1-12.pdf

Mechatronics 55 (2018) 1–12

Contents lists available at ScienceDirect

Mechatronics journal homepage: www.elsevier.com/locate/mechatronics

Modelica-based dynamic analysis and design of lift-generating disk-type wind blade using computational fluid dynamics and wind tunnel test data☆ Yeongmin Yoo, Soyoung Lee, Jaehyun Yoon, Jongsoo Lee∗ School of Mechanical Engineering, Yonsei University, Seoul 120-749, South Korea

a r t i c l e

i n f o

Keywords: Wind power system Life-generating disk-type blade Multi-physics system Computational fluid dynamics Wind tunnel test Integrated Modelica simulation

a b s t r a c t Wind power generation research and application technology have received much attention in the development of renewable energy. However, the traditional blade-rotating type wind power system has a number of drawbacks such as natural landscaping damage, flow-induced noise, and shadow flickering problems. In this paper, we propose a lift-generating disk-type blade power generation mechanism that can effectively generate wind power even with a simple structure considering the problems of the existing systems. Data on the lift force in relation to the shape of the designed blade were derived through a computational fluid dynamics simulation, and the Modelica language was used to model the integrated multi-physics wind power system. Then, a wind tunnel test was conducted using a small-scale model of the disk-type blade created to verify the simulation. The experimental results were in good agreement with the simulated results. Thus, we validated the modeling of the wind power system and applied the law of similarity to obtain the generator power output prediction results for the actual scale model.

1. Introduction Wind power technology is one of the various methods used to obtain electricity from natural energy and has received much attention because it utilizes an infinite and free resource. In Europe, particularly in Germany, Denmark, and the Netherlands, there has been much research on wind power systems (WPSs) since the 1970s. As a result, several megawatt WPSs have recently been commercialized. Thus, wind energy is expected to play a decisive role in the future world energy supply [1–3]. To produce WPSs aligned with this purpose, efficient development is required to lower the development cost. With this objective, computer simulation technology is used in all development fields. Therefore, the simulation technique is a very important means for streamlining development work. When a WPS is being designed, numerous factors must be considered, including various physical phenomena such as the mechanical dynamics, electricity, control, and flow, along with a suitable program for the required integrated simulation of the system. The multi-physics system simulation has been typically integrated under different hardware/software platform environments. Consequently, it is easy for this design method to effectively cope with frequent design changes to the initial conceptual design, which results in reduction unnecessary work and increase efficiency. In the initial design phase, it is necessary to use integrated simulation software because it is important to communicate, coordinate, and cross check between concept verification and development personnel rather than conduct detailed analyses [4,5]. For this reason, the Modelica [6–9] language has been adopted for integrated simulation. Various WPS studies have been conducted using Modelica. Enge-Rosenblatt et al. [10] published a paper on the power energy results of varying the simulation configuration based on the way the WPS operates. Strobel et al. [11] proposed a Modelica library by designing an offshore WPS and verifying the simulation results. Petersson et al. [12] presented a mathematical model of a vertical axis WPS and presented simulation results for it. Eberhart et al. [13] constructed an open source library for the simulation of a horizontal axis WPS. In this study, a new type of WPS utilizing Modelica was designed based on these studies. The conventional WPS is a horizontal-axis-type because the rotation axis of the blade is placed horizontally. These types of blades have suitable characteristics for high-speed rotation, but various problems are raised, such as noise generation due to the large scale wind power system (tip speed ratio, etc.) [14–16], changes in the landscape of the natural environment, and social pressure issues because of the shadows cast by the structures [17]. Therefore, the authors had the goal of designing a

☆ ∗

This paper was recommended for publication by Associate Editor Dr. Yayou Li. Corresponding author. E-mail address: [email protected] (J. Lee).

https://doi.org/10.1016/j.mechatronics.2018.08.003 Received 8 February 2018; Received in revised form 19 July 2018; Accepted 13 August 2018 Available online 1 September 2018 0957-4158/© 2018 Elsevier Ltd. All rights reserved.

