Fluids

Fluids

MRF to simulate a turbine in buoyant flow with solar radiation

    • pcaicedo
      Subscriber

      Hello everybody,


      I have been working in a buoyant flow with solar radiation. The case without MRF gives results as expected and according the physics; however, when I include MRF to simulate a turbine, I got several errors like negative torque and reverse flow in the whole domain. The summary of what I have done is the following:


      1) Simulating the case without MRF to get the airflow to design the turbine. The mass flow rate is 950 kg/s, and the upwind velocity 11 m/s for a diameter of 10 m (vertical component of the geometry). The following slice shows the velocity contours.



      2) Designing a turbine with the previous mass flow rate and considering 15 rpm.


      3) I keep the same BCs. The only difference with the previous simulation is the interfaces at the inlet and outlet of the rotor section. The turbine is placed in the bottom center of the domain. I assign a rotational velocity in that zone and then run the simulation.



      The results are not even close as expected with that kinetic energy. At 5-10 rpm, I got reverse flow (it should be at higher), and at very low 0.1 rpm, negative torque.


      So, in my understanding if the problem without MRF is running ok, and then I am applying normal conditions to simulate rotating flow, where could be the problem?


      I have checked carefully the configuration and everything seems ok.


      Thanks,


      Paul

    • DrAmine
      Ansys Employee

      In which direction does your gravity vector point to? Is it the same as or aligned with the rotational axis?

    • pcaicedo
      Subscriber

      Hi,


      It is the same/aligned with the rotational axis, and for my case is -Z.

    • Karthik R
      Administrator

      Hello,


      I do not fully understand your problem, but could you perhaps work your simulation with just MRF. Could you switch off solar radiation and buoyancy? Perhaps work your MRF first and then all other complexity? I am not sure it is fully helpful.


      Best Regards,


      Karthik 

    • pcaicedo
      Subscriber

      Hi,


      The flow is accelerated by the buoyant effect (density differences), so it is necessary to have radiation and buoyancy in order to have airflow. It is this concept (https://en.wikipedia.org/wiki/Solar_updraft_tower#/media/File:Solar_updraft_tower.svg), but in my case I have only one radial turbine at the bottom center.

    • klu
      Ansys Employee

      Hi Paul,


      As Karthik advised, please turn off the radiation and gravity but keep the same mass flow so that we can verify if the MRF is working properly. Also please double check the MRF settings for example the rotating direction (right-hand rule), wall boundary conditions, etc.

    • pcaicedo
      Subscriber

      Thanks for your answers,


      I have turned off radiation and buoyancy, applied mass flow inlet (the value I got without MRF) and pressure-outlet at the inlet and outlet BCs respectively. 


      I have checked the rotating direction with the velocity vectors and it is ok. The IGVs direct the flow to the blades. 



      But, the torque is negative. At this point, I think I am doing something wrong with MRF.


      Moments - Moment Center (0 0 0) Moment Axis (0 0 1)


      Moments (n-m) Coefficients


      Zone Pressure Viscous Total 


      blades -31235.298 0.20310766 -31235.095


       


      The MRF setup is the following


       



       


      This is the plot of the torque



       


      I added two interfaces as BCs of the rotor zone. Do I need to consider / apply something else?


       


      Regards,


      Paul

    • klu
      Ansys Employee

      Hi Paul,


      I would suggest start with a mass flow inlet plus a pressure outlet. The purpose is to verify how much work should be extracted by turbine at the "design" condition. Also can you please attach screenshots for the BCs of turbine wall surfaces and torque calculation settings? Thanks.

    • pcaicedo
      Subscriber

      Hi,


      Sorry, I have edited the previous comment after running a quick simulation. 


      BCs of turbine wall surfaces



      The hub and shroud



      The torque calculation settings



      Thanks,


      Paul

    • Karthik R
      Administrator

      Hello Paul,


      Here are a couple of video tutorials on rotating flow simulation using MRF. These might be useful resources if you are looking to double-check and debug your settings.




      I hope this helps you.


      Best Regards,


      Karthik

    • pcaicedo
      Subscriber

      Hello Karthik,


      I have watched those videos previously. I am doing exactly the same thing as the video #1 (left the walls created by the interfaces as static).




      Also, as Keyur told me in the previous post (https://forum.ansys.com/forums/topic/mesh-interface-generation-with-extra-walls/), those walls don't have influence in the case.


      Regards,


      Paul

    • DrAmine
      Ansys Employee

      Might it be that your negative torque is due to your angular velocity as it is negative direction relative to the axis of rotation (counter-clock-wise)

    • klu
      Ansys Employee

      Hi Paul, 


      As abenhadj suggested, the directions of rotation and torque should be the same therefore I do not see any problem with the negative sign but please check if the magnitude meets your design or makes sense. Thanks.

    • pcaicedo
      Subscriber

      Hi,


      Thanks for your help.


      The current angular velocity is clockwise (negative), but I have tried with 10 rpm (counter-clockwise) and even at 0 rpm (without MRF), the torque is negative.


      I computed at 80 rpm (clockwise), and now I have positive torque, but in my understanding that doesn't make sense because in normal conditions the negative torque appears at high rpm since the turbine decelerate. Also, the maximum power without the turbine is around 56 kW (considering the kinetic energy), and with that torque (79000 Nm), I got 661 kW. It is not realistic.



      I simulated considering radiation and buoyancy with that angular velocity (80 rpm), and the flow is completely reverse.


      Regards,


      Paul

Viewing 13 reply threads
  • You must be logged in to reply to this topic.