many question(help)

mehrdadeng · April 1, 2009, 04:41

hi to all

i am beginner in cfx.
i am simulating turbo charger of automobile model wrought from blade & shell.i have some problem in this process please help me.

1- i think blade domain is solid but i cant define rotating condition for solid domain in general mode of cfx.

2-how i can get variable rate from variable inlet boundry condition??

3-how i can apply variable rate results to blade in next step??

best regard

mehrdadeng · April 1, 2009, 16:43

are there any remark???

ghorrocks · April 2, 2009, 20:57

Hi,

1) Do you need to model the solid domain? Unless heat transfer or deflections are important then don't model it.
2) You can specify pretty much any boundary condition parameter as a CEL expression. CEL can be either a function (pressure=1+2*temperature+3*X for instance) or an interpolation function.
3) I don't understand this question.

Glenn Horrocks

mehrdadeng · April 3, 2009, 09:17

Quote:

Originally Posted by ghorrocks

Hi,

1) Do you need to model the solid domain? Unless heat transfer or deflections are important then don't model it.
2) You can specify pretty much any boundary condition parameter as a CEL expression. CEL can be either a function (pressure=1+2*temperature+3*X for instance) or an interpolation function.
3) I don't understand this question.

Glenn Horrocks

thanks,ghorrocks

3)BC is variable,so the torque exerted to blade will be variable.i shoud calculate rate of blades from variable torque & apply this rate to blade in next step of simulation.because of variable torque ,the rate of blade will be variable too.now i dont know that how i can apply variable rate to blades??.
my english is not well ,excuse me

thanks for reply

ghorrocks · April 6, 2009, 01:54

hi,

I still don't understand your question but I will have a guess at what you are asking.

If you mean that you need to find the operating speed for a given flow condition, the best way to do this is to assume some speeds and plot torque imbalance versus speed. Torque imbalance is the net torque on the rotor. Interpolate to where the torque imbalance is zero and that's your steady state operating condition.

Glenn Horrocks

yochule · June 1, 2009, 16:16

mehrdadeng , ghorrocks i think you have an answer for my questions!
i have trying to do this! i'm investigating a hairdryer fan. i want to optimize the geometry and i need to specify a motor (torque - rpm) for a desired geometry design.
i think, about ghorrocks post, i need to run a geometry1 in deferents rpm and mass flow, and evaluating the torque invalance. i will be able to determine the point of work as a whole pieces!
i redesign the geometry and run again with geometry2, different rpm and flow rate, and find a new point work...
so i can compare the geometry1 to geometry2 performance and decide which is the best...
is this correct?

thanks!

ghorrocks · June 1, 2009, 18:27

Hi,

Yes, that is it. You are trying to model enough points on the pressure versus flow rate curve so you know what it looks like. You can then find the operating point where the motor torque equals the fans torque + losses (for steady state running).

Glenn Horrocks

yochule · June 2, 2009, 11:49

glenn, i have a new doubt.
when i set the convergence criteria i can set the invalance to zero, so i could obtain zero torque of imbalance in each run, so: how do i know if the boundary conditions that i set are the correct? i mean, if i set a 10 m3/h (flow rate) and 1000 rpm, how can i know if this is the flow rate that i really i would have with 1000 rpm in a test?
i can´t make out if with this sets i am forcing a flow pattern that is not the same that i would obtain in one experiment.

regards!

ghorrocks · June 2, 2009, 18:47

Hi,

There are three main parameters to be determined - rotor speed, flow rate and pressure change. Generally you know one parameter, in this case it looks like you know the flow rate is 10 m3/h. So using this flow rate you can do a series of pressures (or rotor speeds) which give you a pressure versus rotor speed line.

You then establish the system line for pressure versus speed and where the two lines intersect is the operating point - that is the speed and pressure difference where everything is balanced.