Y. Yoo et al.

Mechatronics 55 (2018) 1–12

Fig. 1. Schematic diagram of the lift-generating disk-type WPS.

lift-generating disk-type WPS that could generate electricity in response to an omni-directional wind instead of a conventional WPS. The disk-type blades were first designed using computer-aided design (CAD). Then, they were modified based on the design parameters. We designed the WPS by applying the designed blade in a Modelica simulation. Wind tunnel tests were performed to verify the simulation, and the results were in good agreement. 2. Mechanism of wind power system The lift-generating disk-type WPS is a system in which the blade is disk-shaped, and power is generated while it moves in a vertical direction. The basic principle of the system is that along the vertical axis of the supported tower, the disk-type blade is lifted up and down by the wind. This type of system generates electric energy through a power transmission system (PTS) that converts such motion into rotary motion, along with a power conversion system (PCS) such as a generator. This type of model is very different from the typical WPSs such as the conventional horizontal or vertical axis turbines. The proposed WPS is advantageous for areas in which flow-induced noise and blade flickering/shadows are critical. The small-scale WPS developed by the authors includes a disk-type blade, spring for managing the vibration generated by the vertical movement of the blade, crank-rod mechanism [18,19] for converting the vertical motion into rotational motion, and tower for installing the system. Thus, through the PTS, power is produced by the generator. The proposed WPS is shown in Fig. 1. The proposed disk-type blade has some distinct characteristics. First, the disk can move up and down to generate a lift force that is converted into electrical energy. Second, because of its symmetrical shape, it is easily activated by the wind coming from any direction. The proposed WPS can generate the power from the arbitrarily directed wind. The additional conversion equipment is not necessary and the power generation is available with the inverter equipment only. Finally, the proposed model makes a relatively weak contribution to aerodynamic noise and shadow flickering because it does not feature a multi-blade rotation. In the conceptual design, the space required for realizing its motion could be comparatively small in the vertical direction. 3. System modeling 3.1. Shape design of disk-type blade The shape of the blade was designed using CAD and is shown in Fig. 2. The material used for the blade is a polymer-type polycarbonate [20] with a density of 1.12e3 kg/m3 , Young’s modulus of 2.3e9 Pa, and Poisson’s ratio of 0.33. The blade configuration is assumed as follows: the disk-type blade is hollow, and its shell thickness is 10.3 mm. The shape and dimension were determined by blade loads and material costs. For instance, more lift can be generated with a longer airfoil chord length for the blade. However, such a longer length increases the weight of the product, resulting in a non-economical design. Considering these conditions, an analysis of the parameters that affected the shape design was carried out in order to create a lightweight blade with a minimal airfoil chord length. A sensitivity analysis was performed to derive the maximum lift generation shape parameters by varying the geometric shape of the blade. Additionally, the fixed variables and control variables to be changed were accordingly selected. As fixed variables, the radius of the center tower connected to the blade is 0.3 m, thickness is 0.05 m, and inner radius between the tower and disk-type blade is 0.55 m. For the aforementioned fixed parameters, a total of three disk-type blade parameters, including the airfoil chord length, angle of attack (AOA), and NACA series airfoil cross-sectional shape of the blade, were designated as design variables [21–23], as shown in Fig. 3 and listed in Table 1. For each parameter, the three values represent the three levels considered when varying the design of the experiment [24]. The chord lengths were 0.2 m, 0.45 m, and 0.7 m. The values of the AOA were 5°, 10°, and 15°. In the present study, most commonly used NACA series airfoils of NACA632615, NACA64A410, and NACA631412 were selected. In this study, we used a central composite design (CCD) table that allowed the

2

Y. Yoo et al.

Mechatronics 55 (2018) 1–12

Fig. 2. Shape of disk-type blade.

