Back to the programme printer.gif Print




Delegates are invited to meet and discuss with the poster presenters in this topic directly after the session 'Aerodynamics and rotor design' taking place on Wednesday, 12 March 2014 at 09:00-10:30. The meet-the-authors will take place in the poster area.

Ricardo Pereira TUDelft, The Netherlands
Co-authors:
Ricardo Pereira (1) F P Gerard van Bussel (1) Weinand Timmer (1)
(1) TUDelft, Delft, The Netherlands

Printer friendly version: printer.gif Print

Abstract

Global optimization of horizontal axis wind turbine rotors including active stall control

Introduction

Increasing size of HorizontalAxis Wind Turbines (HAWT) together with the trend of going offshore requires robust designs. Modern offshore HAWTs are variable speed and pitch-controlled. If the pitch system is eliminated the maintenance costs may decrease and the availability will increase, which might result in a lower cost of energy(COE). This paper investigates active stall control as an alternative to pitch control. Previous research [1] showed active stall control of HAWT appears feasible only if the rotor is re-designed to incorporate active-stall devices. A blade planform optimizer
is thus developed, including rotational speed and actuator employment as variables.


Approach

The optimizer iterates on a solution vector with values corresponding to twist and chord spanwise distribution, and rotational speed and actuator lift change, dCl, imposed at different wind speeds.
Each iteration uses a Blade Element Momentum (BEM) code to derive what is the overshoot in Torque, Root Flapwise Bending Moment and Thrust over the range of considered wind speeds, compared to the values obtained with the National Renewable Energy Laboratory (NREL) 5 MW pitch controlled machine[2]. The required increase in blade laminate skin thickness to meet the tip clearance requirement is also computed.
Based on these overshoots, the increase in the turbine capital cost (TCC) is computed using mostly the turbine cost model of [3]. The power versus wind speed curve computed with the BEM code is multiplied by the Weibul wind speed distribution of a North Sea offshore site, and used to compute the annual energy production (AEP).
Two different actuator types are considered, namely boundary layer transpiration(BLT) and trailing edge jets(TEJ). The influence of the actuators is considered on the change in local lift coefficient dCL they can induce and also on their power consumption. From each considered wind speed, the actuator power consumption is subtracted from the rotor aerodynamic power, thus calculating the AEP.
The COE is determined from the new TCC and AEP, and is chosen as the function to minimize. Figure illustrates the calculation procedure. To ensure the solution found is not local minimum, the cost function's response surface is covered using splines [4], effectively sweeping across the whole design space and thus performing a global optimization.


Main body of abstract

The optimization is performed with the matlab function’fmincon’, which finds a solution to a constrained problem by iterating on the design space and going in the direction of decreasing cost function. The twist and chord spanwise distribution were constrained to yield "reasonable" results, the maximum and minimum allowed values were chosen rather arbitrarily, simply to allow sufficient freedom for the solution to converge such that the final design point is not located at the edges of the domain.The inequality matrix in the algorithm formulation was used to enforce the trend present in an optimum rotor, [5] that is both twist and chord decreasing from the root to the tip of the blade. A non-linear constraint was used to limit the radial variation of chord and twist. This was imposed both due to manufacturability issues and to limit the circulation shedding.The radial derivatives are limited to twice the maximum occurring at the reference blade.
It is assumed that the actuators change the lift of the airfoil section, but not the drag. BLT is modelled by using a panel code which includes boundary layer suction and blowing, validated in [6]. It is assumed that the change in lift coefficient imposed by the actuation is proportional to the transpiration velocity at angles of attack beyond stall, while not changing in the linear part of the lift polar. This is illustrated in figure for the DU93-W210 profile. An airfoil optimizer incorporating BLT [6]is used to design airfoils which are sensitive to transpiration, while being robust enough to handle angle of attack changes from wind shear or turbulence and having a satisfactory rough performance, as imposed to any wind turbine airfoil. These tailored airfoils ensure the potential for using BLT is maximized. The actuator’s power consumption was computed by assuming a known porous material [6] and calculating the required pressure change and mass flow rate of the transpired air mass.
TEJs were modelled following [7]. For this actuator, it is assumed that the same change in lift coefficient is obtained by imposing the same jet momentum coefficient. Increasing this momentum coefficient increases the change in the lift coefficient obtained in the attached regime. As the angle of attack increases, the flow separates in the trailing edge region rendering TEJs ineffective. The airfoil geometry and TEJ angle play a role in the actuators’ influence, but this trend was observed for different airfoils and was accordingly implemented [8]. The power consumption of the TEJ is computed using the required mass flow rate and the required velocity (i.e. momentum) of the jet.
The blade planform solutions obtained tend to increase the yearly energy capture as much as possible, resulting in large torque overshoots. These overshoots increase the ICC, but the increase in AEP is much larger and thus the COE is decreased. This type of solution leads to larger rating machines, for which other design parameters would be altered, such as the rotor radius. Moreover the model employed to determine the COE increases the cost of each of the components independently, meaning that an increase in the aerodynamic torque implies an increase in the generator cost, but its weight increase is not used to resize the tower. Considering this, the maximum overshoot in torque and power is constrained to 1.5, and different values for the maximum allowed overshoot are investigated. Figure displays the different COE obtained as different torque overshoot constraints are applied, for the cases with and without actuation, obtained with the NREL reference turbine geometry as initial guess for the optimizer. The relative COE is simply the ratio of the new COE compared to the COE obtained with the reference pitch-controlled turbine. Results indicate that the COE may decreaseby approximately 2-7% when using active stall control. Figure shows the blade planform optimized solution, together with the geometry of the reference blade, obtained by imposing a torque overshoot of 40% and employing TEJs.