Glenn Horrocks

yochule · June 2, 2009, 22:24

Suppose I wan to know if a designed geometry can reach a desired flow rate.
So I run one set up with this (desired) flow rate, and a proposed rpm, then using the imbalance of torque I will know if this rpm is the correct. If not, I would need to run the same flow rate, but with different rpm, and see the torque imbalance again, and so on. In the end I will have one point with this flow rate and rpm.
If I need to obtain the rpm vs flow rate curve (for one geometry) I need to exercise several run like the described above to have point enough.

ghorrocks · June 2, 2009, 23:55

Hi,

Yes, that's the idea.

Glenn Horrocks

yochule · July 29, 2009, 16:39

how can i determine the invalance?
the torque (tool--function calculator--torque) is the same of the net torque? (shaft - vs - load)
i'm not sure about this point..

ghorrocks · July 29, 2009, 18:53

Hi,

I assume you are analysing the compressor side of the turbocharger. If you are doing the turbine side just swap the torque signs around on the following comment.

The torque absorbed by the compressor rotor can be determined by the torque()@ function. That will give you the CFX predicted torque at the pressure, flow and rotor speed you modelled. You then look at the turbine operating curve and see what torque the turbine produces at the speed and operating conditions. If the turbine is producing more torque than the compressor is absorbing then the operating speed is faster, so redo the simulation at a faster speed. If the turbine is producing less torque redo the simulation at a lower speed.

Keep iterating this process until you have balanced the torque to the level of accuracy you desire.

Glenn Horrocks

yochule · July 30, 2009, 10:06

hi all...
i'm modeling a rotor-stator with a DC motor driven the rotor.
so i need to compare the torque in the rotor (tool--function calculator--torque) to the torque shaft to this RPM in the data sheet of the motor? what happened if i had not this information?
there is not a direct way to determine the equilibrium between rpm-delta P-flow rate?

ghorrocks · July 30, 2009, 21:15

Hi,

The equilibrium is established between the motor producing torque and the rotor absorbing it. Obviously you cannot know the operating point without knowing the torque of both of them.

There are some exceptions - for instance if the torque curve of the motor is very flat (ie does not change with speed significantly) then all you need to do is find at what speed the rotor absorbs that much torque. But whether this is true of not will depend on the type of motor, your operating point and many other factors. That's why engineering is challenging.

Glenn Horrocks

yochule · July 31, 2009, 11:27

so, i need the motor data sheet, and to measure the delta P?
and i can make rpm-flow rate curves.
i´m confused!

yochule · August 3, 2009, 12:14

there is no problem to me at all because i have the data sheet of the motor, and i can measure the pressure loss (because i have a prototype of these rotor/stator)
but, if i had not this pieces to measure or had not the data sheet of the motor, how can i simulate this?
if i decide to modify some dimensions of the rotor (o the stator) o to change the motor, how can i simulate a piece that i have not information (except the geometry)?

can anybody understand me?

yochule · September 28, 2009, 14:18

hello guys...
i´m back studyng this problem

. i don't understadn two points:
1 - i would be able to make a curve of rotor-shaft vs cfx torque & rpm. but i can't determine the pressure of the equilibrium..
2 - will i need to measure the pressure dropp in the system to fix it? if i had not the pieces, how i must to continue

?
i attached the points i can reach in the run vs the motor data sheet.

ghorrocks · September 28, 2009, 19:13

So you now need an additional piece of information to fix the remaining variable. It could be the pressure, the flow rate, whatever. You just need one more piece of information to fix it to a point.

It may be easier to get the flow rate from the engine characteristics.

yochule · September 28, 2009, 21:00

yes, i can measure the flow rate with a rpm set. but this is because i have the prototype. how can i predict the flow rate with another prototype?
if i want to change some geometry characteristics, i will not have the same flow rate for the same rpm, or not? how can i be sure that this last change will be possitive or negative?
sorry about this large talks, but i really cant understand how to predict the real caracteristics in base of the cfd.
it is clear to me that we need to measure some real thing to be sure of our cfd, but there be some thing that i had missing in this procedure...

