Software
CARDAAV
Several aerodynamics prediction models currently exist for studying vertical-axis wind turbines and a complete state of the art review including the appropriate references is given by Turyan et al., Strickland and Paraschivoiu. Presently, various aerodynamic methods, appropriate for the conception of turbines, are available to designers. Theses models are based on several methods which can be classified into three categories:
- Navier-Stokes / Source-in-cell models ,
- Vortex models, and
- Streamtube models.
The fundamental objective of each model is the calculation of the induced velocities on the upwind and downwind zones of the rotor and the determination of loads and performance generated by the turbine. In Navier –Stokes / Source-in-cell models, the flow field around the VAWT is computed by solving the steady-state Navier-Stokes equations. The tubines are represented by distribution of momentum sources and sinks that are computed using blade element theory. Those models are very CPU-intensive, and therefore are not appropriate for preliminary design purposes. Their main advantage lies in their ability to predict the details of the flow field around the rotor. This capability makes this type of model appropriate for the study of rotor and wake interferences.
The CARDAAV computer code, based on the Double-Multiple Streamtube model with Variable upwind- and downwind-induced velocities in each streamtube (DMSV) model, has been constantly improved in order to produce an efficient software package appropriate for the needs of VAWT designers. The software is in International Standard Units and runs under Microsoft Windows environment, is user-friendly in the sense that the rotor geometry and its operating conditions can easily be prescribed by the user though the interface windows. The results can be directly visualized on the screen or stored in ASCII files in a format compatible with the graphic processing software TECPLOT® (Amtec Engineering Inc.) for detailed interpretations. Those ASCII files can also be interpreted to graphics by any other software as they are defined in text format.
Top
Software Capability
CARDAAV is a software for the prediction of the aerodynamic forces developed on the blades of a Darrieus turbine, hence the torque and the mechanical power produced by the rotor but also the structural forces withstood by the machine. The compute code takes into account the:
- Shape of the rotor and the blades.
- Airfoil used at each blade station.
- Dynamic stall effects.
- Secondary effects as: the tower shadow and the struts and spoilers drag.
- Ground shear effects.
- Blade-tip loss of straight bladed turbine.
- Tri-dimensional atmospheric turbulences in the wind.
- Skew angle of the wind velocity.
As mentioned in the introduction of the user’s manual, CARDAAV is based on the Double-Multiple Streamtube model upwind- and downwind-variable induced velocities in each streamtube (DMSV), which is well defined in the reference. That model is considers the following assumptions:
- The rotor can be assimilated to two actuators disks in tandem distanced by the diameter of the rotor.
- Each one of the actuator disks is divided in many streamtubes.
- The local flow velocity and the aerodynamic forces are computed at NP points on the area defined by the intersection of an actuator disk and a streamtube. NP is the numerical parameters usually fixed at 4 or 6.
- The flow through the streamtube is considered steady.
- The pressure variations along the streamtubes boundaries are aerodynamically independent.
- The streamtube have equal and fixed width and height in the upwind and the downwind section of the rotor.
- The tri-dimensional stochastic wind generation module considers the Taylor’s Frozen turbulence hypothesis.Â
Additional assumptions are induced by the numerical computation of the quantities:
- The lift and drag coefficients are listed in resources files. They are given as a function of Reynolds number and angle of attack. A linear interpolation is executed by the code for any Reynolds number and incidence angle that is not listed in the file.
The computation procedures of the DMSV is well explained in the reference mentioned above but consist mainly in the following steps:
- The user provides all the information on the geometry of the Darrieus turbine as well as its operating conditions.
- The rotor is divided in a number of streamtube defined by the user. The user also defines the number of integration point per streamtube NP.
- If the ground shear effect is considered, the wind velocity is evaluated at the entrance of each upwind streamtube.
- If the turbulent wind option is activated, the perturbations in the x and y directions are generated at each grid point (will be defined further) in terms of complex Fourier series, which are evaluated at blade element position and summed to the constant component of the wind.
- Initial unitary values are given to the upwind section interference factor.
- The local resultant flow velocity and angle of attack are evaluated at each integration point and the corresponding lift and drag coefficients are computed using the blade element theory
- The relation that equates the variation of the streamwise momentum to the projection of the blade element forces is solved iteratively and new interference factors are obtained.
- The local flow velocity, angle of attack and the aerodynamic forces are revaluated.
Considering the named model and the calculation procedure, the following paragraphs will present the possibilities and limitations of CARDAAV.
Top
What CARDAAV can do
The CARDAAV v.3.00 software will allows the users to:
- Choose between the five available blade shapes (straight, parabolic, catenaries, Troposkian with weight or without) or define a custom shape of blade composed of straight lines and curves using the blade-drawing tool of the guided user interface.
