See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/312530787
Investigation of Blade Geometry Linearization on Performance of Small Wind Turbine Conference Paper · December 2016
CITATIONS
READS
0
366
2 authors: Shubham Deshmukh
Manabendra De
Motilal Nehru National Institute of Technology
National Aerospace Laboratories
1 PUBLICATION 0 CITATIONS
4 PUBLICATIONS 0 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
Development of 10/10 Wind Solar Hybrid System View project
Aeroacoustic optimization of a Horizontal Axis Wind Turbine View project
All content following this page was uploaded by Manabendra De on 19 January 2017. The user has requested enhancement of the downloaded file.
SEE PROFILE
61st Congress of the Indian Society of Theoretical and Applied Mechanics 11-14 December 2016, Vellore, Chennai
Investigation of Blade Geometry Linearization on Performance of Small Wind Turbine Shubham Deshmukh 1*, Manabendra M. De 2 1
Undergraduate Student, Mechanical Engineering Dept., MNNIT, Allahabad - 211004, India 2 Scientist, CSMST, CSIR-NAL, Bengaluru - 560017, India * E mail of corresponding author :
[email protected]
1. ABSTRACT With emphasis on combating climate change at both National as well as Inter-national level, research on renewable energy is getting a strong impetus. Wind Energy is one of the promising renewable energy resources, where lot of research is being carried out. Ample amount of research goes into developing techno-economically superior wind turbine blade shapes. While various blade optimization paradigms result in blades with non-linear geometry, in order to ease the manufacturing of wind turbine blades and decrease production costs, a standard industrial practice has been to linearize the blade geometry. However, this results in performance deterioration of the wind turbine. The present work is aimed to assess the effect of linearizing the chord distribution on the performance of a wind turbine blade. Performances of two candidate blade geometries, with nonlinear and linear chord distributions, were studied by two different approaches. The first approach was based on Blade Element Momentum Theory (BEMT) and the second approach employs principles of Computational Fluid Dynamics (CFD). In the normal operational envelope of the considered rotors, BEMT based approach indicated an average dip of about 2.3 % in power & CFD approach indicated an average dip of about 6.7 % in power. 2. INTRODUCTION Wind turbines, which offer a clean and effective way to harness the renewable energy of wind, have been attracting the attention of researchers. One of the critical components of a wind turbine is the blade, which converts the kinetic energy available in the wind into useful mechanical shaft energy. Hence, it is very important to design the blades correctly to optimize the energy yield. Linearization of chord for easing the manufacturing of turbine blade has been a common practice in the industry. However, this leads to a deviation in the aerodynamic performance. Wang et al. [1] demonstrated an increase in power coefficient (Cp) and a reduced blade surface area using non-linear blade design in a multi-objective optimization through BEMT. Liu et al. [2] adopted linear radial profiles of blade chord and twist angle on a heuristic basis and optimized the slope of these two lines based on BEM theory, with both Prandtl tip loss correction and wake consideration. Velázquez et al. [3] indicated a decrease in production cost by linearization of chord geometry between 70% and 90% of radial length at the expense of blade performance. A decrease in power coefficient from 0.49 to 0.36 was observed by Mirhosseini et al. [4] due to linearized chord distribution. Sugathapala [5] showed a similar decrease from 0.461 to 0.43 in the Cp values of linearized blade for local manufacturing of small-scale wind turbines. Mendez and Greiner [6] showed a method to optimize chord and twist distributions in wind turbine blades by using genetic algorithms and signified that high-quality results were obtained until the stall zone. They suggested the use of three-dimensional Navier - Stokes equations in a rotating frame to obtain precise solutions.
In light of the above literature, it is quite obvious that while this particular aspect of blade design is very important, detailed understanding of the flow physics is further needed to evolve the design process. In view of that, the present work attempts to assess the effect of linearizing the chord distribution by two different approaches, a BEMT based methodology and a CFD based approach. Losses in performance, as predicted by CFD, were found to be in line with that reported in the literature [5]. BEMT based approach under predicted the losses vis-à-vis the CFD approach. This difference is primarily attributed to the fact that while CFD resolved the flow physics in totality i.e. by solving the Navier - Stokes equations, BEMT used the averaged approximations for predicting the performance, i.e. use of Cl and Cd values of an airfoil to compute the forces and without considering the span wise flow. At the same time, while BEMT approach is computationally inexpensive, CFD approach is relatively computationally demanding. However, both the approaches clearly demarcated a decrease in power output of the wind turbine due to chord linearization, with magnified effects at relatively higher wind speeds. Choice of the approach is governed by the end context i.e. applied industrial research or time-bound product development. 3. PROBLEM FORMULATION In this study, a 7.5 m diameter wind turbine blade, designed and developed at CSIR National Aerospace Laboratories (NAL), Bengaluru, is considered. It is a part of the Wind-Solar Hybrid System (WiSH) being developed at CSIR-NAL. Chord distributions of two candidate blades that have been considered for the study are shown in figure 1. Blade geometries were described at 20 radial stations. A non-linear chord distribution arrived, based on a methodology for maximizing the Coefficient of power (Cp) at the design Tip Speed Ratio (TSR) and considering the effect of wake rotations. Linearization was based on a simple two-point scheme, where the linear slope of chord distribution variation along the span was obtained by joining the tip chord length with the chord length at the 10th radial station. These two lengths were computed from the previous methodology.
