problem moment on vertical axis wind turbine

[M1000]

Hello everyone, I've been working with gambit and fluent for my university thesis.
I am performing cfd analysis on a vertical axis wind turbine.
the problem is that I get values for the moment along the unacceptable z axis.
in practice for a 9m/s wind I get a moment of 4155 nm.
multiplying for 120*2*pi/(60), since the trubine has rotation speed of 120 rpm, I get:

pe=extracted power=52213 w

against a available power given by the report:

♪

in which air density is 1.225 kg/m^3) for ambient temperature
v is speed
a is the front surface of the turbine.

in my case a=9m*5m=45m^2
v=9m/s
and then I get:

pw=20093 w

clearly I can't extract more power than that available from the wind.

the data is even more evident considering the cp power coefficient of the turbine, defined as:

cp=pe/pw

his theoretical masimo limit according to betz theory is 0.59, and no turbine can overcome it.

for vertical axis wind turbines as mine should be 0.26<cp<0,37. -="" che="" con="" condizioni="" cp="52213/20093=2,59." define="" di="" fluent="" ho="" i="" inserito="" io="" le="" non="" prova="" quadra.="" qualcosa="" quindi="" sicuramente="" sono:="" sopracitati="" su="" un="" valori=""> model -> solver: pressure based solver, unsteady
define -> model -> viscous: k-epsilon standard
define -> model -> energy equation on

boundary condition:
definition -> boundary condition
velocity inlet: velocity specification method -> components, reference frame absolute
the zone of the domain in which the turbine is inserted (fluid zone): motion tipe -> moving mesh, rotational velocity -> 120 rpm
turbines (wall): wall motion -> moving wall, motion -> relative to housing cell zone.

solve:
solve -> control -solution
urf(under-realaxation factors) the values that fluent gives
pressure-velocity coupling -> simple
discretization: pressure -> early
all values on secon.order upwind

solve -> monitors -> residuals:I left the values indicated plotting the window
solve -> monitors -> force: options: print, plot and write
wall zones: turbines
coefficient: moment
moment center (0.0.0) turbine has rotation axis in origin

solve -> initialize -> initialize: all zones

solve -> iterate

as time step size I entered 0,005556s or corresponding to 4 degrees of rotation, while as time step number 1080 or 12 rotations.
the data of 4155nm is reported propio to the last time step.

What's wrong?
Maybe I'm wrong to interpret the data?</cp<0,37.>

 

[Franz 2.0]

use fluent without doing physical examination 1?

 

[M1000]

I did, but if you could clarify where the problem is, it would be better... .

 

[Drag]

but of the turbine you know the real data, ie you know for certain that with 9 m/s to 120 rpm absorbs that power?

 

[M1000]

I do not have real data, but realistically the turbine cannot absorb more energy than that provided by the air mass, and as I wrote in the first post for speed of 9m/s the wind has a power equal to:

pw=1/2*rho*v^3*a=20093 w

while from fluent reports along the z axis I have a moment of 4155nm multiplied by 120 rpm, clearly expressed in rad/s (120*2*pi/60) provide me with a power of

pe=52213w.

So either I solved the problem of world energy supply or there is something wrong.

 

[Drag]

pw=1/2*rho*v^3*a

in this formula only the speed v :
then I wonder if instead of 120 rpm put 10 rpm and repeat the calculation with fluent
the absorbed pair changes, but not the v=9 m/s.
will not be that 120 rpm is not a proper data?

 

[M1000]

in the formula the other characteristic is the frontal surface a.

definitely changing the rotation speed the absorbed torque changes,
120 rpm could also be wrong, and at this point how do I establish the correct rotation speed?

 

[Drag]

you have to do a test with different rpm and see if the absorbed pair rises or falls, then go for iteration.
Maybe if the mesh is very thick you should work on a lighter model to do first, then take back the good one.
greetings

 

[M1000]

Okay, I launched a 60 rpm simulation, let's see what happens.
Unfortunately I have to launch it with the finished mesh model, because for example changing only the parameters of discretization for wind to 5m/s I have a remarkable difference, that is, for discretization to the first order 6100 nm, for discretization to the second order 3800nm.
Anyway, thank you for the advice.

 

[paolina]

Hello everyone, I've been working with gambit and fluent for my university thesis.
I am performing cfd analysis on a vertical axis wind turbine.
the problem is that I get values for the moment along the unacceptable z axis.
in practice for a 9m/s wind I get a moment of 4155 nm.
multiplying for 120*2*pi/(60), since the trubine has rotation speed of 120 rpm, I get:

pe=extracted power=52213 w

against a available power given by the report:

♪

in which air density is 1.225 kg/m^3) for ambient temperature
v is speed
a is the front surface of the turbine.

in my case a=9m*5m=45m^2
v=9m/s
and then I get:

pw=20093 w

clearly I can't extract more power than that available from the wind.

the data is even more evident considering the cp power coefficient of the turbine, defined as:

cp=pe/pw

his theoretical masimo limit according to betz theory is 0.59, and no turbine can overcome it.