Conclusion

The trend to go offshore demands for robust solutions. If power regulation could be achieved through active-stall control, the pitch system could be mitigated, making the HAWT more robust and perhaps contributing to a decreased COE. To assess the feasibility of using active-stall control to replace the pitch system for power regulation, a blade planform optimizer incorporating active-stall actuators was developed. The tool minimizes the COE by varying the chord and twist spanwise distribution and by considering also actuator employment and rotational speed variation. The optimizer evaluates the AEP, while using a cost model to determine the expenses of the new design solution. Two types of actuators are modelled, BLT and TEJ, which are evaluated in terms of their power consumption and lift coefficient sectional change.
Results show the torque overshoot has a very large influence on the final COE, and by allowing an overshoot of 40% the COE can be decreased even without employing actuation. When actuators are employed, results indicate that the COE may decrease by 2-7 % compared to pitch controlled HAWT, depending on the actuator employed and maximum allowed torque overshoot. TEJs appear to bring a larger decrease in the COE compared to BLT, possibly related with the fact that they change the lift coefficient below stall. Regarding the blade planform, the optimized solution generally tends to decrease the chord and the twist angle, leading to an increase in angle of attack along the blade,using the decreased lift coefficient of stalled airfoils to regulate power above rated wind speed.
In the future other types of actuators will be considered, namely Dielectric Barrier Discharge plasma actuators. These actuators will be considered since they have no moving parts and require no internal piping to feed a pneumatic system, which is attractive for offshore HAWTs. Future research could also focus on estimating how much of a reduction in the pitch system cost could be achieved by employing active stall control, since in this case the pitch system would only be used in emergency situations and could thus perhaps be designed to be lighter and cheaper.



Learning objectives
- Assess the feasibility of employing active stall control for power regulation, mitigating the pitch system.
- Gain insight on which rotor design would be optimal for such a task, in terms of chord and twist distribution, and actuator employment.



References
[1] - Pereira, R., van Bussel, G. J. W., andTimmer, W. A., 2012. “Active stall control for large offshore horizontal axis windturbines;a conceptual study considering different actuationmethods”. IOP-Science of Making Torque, Oldenburg 2012.

[2] - Jonkman, J., Butterfield, S., Musial, W., and Scott, G., 2009. “Definition of a 5-mw reference wind turbine for offshore system development”. NREL-Technical Report, 500-38060.

[3] - Fingersh, L., Hand, M., and Laxson, A., 2006. “Wind turbine design cost and scaling model”. NREL Technical Report, 500-40556.

[4] - de Visser,C., Chu, Q.P. and Mulder, J.A., 'A new approach to linear regression with multivariate splines', Automatica 45 - 2903 2909, Elsevier 2009

[5] - Manwell, J., McGowan, J., and Rogers, A., 2002. “Wind energy explained - theory, design and application”. NRELTechnicalReport, 500-38060.

[6] - Oliveira, G., 2011. “Wind turbine airfoils with boundary layer suction - a novel design approach”. TUDelft Master Thesis.

[7] - Brunner, M., Blaylock, M., Cooperman, A., and van Dam, C., 2012. “Comparison of cfd with wind tunnel tests of microjets for active aerodynamic load control”. 50th AIAA Aerospace Sciences Meeting.

[8] – Blaylock, M., Chow, R., Cooperman, A. and van Dam, C.P., “Comparison of Pneumatic Jets and Tabs for Active Aerodynamic Load Control”, Wind Energy , DOI:10.1002/we 1638, John Wiley & Sons, 2013