Figure 1 : Chord distribution of two candidate blades
4. METHODOLOGY The approaches to study the effect of linearizing the chord distribution of blade onto its performance are hereby presented. 4.1
BEMT ANALYSIS
BEMT methodology is an amalgamation of Blade Element Theory and Momentum Theory. Wind turbine blade is divided into elemental sections. Based on the local flow conditions and for the local blade geometry, the lift and drag forces are transformed into forces along in-plane and out-of-plane directions of the blade. All these elemental forces are then numerically integrated over the blade span to get the cranking force and thrust force. Cranking force and the operating rotor speed condition is used to compute the power. Since the entire methodology involves an iteration to find out consistent axial induction factor and tangential induction factor, an industrial standard commercial code GH-Bladed, available at CSIR-NAL, was used for analyzing the candidate blade geometries. Graphical user interface (GUI) of GH-Bladed is shown in figure 2.
Figure 2 : GUI of GH-Bladed Overall specification of the wind turbine as well as details of the blade geometry are fed into the software, operating conditions are specified and performance of the rotor is evaluated. Before the software was used for carrying out the effect of linearization of chord on performance of a wind turbine, its capabilities were assessed by simulating the performance of a 2.05 m diameter wind turbine, designed, developed and operational at CSIR-NAL. Comparing of the diurnal variation of power, as predicted by GH-Bladed, was carried out vis-à-vis the experimentally observed performance, as captured at CSIR-NAL. Comparisons of the diurnal power generation are presented in figure 3. Except for few instances, the trend of diurnal power generation, as predicted using GHBladed, is reasonably similar with the experimental observations. This confirms the prediction capability of GH-Bladed software.
Figure 3: Validation of diurnal power generation results of GH-Bladed with experimental data 4.2
CFD ANALYSIS
The CFD analysis was carried out with Shear-Stress Transport (SST) k-w model, for its ability to predict adverse pressure gradient boundary layer flow and separation. A rotating reference frame motion was adopted with periodic boundary condition and coupled pressure-velocity scheme was chosen with second order spatial discretization for robust computation. An unstructured mesh containing 4.4 million tetrahedral cells was generated using ANSYS ICEM CFD mesh generation tool, which was then imported into ANSYS Fluent environment and converted into polyhedral mesh. The polyhedral mesh was chosen due to its suitability for external aerodynamic analysis [7,8]. The conversion of tetrahedral mesh to polyhedral mesh gave an added advantage of reduced cell count from 4.4 million to approximately 1.4 million, reducing the computational time substantially.
Figure 4 : Domain size and boundary description
4.2.1
Validation Validation studies were carried out with NREL Phase VI turbine blades at 7 m/s and surface pressure coefficient was plotted against normalized chord length at 30%, 46.7%, 80% and 95% of blade length [9]. The present approach predicts the aerodynamic performance reasonably well as compared to the experimental data [10], except those zones where there is a sharp velocity gradient, as is observed at the leading edge on the suction side of the blade. Some discrepancy is also observed at the trailing edge at 80% and 95% of the blade sections which is expected due to blunt trailing edge, as a fillet radius of 1mm was used to increase the mesh quality at the trailing edge. The surface pressure coefficient has been defined as: 𝐶𝑝 =
𝑃 − 𝑃0 1 𝜌 (𝑢𝑖 2 + (𝑟𝜔)2 ) 2
where 𝑃0 is the ambient pressure, 𝑢𝑖 is the inlet velocity of wind , 𝑟 is the radius of blade section and ω is the angular velocity. r/R = 46.7%
r/R = 30% 6
6
k-w SST
k-w SST
5
Experiment
Experiment
4
4
3
3
-Cp
-Cp
5
2
2
1
1
0
0
-1
-1
-2
-2 0
0.2
0.4
0.6
0.8
1
0
0.2
0.4
x/c
0.8
1
x/c
r/R = 80%
r/R = 95%
6
6
k-w SST 5
k-w SST 5
Experiment
4
4
3
3
-Cp
-Cp
0.6
2
2
1
1
0
0
-1
-1
-2
Experiment
-2 0
0.2
0.4
0.6
x/c
0.8
1
0
0.2
0.4
0.6
0.8
x/c
Figure 5 : Distribution of Pressure Coefficient for NREL Phase VI blade at 7 m/s and 72 rpm
1
4.2.2
Domain Independency Study
Domain dependency studies were carried out for 7 candidate domains (sizes: n x 2n; 1≤ n ≤7) in order to ensure that the results of the analysis were independent of domain size. The dimensions of the domain were chosen as multiples of blade length, as demonstrated in figure 4. A 7x14 domain size was adopted as the consecutive predicted value of torque indicated a nominal difference of 1.4 % from its last lower domain size.
