking about? This is exactly the sort of thing the particle tracking model is in CFX for. And if you have more particles than the particle tracking can handle then CFX also has eularian approaches.

Also note that particle tracking models do not model every particle. The modelled particles represent a population of particles, so you just need enough particles to describe the population. It is unlikely you will need 5 million particles to get adequate resolution of the population.

Either way there should be no need for you to write your own stuff.

alastormoody11 · June 23, 2010, 09:39

Respected Sir,

The problem is that at t=0s the particle injection starts and it ends at t=20milliseconds, however the particles that are injected during this interval at least 0.14s to traverse the control volume, now if I stop the particle injection at t=20ms I track only 100000 particles however if the particles are continually injected I have to track 5000000 particles which I think will overwhelm the system I am working on. Also I realised that 100 particles per time step was too high a number so have diminished the number of particles emitted per time step to 10. However it would be more realistic, and compuationally helpful if I could stop the particles from flowing in at t=20ms I hope that I have made the problem clear and await your reply.

Regards,

Ashish

alastormoody11 · June 23, 2010, 09:44

Also checked out CEL functions for the above purpose they have a step function but I dont think it can be used as it is since it gives some odd values at zero.

ghorrocks · June 23, 2010, 18:36

You don't seem to understand how particle tracking works. It works by injecting a representative sample of particles and tracking them. Each modelled particle represents a large number of particles, not one single particle. This means to model 5M particles you probably only need 2k modelled particles (that's just a guess, you need to do a sensitivity study to work out the true required number) - and each modelled particle represents 500 true particles and therefore has momentum coupling of the full 500 particles.

You can give particles a maximum life so they disappear at a given time or particle lifetime.

As for injection times, yes that is easily done by step functions, if statements or interpolation functions. I have no idea what you mean by "odd values at zero", it works just fine.

June 22, 2010, 05:27	stopping mass flow of droplets in the midst of a simulation	#1
alastormoody11 Senior Member Join Date: Jun 2009 Posts: 100 Rep Power: 16	Hi, I am modeling a spray injection and the injection lasts only for 20ms and the injected particles take 0.14s to traverse the flow domain, to make the simulation realistic as well as to reduce the computational cost(I will be tracking 5million particles instead of 1 lakh particles) I guess I would have to write a user subroutine for this since there is no readymade function for this sort of thing, if anyone has faced a similar sort of problem can they share their code or suggest any other method of achieving this sort of mass flow.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
mass flow rate not conserved in turbomachine, interface defined wrong?	wildli	FLUENT	3	September 15, 2022 12:19
Domain Reference Pressure and mass flow inlet boundary	AdidaKK	CFX	75	August 20, 2018 05:37
Target Mass Flow Rate	Nitin	FLUENT	9	June 17, 2017 10:30
Mass flow rate graph	Lior	FLUENT	0	March 7, 2004 10:02
Determine mass flow rate along flow path	Atit Koonsrisuk	CFX	2	October 5, 2003 04:47

June 22, 2010, 19:58		#2
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,716 Rep Power: 143	? What are you talking about? This is exactly the sort of thing the particle tracking model is in CFX for. And if you have more particles than the particle tracking can handle then CFX also has eularian approaches. Also note that particle tracking models do not model every particle. The modelled particles represent a population of particles, so you just need enough particles to describe the population. It is unlikely you will need 5 million particles to get adequate resolution of the population. Either way there should be no need for you to write your own stuff.

June 23, 2010, 09:39		#3
alastormoody11 Senior Member Join Date: Jun 2009 Posts: 100 Rep Power: 16	Respected Sir, The problem is that at t=0s the particle injection starts and it ends at t=20milliseconds, however the particles that are injected during this interval at least 0.14s to traverse the control volume, now if I stop the particle injection at t=20ms I track only 100000 particles however if the particles are continually injected I have to track 5000000 particles which I think will overwhelm the system I am working on. Also I realised that 100 particles per time step was too high a number so have diminished the number of particles emitted per time step to 10. However it would be more realistic, and compuationally helpful if I could stop the particles from flowing in at t=20ms I hope that I have made the problem clear and await your reply. Regards, Ashish

June 23, 2010, 09:44		#4
alastormoody11 Senior Member Join Date: Jun 2009 Posts: 100 Rep Power: 16	Also checked out CEL functions for the above purpose they have a step function but I dont think it can be used as it is since it gives some odd values at zero.

June 23, 2010, 18:36		#5
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,716 Rep Power: 143	You don't seem to understand how particle tracking works. It works by injecting a representative sample of particles and tracking them. Each modelled particle represents a large number of particles, not one single particle. This means to model 5M particles you probably only need 2k modelled particles (that's just a guess, you need to do a sensitivity study to work out the true required number) - and each modelled particle represents 500 true particles and therefore has momentum coupling of the full 500 particles. You can give particles a maximum life so they disappear at a given time or particle lifetime. As for injection times, yes that is easily done by step functions, if statements or interpolation functions. I have no idea what you mean by "odd values at zero", it works just fine.