Fig. 3. Geometric model and design parameters of a disk-type blade. Table 1 Design variables of disk-type blade.

x1 x2 x3

Factor

Unit

Level-1

Level-2

Level-3

Chord length Angle of attack Shape of airfoil

m ° NACA

0.20 5 632,615

0.45 10 64A410

0.70 15 631,412

smallest number of numerical experiments [25,26]. In the CCD, the overall number of experiments was 15, wherein the value of 𝛼 for each design variable was ±4.20 in x1 , and ±8.41 in x2 . Particularly in x3 of the NACA series, -𝛼 was selected using the level-1 value of NACA632615, and +𝛼 was selected using the level-3 value of NACA631412. 3.2. Computational fluid dynamics simulation A flow simulation was conducted using ANSYS ICEM-CFD, which is commercial software for computational fluid dynamics (CFD) simulations [27]. The governing equations solved in the fluid domain were the incompressible Navier–Stokes equations with the Reynolds stress term. The Reynolds stress was treated as a shear gradient and eddy viscosity in the Boussinesq approximation. The inlet wind speed in the present study was 12 m/s, which is the rated speed for a small-scale WPS. When the projection diameter of the blade was set to be the characteristic length with this inlet velocity, the Reynolds number became 861,000–2,100,000 considering the wind speed, disk diameter, and air viscosity. This Reynolds number was in the range corresponding to a turbulent flow; therefore, a model of turbulence was required for the numerical simulations. Moreover, the Reynolds averaged Navier–Stokes (RANS) model was not adequate in this situation because the purpose of this study was to observe the pressure generated by the boundary layer flow; however, this model could not resolve the Reynolds stress and over-predicted the flow separation. Therefore, in this study, the RANS turbulence model was used for resolving the Reynolds stress [28–30]. The horizontal lengths of the fluid domain were 6 m and 4 m in the X- and Z-axis directions, respectively. The vertical length was 3 m. The flow analysis domain and boundary condition for the disk-type blade are shown in Fig. 4, wherein the CFD domain size was selected to accommodate the development of a downstream vortex. Using ANSYS ICEM-CFD, we built a tetrahedral mesh in the fluid domain. The unit size for this mesh was 0.012 m, and the ratio of the maximal size to the unit size was 20. The fine mesh near the disk-type blade and tower is 4.05e-13 m3 while the coarse mesh for the region far away from the blade is 6.76e-6 m3 . The pressure generated around the blade is an important factor for aerodynamic performance and the flow field is composed of fine mesh. The coarse mesh is used far-field to shorten the simulation time. There were 7 million mesh cells. The boundary conditions of the flow simulation were location-dependent. The inlet condition was Dirichlet’s boundary condition, according to which the boundary nodes had a constant velocity. The outlet condition was the open boundary condition, which prescribes zero static pressure for no disturbance of the fluid motion by the boundaries. The boundary condition at the surface was the no-slip condition, according to which the flow velocity was zero. Other boundary conditions were related to some initial conditions. Fig. 5 and Table 2 show the CFD simulation results for 3

Y. Yoo et al.

Mechatronics 55 (2018) 1–12

Fig. 4. Flow analysis domain and boundary condition.

Fig. 5. CCD results for lift and drag forces. Table 2 CFD simulation results using CCD method. Case

x1 [m]

x2 [°]

x3 [NACA]

Lift force [N]

Drag force [N]

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

0.2 0.2 0.2 0.2 0.7 0.7 0.7 0.7 0.0295 0.8704 0.45 0.45 0.45 0.45 0.45

5 5 15 15 5 5 15 15 10 10 1.591 18.409 10 10 10

632,615 631,412 632,615 631,412 632,615 631,412 632,615 631,412 64A410 64A410 64A410 64A410 632,615 631,412 64A410

5.85399 4.04579 5.21789 4.70625 12.34182 7.35633 22.19112 12.28102 0.322144 7.13697 9.596121 21.66749 14.1654 7.79639 6.94187