Torque (N-m)
200 150 100 50 0 1x2
2x4
3x6 4x8 Domain size
5x10
6x12
7x14
Figure 6 : Domain dependency study for non-linear blade at 8m/s and 177 rpm 5. RESULTS Static Pressure contours on pressure and suction side of both the blades with linear and non-linear chord distribution are shown in figure 7. Power curve predictions from BEMT & CFD based approach are shown in figure 8 & 9.
PS
(i)
SS
PS
(ii)
Figure 7 : Static Pressure contours on Pressure Side (PS) and Suction Side (SS) of (i). Non-linear blade and (ii). Linearized blade, at 8m/s and 177 rpm
SS
Figure 8 : Power Curves from BEMT approach
Figure 9 : Power Curves from CFD approach
6. CONCLUSIONS Assessment of the aerodynamic performance of a small scale wind turbine has been documented in the present paper. Effect of linearization of chord distribution on the power curve of two candidate blades were studied using BEMT and CFD based methodologies. Power produced by the blade geometry with linear chord distribution was found to be lower than the actual blade, with decreased performance at higher relative wind speeds. These losses are met due to geometry simplification for ease of production of turbine blades. Differences are found to be more pronounced in CFD results as compared to BEMT data. This is due to the inability of BEMT to resolve the flow physics as well as non-consideration of the span wise flow. Future scope of research includes continuing the studies further with CFD based approach for a set of candidate blades with linearized chord distributions and finding out which candidate would have minimal loss due to geometry simplification. Continuing the work further, it is envisaged to carry out experimental evaluation on full scale prototypes and document the findings so as to serve as benchmark cases for the research community. 7. REFERENCES [1]. W. Quan, J. Wang, J. Chen,S. Luo and J. Sun, Aerodynamic shape optimized design for wind turbine blade using new airfoil series, Journal of Mechanical Science and Technology, (2015). [2]. L. Xiongwei, W. Lin and T. Xinzi, Optimized linearization of chord and twist angle profiles for fixed-pitch fixed-speed wind turbine blades, Renewable Energy, (2013). [3]. M.T. Velázquez, M.V.Carmen, J.A.Francis, L.A.Pacheco and G.T.Eslava, Design and Experimentation of a 1 MW Horizontal Axis Wind Turbine, Journal of Power and Energy Engineering, (2013). [4]. M. Mirhosseini, A. Sedaghat and A.A. Alemrajabi, Aerodynamic modelling of wind turbine blades and linear approximations, 10th International Conference on Sustainable Energy Technologies, (2011). [5]. A. Gamarallage and T. Sugathapala, Aerodynamic Performance Modeling and Optimization of Small Scale Wind Turbine Rotors, International Conference on Modelling, Simulation and Applied Mathematics (MSAM 2015), (2015). [6]. J. Méndez and D. Greiner, Wind blade chord and twist angle optimization by using genetic algorithms, Proceedings of the Fifth International Conference on Engineering Computational Technology. - Las Palmas : Civil-Comp Press, 12-15, (2006).
[7]. M. Peric and S. Ferguson, The advantage of polyhedral meshes, CD Adapco Group, (2005). [8]. M. Spiegel, T. Redel, J. Zhang, T. Struffert, J. Hornegger, R.G. Grossman, A. Doerfler and C. Karmonik, Tetrahedral vs. polyhedral mesh size evaluation on flow velocity and wall shear stress for cerebral hemodynamic simulation, Computer Methods in Biomechanics and Biomedical Engg., 9-22, (2010) [9]. M. Ghasemian and A. Nejat, Aerodynamic noise prediction of a Horizontal Axis Wind Turbine using Improved Delayed Detached Eddy Simulation and acoustic analogy, Energy Conversion and Management, (2015). [10]. D. Simms, S. Schreck, M. Hand, L.J. Fingersh, NREL Unsteady Aerodynamics Experiment in the NASA-Ames Wind Tunnel: A Comparison of Predictions to Measurements, Colorado, National Renewable Energy Laboratory, (2001).
View publication stats