April 1, 2009, 04:41	many question(help)	#1
mehrdadeng Member mohsen Join Date: Mar 2009 Posts: 53 Rep Power: 17	hi to all i am beginner in cfx. i am simulating turbo charger of automobile model wrought from blade & shell.i have some problem in this process please help me. 1- i think blade domain is solid but i cant define rotating condition for solid domain in general mode of cfx. 2-how i can get variable rate from variable inlet boundry condition?? 3-how i can apply variable rate results to blade in next step?? best regard Last edited by mehrdadeng; April 1, 2009 at 13:06.

June 1, 2009, 16:16	very interesting!!	#6
yochule Member nazareno mancinelli Join Date: Mar 2009 Location: argentina Posts: 35 Rep Power: 17	mehrdadeng , ghorrocks i think you have an answer for my questions! i have trying to do this! i'm investigating a hairdryer fan. i want to optimize the geometry and i need to specify a motor (torque - rpm) for a desired geometry design. i think, about ghorrocks post, i need to run a geometry1 in deferents rpm and mass flow, and evaluating the torque invalance. i will be able to determine the point of work as a whole pieces! i redesign the geometry and run again with geometry2, different rpm and flow rate, and find a new point work... so i can compare the geometry1 to geometry2 performance and decide which is the best... is this correct? thanks!

June 2, 2009, 11:49	question	#8
yochule Member nazareno mancinelli Join Date: Mar 2009 Location: argentina Posts: 35 Rep Power: 17	glenn, i have a new doubt. when i set the convergence criteria i can set the invalance to zero, so i could obtain zero torque of imbalance in each run, so: how do i know if the boundary conditions that i set are the correct? i mean, if i set a 10 m3/h (flow rate) and 1000 rpm, how can i know if this is the flow rate that i really i would have with 1000 rpm in a test? i can´t make out if with this sets i am forcing a flow pattern that is not the same that i would obtain in one experiment. regards!

June 2, 2009, 22:24	So It stem's that I was wrong!	#10
yochule Member nazareno mancinelli Join Date: Mar 2009 Location: argentina Posts: 35 Rep Power: 17	Suppose I wan to know if a designed geometry can reach a desired flow rate. So I run one set up with this (desired) flow rate, and a proposed rpm, then using the imbalance of torque I will know if this rpm is the correct. If not, I would need to run the same flow rate, but with different rpm, and see the torque imbalance again, and so on. In the end I will have one point with this flow rate and rpm. If I need to obtain the rpm vs flow rate curve (for one geometry) I need to exercise several run like the described above to have point enough.

July 29, 2009, 16:39	calculating invalance	#12
yochule Member nazareno mancinelli Join Date: Mar 2009 Location: argentina Posts: 35 Rep Power: 17	how can i determine the invalance? the torque (tool--function calculator--torque) is the same of the net torque? (shaft - vs - load) i'm not sure about this point..

April 1, 2009, 16:43		#2
mehrdadeng Member mohsen Join Date: Mar 2009 Posts: 53 Rep Power: 17	are there any remark???

April 2, 2009, 20:57		#3
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,700 Rep Power: 143	Hi, 1) Do you need to model the solid domain? Unless heat transfer or deflections are important then don't model it. 2) You can specify pretty much any boundary condition parameter as a CEL expression. CEL can be either a function (pressure=1+2temperature+3X for instance) or an interpolation function. 3) I don't understand this question. Glenn Horrocks

April 6, 2009, 01:54		#5
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,700 Rep Power: 143	hi, I still don't understand your question but I will have a guess at what you are asking. If you mean that you need to find the operating speed for a given flow condition, the best way to do this is to assume some speeds and plot torque imbalance versus speed. Torque imbalance is the net torque on the rotor. Interpolate to where the torque imbalance is zero and that's your steady state operating condition. Glenn Horrocks

June 1, 2009, 18:27		#7
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,700 Rep Power: 143	Hi, Yes, that is it. You are trying to model enough points on the pressure versus flow rate curve so you know what it looks like. You can then find the operating point where the motor torque equals the fans torque + losses (for steady state running). Glenn Horrocks