1.883382 1.796428 4.80105 4.482469 7.999032 6.703186 18.77501 17.830649 0.812467 15.39085 2.780241 20.608413 7.464796 7.099322 7.065946

the blade designed by the CCD method. The steady state CFD simulation was conducted to see the pressure distribution around the blade. The lift and drag forces according to the blade shape were calculated instead of thrust force. 3.3. Integrated dynamic analysis using modelica Modelica is an object-oriented simulation language developed to simulate complex and large heterogeneous physics. Using Modelica, it is possible to reuse the components used in one model, easily model a system expressed in mathematical language, and perform a multi-dimensional analysis of the model. In addition, it is more practical to implement a simulation because it has an acausal relationship with the elements of the before and after steps, even if the elements needing modifications are changed during the course of the simulation design through acausal programming [31–33]. To conveniently model the WPS system, including the multi-physics, a study was conducted using a multibody system library that provides a three-dimensional (3D) mechanical component. The main component design of the WPS used the multibody system library [34,35] in Modelica. The two primary forces acting on the disk-type blade are lift and drag. To simulate the lift-generating disk-type WPS, the lift force in the Y-axis direction and drag in the X-axis direction were selected as input values which were derived from computational fluid dynamics. The design was carried out with the force acting on the center of mass of the blade [36]. Because the blade model was considered to be a rigid body in Modelica, the mass, center of mass, and inertia tensor were selected as input variables. These data were applied as parameter values to blade model. The tower was selected as a cylindrical tower with a diameter of 0.25 m and 4

Y. Yoo et al.

Mechatronics 55 (2018) 1–12

Fig. 6. Diagram of crank-rod mechanism. Table 3 Parameters for the PTS and PCS mechanism. Component

Parameter

Unit

Values

Crank

Mass Length Mass Length Mass Radius Moment of inertia Gear ratio Reference torque Moment of inertia

kg m kg m kg m Kg m2 – Nm Kg m2

3.08 1 9.24 4 88.2 0.18 1.43 5 21.3 0.0035

Rod Flywheel

Gear box Generator

height of 10 m. To simulate the dynamic behavior of the blade, it was designed to generate lift and drag when the wind acts on the blade, which was initially prevented. The upper part of the blade was equipped with a spring that vibrates as a result of the wind force to cause periodic motion [37,38]. The spring constant is selected as the following steps: First, a linear type spring is chosen with an initial length of 1 m. Second, the range of spring deflection is determined in accordance with the deflection at the rod end point when the crank rotates; the range of spring deflection is 2–4 m. Third, the spring constant is calculated using blade weight and deflection values, and is selected to derive the maximum power within the deflection for each of blade design Cases. Fourth, the damping constant is assumed to be the same regardless of the blade shape. The conventional PTS has the function of transferring the output torque from the rotation axis of a WPS rotor to a generator [39,40]. However, its configuration and role significantly change depending on the mechanical characteristics of the transmitted torque. The rotor and blade combination is an important factor because it converts the kinetic energy of an air flow into mechanical energy by rotating the rotor shaft. The energy conversion efficiency depends on the rotational force, which in turn affects the performance of the system. However, the disk-type blade is a system that reciprocates vertically along the vertical axis and cannot utilize the rotational force generated at the rotor shaft, as is done with the blades and rotor of a conventional WPS. Therefore, a crank-rod mechanism was considered to convert the motion of the disk-type blade into rotational motion. An object diagram of this system is shown in Fig. 6. From the overall view point of the proposed WPS, the crank-rod mechanism is one of significant components affecting the power generation as well as disk-type blade. The length of the crank with the rod was calculated to allow the crank to rotate in a circle according to the displacement of the blade. The rotational speed of the low-speed shaft was increased by the gearbox ratio, whereas the torque was reduced. The torque on the high-speed shaft connected to the gear model was used for the rotation of the generator. Power was finally obtained from the system using the angular velocity and torque output from the generator. Table 3 lists the parameters for the PTS and PCS mechanism design included in the small-scale WPS [41]. Fig. 7 shows a model of the whole system in OpenModelica. 4. Simulation verification 4.1. Experiment composition A wind tunnel test was conducted to verify the simulation of the WPS designed using Modelica, and the data from the experiment and simulation were compared. Among the 15 blade types, the Case where the most lift occurs was Case 7, and the AOA was 15°. Based on this, the blade was reduced in size by 1:10 and 1:15. The small scale model was manufactured using a stereolithography apparatus (SLA) type 3D printer, and its shape is shown in Fig. 8. An SLA type 3D printer can output models with high precision, light weight, and strength. Table 4 lists the specifications of the small-scale model made with the 3D printer. In the Case of the blade shape derived from the CCD method, the reduced model corresponding to Case 7 was Case C. Based on this, Case A and Case D had the same AOA. To compare various results, Case B was made by greatly increasing the AOA. In addition, Case A and Case B were solid types, and C and D were hollow. 5

Y. Yoo et al.

Mechatronics 55 (2018) 1–12

Fig. 7. Model of lift-generating disk-type wind power system in OpenModelica.