for vertical axis wind turbines as mine should be 0.26<cp<0,37. -="" che="" con="" condizioni="" cp="52213/20093=2,59." define="" di="" fluent="" ho="" i="" inserito="" io="" le="" non="" prova="" quadra.="" qualcosa="" quindi="" sicuramente="" sono:="" sopracitati="" su="" un="" valori=""> model -> solver: pressure based solver, unsteady
define -> model -> viscous: k-epsilon standard
define -> model -> energy equation on

boundary condition:
definition -> boundary condition
velocity inlet: velocity specification method -> components, reference frame absolute
the zone of the domain in which the turbine is inserted (fluid zone): motion tipe -> moving mesh, rotational velocity -> 120 rpm
turbines (wall): wall motion -> moving wall, motion -> relative to housing cell zone.

solve:
solve -> control -solution
urf(under-realaxation factors) the values that fluent gives
pressure-velocity coupling -> simple
discretization: pressure -> early
all values on secon.order upwind

solve -> monitors -> residuals:I left the values indicated plotting the window
solve -> monitors -> force: options: print, plot and write
wall zones: turbines
coefficient: moment
moment center (0.0.0) turbine has rotation axis in origin

solve -> initialize -> initialize: all zones

solve -> iterate

as time step size I entered 0,005556s or corresponding to 4 degrees of rotation, while as time step number 1080 or 12 rotations.
the data of 4155nm is reported propio to the last time step.

What's wrong?
Maybe I'm wrong to interpret the data?</cp<0,37.>

hmmm
check on grid -> stairs
you would have imported a mesh in [mm] fluent the law in meters... so try to change it from the same window in [mm] And tell us if anything...
Hi.

 

[M1000]

hmmm
check on grid -> stairs
you would have imported a mesh in [mm] fluent the law in meters... so try to change it from the same window in [mm] And tell us if anything...
Hi.

hello paolina and thank you,
I checked but, but on gambit the mesh should be in meters, and fluent the law as such, in fact on grid->scale says that the mesh was generated in meters.
60rpm tests have again resulted in a negative outcome.
I got a moment again, on the tree of the turbine, negative and however unacceptable because reporting it in w much greater than the power made available by the air mass that invests the turbine.
for now I am doing some tests with a model of the turbine in 2d, but I am using a different approach, that is according to the speed tip ratio, defined as a relationship between the angle speed of the turbine at the end of the blade and the speed of the wind:

==============___________________________________________________________

in literature I found that for darrieus type turbines this parameter should be between 4 and 7 but I carried out some tests with λ=4 that gave me a negative moment.
now I am running one with λ=1.6, which by wind at 5m/s means 17rpm.
The truth is I don't know what fish to take anymore.

again thanks for the help

 

[paolina]

yes pre processor (gambit) creates mesh in [m] but fluent works better in [mm] (this only to simplify many setup operations) cmq can you repeat to me step by step the way you get the time report? :smile:
Good job

 

[M1000]

to obtain the report of the moment in settling as well as to activate the monitor of residuals active also that relative to the forces, in the case the cm along the axis z and in the origin of the axes, point in which the shaft of the turbine is placed.

when fluent finished working I go to the menu:
report->forces->options:moment,
I insert the coordinates of the position of the tree (0.0) and remove the surfaces (i.e. the blades of the turbine) of which I am interested in the report.
then click on "print" and fluent displays pressure moment values, viscosity moment and total moment that is the sum of the first two.

I have performed several tests on the 2d model for winds at 5m/s, 11m/s and 20m/s, and for values of λ=tip speed ratio equal to: 1.6, 1.7, 1.8, 1.9, 2, 2.5, 4.
in theory darrieus turbines as mine should work with 4<λ<7 but 2 then the values of the moment are negative.
the only useful results I obtained them for 1.6<λ<1.8 but in these cases I have yields or for better to say very low power coefficients or the order of cp=0.15.

I always thank you for the attention.

 

[stevie]

I do some considerations. . .
the unknown in the simulation of the problem, are both the available torque to the shaft and the rotation speed of the turbine.
or, for a certain wind speed value, your turbine will generate a date torque by turning at a certain speed.
classic case in which you have to use a dynamic mesh considering the structure fluid interaction. does not make sense to set values at case of rotation speed.
other consideration on the generated couple. . This varies with the shift of the angle of orientation of the turbine compared to the wind.
the instant torque product for the rotation speed will give you a generated instant power term, very variable in the rotation arc of the turbine.

 

[M1000]

In part you're right about everything, I'll explain.
In fact, I am not giving propio to cases the rotation values of the turbine, I refer to the material found in literature and compared to the values of some manufacturers.
in the sense that there is a parameter called tip speed ratio, defined as a ratio between the rotation speed of the turbine and the speed of the fluid current.
Unfortunately there is something wrong with this parameter as I wrote in the previous post.
with regard to the analysis fsi, i.e. fluid structure interaction, this exulates from my work (and also from fluent that is not suitable for this type of analysis, a software valid as far as I know could be adina, but I might be wrong on fluent).
Also regarding the pair is as you say, the pair provided by fluent is instantaneous however since my turbine has 5 blades has a very constant torque distribution (for this I am using this geometry), in practice after a number of rotations the value of the cm tends to an asymptotic value without oscillations.
stevie, I thank you, however, for the very precise and relevant details.