June 2, 2009, 18:47		#9
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,700 Rep Power: 143	Hi, There are three main parameters to be determined - rotor speed, flow rate and pressure change. Generally you know one parameter, in this case it looks like you know the flow rate is 10 m3/h. So using this flow rate you can do a series of pressures (or rotor speeds) which give you a pressure versus rotor speed line. You then establish the system line for pressure versus speed and where the two lines intersect is the operating point - that is the speed and pressure difference where everything is balanced. Glenn Horrocks

June 2, 2009, 23:55		#11
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,700 Rep Power: 143	Hi, Yes, that's the idea. Glenn Horrocks

July 29, 2009, 18:53		#13
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,700 Rep Power: 143	Hi, I assume you are analysing the compressor side of the turbocharger. If you are doing the turbine side just swap the torque signs around on the following comment. The torque absorbed by the compressor rotor can be determined by the torque()@ function. That will give you the CFX predicted torque at the pressure, flow and rotor speed you modelled. You then look at the turbine operating curve and see what torque the turbine produces at the speed and operating conditions. If the turbine is producing more torque than the compressor is absorbing then the operating speed is faster, so redo the simulation at a faster speed. If the turbine is producing less torque redo the simulation at a lower speed. Keep iterating this process until you have balanced the torque to the level of accuracy you desire. Glenn Horrocks

July 30, 2009, 10:06		#14
yochule Member nazareno mancinelli Join Date: Mar 2009 Location: argentina Posts: 35 Rep Power: 17	hi all... i'm modeling a rotor-stator with a DC motor driven the rotor. so i need to compare the torque in the rotor (tool--function calculator--torque) to the torque shaft to this RPM in the data sheet of the motor? what happened if i had not this information? there is not a direct way to determine the equilibrium between rpm-delta P-flow rate?

July 30, 2009, 21:15		#15
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,700 Rep Power: 143	Hi, The equilibrium is established between the motor producing torque and the rotor absorbing it. Obviously you cannot know the operating point without knowing the torque of both of them. There are some exceptions - for instance if the torque curve of the motor is very flat (ie does not change with speed significantly) then all you need to do is find at what speed the rotor absorbs that much torque. But whether this is true of not will depend on the type of motor, your operating point and many other factors. That's why engineering is challenging. Glenn Horrocks

July 31, 2009, 11:27	The more I learn, I realize that less is	#16
yochule Member nazareno mancinelli Join Date: Mar 2009 Location: argentina Posts: 35 Rep Power: 17	so, i need the motor data sheet, and to measure the delta P? and i can make rpm-flow rate curves. i´m confused!

August 3, 2009, 12:14		#17
yochule Member nazareno mancinelli Join Date: Mar 2009 Location: argentina Posts: 35 Rep Power: 17	there is no problem to me at all because i have the data sheet of the motor, and i can measure the pressure loss (because i have a prototype of these rotor/stator) but, if i had not this pieces to measure or had not the data sheet of the motor, how can i simulate this? if i decide to modify some dimensions of the rotor (o the stator) o to change the motor, how can i simulate a piece that i have not information (except the geometry)? can anybody understand me?

September 28, 2009, 19:13		#19
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,700 Rep Power: 143	So you now need an additional piece of information to fix the remaining variable. It could be the pressure, the flow rate, whatever. You just need one more piece of information to fix it to a point. It may be easier to get the flow rate from the engine characteristics.

September 28, 2009, 21:00	confused	#20
yochule Member nazareno mancinelli Join Date: Mar 2009 Location: argentina Posts: 35 Rep Power: 17	yes, i can measure the flow rate with a rpm set. but this is because i have the prototype. how can i predict the flow rate with another prototype? if i want to change some geometry characteristics, i will not have the same flow rate for the same rpm, or not? how can i be sure that this last change will be possitive or negative? sorry about this large talks, but i really cant understand how to predict the real caracteristics in base of the cfd. it is clear to me that we need to measure some real thing to be sure of our cfd, but there be some thing that i had missing in this procedure...