Fig. 8. 3D printing models of disk-type blade. (a) Case A, (b) Case B, (c) Case C and (d) Case D. Table 4 Specifications of small scale model made with 3D printer. Case

A

B

C

D

Angle of attack [deg] Chord length [mm] Weight [g] Type

15 42 69 Solid

45 83 366 Solid

15 61 214.5 Shell

15 82.8 195.5 Shell

The wind tunnel test is shown in Fig. 9. The wind tunnel test equipment was 300 mm in height, 400 mm in width, and 1200 mm in length, and the maximum wind speed was increased to 20 m/s. Experiments were carried out to observe the upward displacement of the fabricated small-scale model blades with varying wind speeds. As the vertical displacement of the disk-type blade increased, upward movement was observed in Case C and Case D, whereas no movement was seen in Case A and Case B. Thus, no displacement result could be obtained. Case A was a 1:15 small scale model. Because the area under pressure was small, the generation of lift for disk movement was insufficient. In Case B, there was no change in motion up to a wind speed of 20 m/s, because the weight increased with the increase in the AOA. In Case C, movement was observed at wind speeds of 12 m/s or more, and it was confirmed that the maximum movement was 110 mm. In Case D, movement was observed at wind speeds of 12 m/s or more, and it rose up to 40 mm, but increased at a much lower rate than Case C. The results for the displacement of the blade are shown in Fig. 10. 6

Y. Yoo et al.

Mechatronics 55 (2018) 1–12

Fig. 9. Wind tunnel test about disk-type blade model.

Fig. 10. Displacement results for different wind speed in wind tunnel test.

Fig. 11. Model of 3D printing disk-type blade in OpenModelica for simulation verification.

4.2. Validation We compared the experimental values from the wind tunnel test and the simulation value for Case C, which showed a large upward movement in the wind speed range of 12–20 m/s. As shown in Fig. 11, the simulated design was constructed considering the wind tunnel test environment. Experiments were performed to obtain the upward displacement of the blades at wind speeds of 12, 15, and 20 m/s. The lifting force derived from the CFD simulation under the same conditions as the experimental environment was applied to Modelica to obtain the upward displacement of the blade. The comparison between the wind tunnel test result and Modelica simulation result is shown in Fig. 12. The initial position of the small-scale 7

Y. Yoo et al.

Mechatronics 55 (2018) 1–12

Fig. 12. Simulation verification of Case C compared with wind tunnel test. Table 5 Simulation verification results compared with wind tunnel test. Wind speed [m/s]

Experimental results

12 15 20

Simulation results

Lift force [N]

Displacement [mm]

Lift force [N]

Displacement [mm]

0.3507 0.7014 1.2859

30 60 110

0.4507 0.6134 1.1845

30.4 53.4 101.1

Table 6 Results of average power according to spring constant and deflection. Case

Spring deflection [m]

Spring constant [N/m]

Average power [kW]

