[mujahidbadshah]

Hi,
I am simulating the performance of a Horizontal Axis Tidal turbine just like a wind turbine operating under water.
I have a rectangular outer domain for the water channel and inside it an inner rotating circular domain enclosing the turbine rotor. I am using steady state and MFR in CFX. Boundary conditions for the simulation are inlet with normal velocity of 0.5m/s, outlet with (0 Pa) static pressure, channel sides and bottom are no slip walls and the top is entertainment with ( 0 Pa) and zero gradient turbulence. I have run a series of 8 simulations for Tip Speed Ratios (TSR) (from 1 to 8) by specifying angular velocity for the rotating inner domain.
Now the issue is that at TSR 7 and 8, my turbine is producing more power than the Bitz limit of 59%. Kindly, some one help me where is the error in my simulations.

[ghorrocks]

FAQ: https://www.cfd-online.com/Wiki/Ansy..._inaccurate.3F

[strobel]

How is your CpxTSR curve?
It is common to reach unreal values of torque at high TSR, because your system start to work as a propeller.

[mujahidbadshah]

Thanks for the reply. The Cp Vs TSR curve is fine and exhibits an increasing trend for Cp with increasing TSR from (TSR 1-5). At TSR 1 the deviation from the experimental results is 50% and becomes only 2% at TSR 5. However, it keeps on increasing and at TSR 7 and 8, the value of Cp exceeds the ideal Cp value of 0.59.
I think that this is a common trouble in wind/tidal turbine simulations. This same issue has also been reported in other threads in this forum given in the links below:
Wind turbine simulation in Ansys CFX

Once a specific rotation speed has been reached (above lambda = 5, where labmda is the ratio between the wind speed and the blade tip speed), the blade's aerodynamic coefficient goes up to 70%.

However, Betz limit says it can never be larger than 59.7 %. And, if I increase lambda a bit more, the blade's aerodynamic coefficient CP keeps getting larger until the wind turbine produces more power than the power available in the wind.

Wind turbine simulation

[ghorrocks]

While Paulo's comment is useful and may well be the source of the problem, unless you have properly validated and verified your simulation you are wasting your time pondering details of the results. You have not mentioned anything about things like the mesh, boundary proximity and convergence tolerance (amungst others) and whether they are suitable for the accuracy you require.

Have you checked that the issues I linked to in the FAQ have been checked and confirmed to be OK? I cannot count the times I have seen discussions on the forum about details of the results of simulations which, in the end, turned out to be artefacts of inadequate mesh resolution or poor convergence.

[mujahidbadshah]

Hi Glenn,
During the course of last week or so, I have read a lot of literature regarding this issue. one of the research paper regarding simulating the performance of tidal turbine have stated that:
"The rotation of the inner domain adds an additional residual torque to the output torque. In a real-life scenario,the flowing fluid would cause the blade to rotate and the net torque obtained is converted into useful power. For our rotating domain simulations, the turbines blades are, however, subjected to an additional net torque which is a sum of the torque due to the flowing water and the torque due to the rotating domain. The residual torque due to the rotating domain was evaluated for each case by running additional simulations with identical rotational speeds where the free-stream velocity was set to zero. This allowed for evaluating the residual torque due to the rotating domain from these computations. This residual torque was subtracted from the total torque to calculate the net effective torque on the rotor which was used for all the power calculations".
Kindly comment on this. To me my issue also seems to be in calculating the torque.
Best Regards

[ghorrocks]

I have no idea what the quote you posted is talking about. In CFX at least, the rotating frame of reference is just that - it is just a frame of reference. There is no torque associated with rotating the frame of reference. The torque should be the same regardless of which frame of reference it is measured in.

I would be very suspicious that the quote you posted is actually covering up an error in the paper associated with how they incorrectly implemented the rotating frame of reference.

[mujahidbadshah]

Hi Glenn,
Thanks for the reply. But there is an issue. When we do this simulation with inlet velocity as 0 m/s and the inner cylindrical domain is rotated at some angular velocity, the torque should be zero. Because, at zero inlet velocity the turbine enclosed in that domain is not producing any torque. But the CFX simulation gives a torque value due to the rotating inner domain. Could you please explain that for me and other users at the forum. And how will we calculate the torque value for the turbine alone.

[ghorrocks]

If you spin a turbine with zero flow through it then it has to push the fluid out of the way as it spins around. This will generate a torque. The torque of doing this has almost nothing to do with the torque generated by the turbine when it is operating at the design flow point. At zero flow it is operating in a grossly separated flow regime and at the design flow point it is operating at a fully attached flow regime - completely different.

You calculate the torque for the turbine alone by torque_x/y/z()@turbine_surface where x/y/z is the rotation axis and turbine_surface is the surface which is the surface of turbine rotating component.

[mujahidbadshah]

Hi Glenn,
Thanks again for your continued support and help. I will look into in it in more detail and hopefully will find out whats wrong that's causing increase in torque generated by the turbine with increasing input r.p.m for the inner domain.

[ghorrocks]

Make sure you know the system performance curve of the device you are modelling. The torque curve you are seeing may be physically correct.

[mujahidbadshah]

Hi Glenn,
Thanks once again for the reply. We have the actual performance curve of the turbine to validate our results. Can you kindly explain it a bit further for me, what do you mean by "The torque curve you are seeing may be physically correct."
One thing more, I have a rectangular stationary domain and inside it a 25 rad/s rotating cylindrical domain that encloses the turbine rotor. The turbine rotor surfaces are assigned no slip wall boundary condition. Kindly, guide me whether the rotor should be assigned a wall velocity as a rotating wall or not? If the rotor is to be specified as a rotating wall then what should be its angular velocity. whether it should also be 25 rad/s or 0 rad/s.

Best Regards
Mujahid Badshah

[ghorrocks]

My comment about the torque curve was just saying that the torque can increase or decrease with speed, depending on the design, speed and many other factors. If you have the performance curve to compare against then you have then you have the data to compare against.

For the rotor blades just leave it at the default option. Relative to the rotating frame of reference they should be no slip boundaries.

[elluthfi]

Originally Posted by mujahidbadshah

Hi Glenn,
During the course of last week or so, I have read a lot of literature regarding this issue. one of the research paper regarding simulating the performance of tidal turbine have stated that:
"The rotation of the inner domain adds an additional residual torque to the output torque. In a real-life scenario,the flowing fluid would cause the blade to rotate and the net torque obtained is converted into useful power. For our rotating domain simulations, the turbines blades are, however, subjected to an additional net torque which is a sum of the torque due to the flowing water and the torque due to the rotating domain. The residual torque due to the rotating domain was evaluated for each case by running additional simulations with identical rotational speeds where the free-stream velocity was set to zero. This allowed for evaluating the residual torque due to the rotating domain from these computations. This residual torque was subtracted from the total torque to calculate the net effective torque on the rotor which was used for all the power calculations".
Kindly comment on this. To me my issue also seems to be in calculating the torque.
Best Regards

Hello Sir, I have the same problem with my simulation. My torque is always increasing whenever I increase the rpm of rotation domain. Meanwhile, in my experiment, I used rope brake dynamometer by applying load gradually and the rpm had the inverse proportional relation to the torque. Have you solved your simulation problem Sir?