- Specify the airfoil used as section at every station of the blade.
- Define the radius of the tower, the airfoil and the chord of the struts and the aerodynamic characteristics spoilers or flaps.
- Add a constant pitch angle to the blades.
- Enter the exponent of the shear-law for the ground shear effect.
- Specify the density and the cinematic viscosity of the considered fluid.
- Enter the skew angle of the wind or a range of skew angle to be considered.
- Choose between the four different models for the generation of the tri-dimensional stochastic wind (Frost, Von-Karman, Kaimal and Solari).
- Define the frequency range of the atmospheric turbulence to be accounted during the computation and also the number of simulation to be executed.
- Obtain the averaged mechanical power (kW) on all the simulations available at a given rotational speed for a range of wind velocities defined by the user or at a given wind velocity for a range of rotational speeds.
- Obtain the averaged efficiency curves on all the simulations of the rotor in terms power coefficient (Cp) as a function of the Tip Speed Ratio (TSR).
- Obtain the distribution of the incidence angle, interference factor and normal/tangential forces (or their dimensionless coefficients) along the span of the blade at any azimuth position for a given TSR. The distribution of the same quantities can also be obtained along the blade path at the equatorial position for a given TSR.
- The mentioned quantities can also be obtained for every successful simulation.
- Visualize the time series of the stochastic wind evaluated at a pre-defined time step.
Top
CANICE 2D/3D
Currently, atmospheric icing is one of the major concerns for the certification authorities as well as the aircraft manufacturers since it presents a major hazard to aircraft operating under natural icing conditions.
CANICE uses an aerodynamic panel method to solve for the potential flow field which is then corrected for compressibility effects. The code uses Lagrangian tracking to determine droplet trajectories and impingement locations. The modified Messenger model is used for ice accretion thermodynamics, in conjunction with an integral boundary-layer solution for heat and mass transfer rates. The code includes roughness, runback and water splash/ice shed models based on water-bead model. The code has the capability of simulating supercooled large droplet conditions. The ice accretion is re-paneled and the airflow field is re-computed at each time step to determine the growth of the ice accretion as time proceeds. The time-step is user specified, however, it is recommended that it be between 30-50 seconds. CANICE 3.0 beta incorporates smoothing of the ice shape between each time step. The smoothing is based on three criteria: a minimum and a maximum allowable panel length and a maximum allowable angle between adjacent panels.
CANICE computer code is presently used for aircraft icing and anti-icing simulation and has been approved by Transport Canada for the certification of Bombardier aircraft in icing conditions.
The first version of CANICE3D, the icing & anti-icing code for complex 3D (full aircraft) configurations, was released and its improvement has already began. SUCTION, a code that can be used to optimize the suction mass flow rate and its distribution over a finite span wing (3D flow), was developed in the Laminar Flow Control area of the project. Numerical prediction of the drag for 3D configurations was performed on the ONERA-M6 and the DLR-F4 geometries.
Even if successful, these attempts to deal with the real, three-dimensional flow phenomena, revealed that extensive work still needs to be done in order to bring the simulation capabilities of the codes to higher levels of accuracy and reliability. Among the objectives that could serve to a substantial improvement of the icing & anti-icing simulation accuracy of the CANICE codes and in the areas of drag prediction and reduction and laminar flow control one can ennumerate the following:
- A more efficient (faster) droplet trajectory module must be developed to enable water collection computations and icing simulations on arbitrary 3D configurations;
- A better physical models for the icing related roughness;
- Development, implementation and validation of a pertinent model for the water drop splashing and shedding to improve the predicted ice shapes in SLD icing conditions;
- Development, implementation and validation of a model for large droplet breakup in SLD icing conditions;
- Modeling and numerical study of the unsteady jets influence on the efficiency of the heat transfer in a hot-air anti-icing system;
- Modeling and numerical study of the inner wall roughness influence on the heat transfer in a hot-air anti-icing system;
- A performing preprocessor (grid generator), along with a 3D compressible Navier – Stokes flow solver available with its source files, are needed to enable more accurate numerical drag prediction and decomposition of complex (full aircraft type) configurations;
- Additional work to identify and reduce the influence of the numerical errors on the drag prediction results and to evaluate and validate different drag reduction solutions;
- Further studies to determine the source of the lack of convergence in the linear compressible transition predictor PSE2DLC and to extend it to 3D flow transition analysis;
Improvement and careful validation of the 2D and 3D suction-based laminar flow control codes, previously developed, aiming at their release for industrial use
Top