1

2 3 4 2 3 4 2 3 4 2 3 4

95.06 63.37 47.53 76.44 50.96 38.22 1519 1012.6 759.5 1465.1 976.73 732.55

0.456 0.415 0.370 0.439 0.393 0.808e-3 0.002e-3 0 0.530 0.04e-3 0 0.528

2

5

7

model was 90 mm. The simulated time was 60 s considering the actual wind tunnel test time. As the wind speed increased, the experimental results and simulation results became closer to each other. Table 5 lists the displacement results for the blade lift in the experiments and simulations. 5. Results and discussion Two simulation results could be derived from this study. First, we applied the lift derived from the CFD simulation to Modelica and compared the power of the WPS according to the blade shape. Second, we predicted the actual scale WPS power by applying the law of similarity to the lift derived in the wind tunnel test. The first result was derived only from the simulation, whereas the second was derived using data obtained from the experiment. 5.1. Comparison of power according to blade shape Among the 15 types of blades designed using the CCD method, the comparison objects were as follows: 1. Case 2 and Case 7 were selected to compare the Cases that generated the lowest and highest lifts. 2. Case 1 and Case 5 were selected to compare the effects of the blade chord length. The remaining variables were the same. 3. Case 5 and Case 7 were selected to compare the effects of the AOA. The remaining variables were the same. The spring constants used in four blade are as follows: 95.06 N/m for Case 1, 76.44 N/m for Case 2, 759.5 N/m for Case 5 and 732.55 N/m for Case 7. The blade motion frequencies are as follows: 0.49 Hz for Case 1, 0.50 Hz for Case 2, 0.42 Hz for Case 5 and 0.43 Hz for Case 7. The generator power results for the WPS in the four blade Cases are shown in Fig. 13. The average power can be described in terms of lift (L), drag (D), blade velocity components (UX , UY ) and spring constant (k) as follows: 𝑃 = 𝑃 (𝐿, 𝐷, 𝑈𝑋 , 𝑈𝑌 , 𝑘)

(1) 8

Y. Yoo et al.

Mechatronics 55 (2018) 1–12

Fig. 13. Simulation results of generator power output.

The average power according to spring constant and deflection is summarized in Table 6 wherein the spring constant can be obtained the following equation in terms of blade mass (M) and deflection (∆x) for an initial length (xi ). 𝑘=

𝑀𝑔 𝑥𝑖 + Δ𝑥

(2)

To compare the blades with the lowest and highest lifts, the average power for 1 h was obtained. Case 9 was excluded from the Case of the lowest lift because the drag force was large. Case 2 (chord length 0.2 m, AOA 5°) had a value of 0.439 kW, and Case 7 (chord length 0.7 m, AOA 15°) had a value of 0.528 kW. From the simulation results, it can be seen that when a larger lift was generated on the blade, the average power was higher. Case 1 (chord length 0.2 m) had a value of 0.456 kW, and Case 5 (chord length 0.7 m) had a value of 0.530 kW as a result of the blade chord length effects. It can be seen that as the area under the flow pressure was increased among the blades having the same AOA, the lift generated in the blades also increased, resulting in a higher total power for the WPS. Case 5 (AOA 5°) had a value of 0.530 kW, and Case 7 (AOA 15°) had a value of 0.528 kW as a result of the blade’s AOA. The average power of the WPS gets larger for the angle of attack of 5° under the same chord length. 5.2. Power prediction for actual scale WPS Applying the law of similarity based on wind tunnel test, it is possible to predict the starting wind speed and generated lift for the actual scale model, which is the most important factor of the airfoil. The AOA (𝛼), chord length (C), wind speed (U), viscosity (v), and density (𝜌) are considered 9

Y. Yoo et al.

Mechatronics 55 (2018) 1–12

Fig. 14. Simulation results of the actual scale model (rated speed = 12 m/s) using the law of similarity.

as the five major variables affecting the lift, and the lift can be constructed as a function of four dimensionless numbers as follows: ( ) 𝑈𝐶 𝑈 𝑈 𝐿=𝐿 , 𝛼, √ , = 𝐿 (𝑅𝑒, 𝛼, 𝐹 𝑟, 𝑀𝑎) 𝑣 𝜌𝐶 𝑐

(3)

Here, Re is the Reynolds number. Fr is a Froude number, which is a dimensionless number representing the buoyancy due to the density change of the inertia relative to the ambient flow. Ma is the Mach number, which is a dimensionless number indicating the compressibility of the flow. c is a symbol representing the velocity of sound. The largest flow velocity for the reduced and actual scale models is 20 m/s. This is Ma number of 0.015, which is much smaller than the 0.4 value for compressible flow. Therefore, the Ma and Fr numbers are not considered, and the following lift equation is derived: ) ( 𝑈𝐶 𝐿=𝐿 , 𝛼 = 𝐿 (𝑅𝑒, 𝛼) (4) 𝑣 Assuming that the AOA of the small-scale model in the wind tunnel test is the same as that of the actual scale model, the dimensionless number affecting the lift is the Reynolds number. When the Reynolds number is constant and the starting wind speed for the small-scale model is 12 m/s, a value of 1.2 m/s can be derived for the actual scale model. The lift at a wind speed of 20 m/s is the same as the lift at a wind speed of 2 m/s. Applying the law of similarity, the lift force is 0.4629 N at a wind speed of 1.2 m/s, 1.2859 N at 2.0 m/s and 46.2924 N at 12 m/s. The computed lift is applied to the simulation, and the power of the actual scale model WPS is shown in Fig. 14. The blade of the actual scale model is the same as that of Case 7, and the average power is 0.534 kW. We can see that the average power in Case 7 is close to the results (0.528 kW) derived from applying only the CFD data in Section 5.1 and the results obtained using the law of similarity in Section 5.2. 6. Conclusions Considering the various problems of the conventional WPS, a generation system mechanism consisting of a disk-type blade capable of generating wind power with a simple structure was proposed. To determine the shape of the disk-type blade, three design parameters affecting the generation of lift were selected, and 15 blade shapes were constructed by applying the CCD approach. CFD simulation was carried out to derive lift and drag forces data for the designed blades, where a rated wind speed condition of 12 m/s was applied. We modeled the disk-type blade, springs, tower, crank, gearbox, generator, etc. for an integrated simulation and verified the simulation through a wind tunnel test. To compare the WPS power values in relation to the influences of the design variables for the blades, a comparison was made with four blade Cases. Based on the experimental results for the small scale model, the lift prediction and power result of the actual scale model were derived by applying the law of similarity. As a result, the mechanism for a new type of blade for a WPS was established, and the power result for the system was derived by applying the simulation technique. Using Modelica, which is suitable for the integrated simulation of multi-physics, modeling results were obtained without defining equations for complex systems like a WPS. In particular, the advantages of Modelica made it very convenient to create and modify a complex system design by understanding the uses and relationships of the components in one-dimensional modeling. Unlike a traditional blade-rotating type in which the wind is blowing into the swept area, the proposed wind power model utilizes the disk-type blade that vertically lifts up and down so that the existing rotor energy coefficient formula cannot be applied. As the future work, a new disk energymeasured coefficient is to be developed as well as lift coefficient. It is also necessary to devise a noise barrier to prevent from the crank-rod induced structural noise for the further study. Acknowledgments This research is supported by the Basic Science Research Program through the National Research Foundation of Korea (NRF), funded by the Ministry of Science, ICT & Future Planning (2017R1A2B4009606). 10

Y. Yoo et al.

Mechatronics 55 (2018) 1–12

References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] [39] [40] [41]

Blanco MI. The economics of wind energy. Renew Sustain Energy Rev 2009;13(6):1372–82. Bitar EY, Rajagopal R, Khargonekar PP, Poolla K, Varaiya P. Bringing wind energy to market. IEEE Trans Power Syst 2012;27(3):1225–35. Hwangbo H, Johnson A, Ding Y. A production economics analysis for quantifying the efficiency of wind turbines. Wind Energy 2017;20(9):1501–13. Reiner MJ, Zimmer D. Object-oriented modeling of wind turbines and its application for control design based on nonlinear dynamic inversion. Math Comput Model Dyn Syst 2017;23(3):319–40. Jin JH, Park BK. Development of integrated simulation program for artificial satellite operations by Modelica. J Soc Aerosp Syst Eng 2015;9(3):39–46. Open Source Modelica Consortium. OpenModelica user’s guide version 1. 12. 0 , https://openmodelica.org/doc/OpenModelicaUsersGuide/latest/; 2017 [accessed Accessed 11 October 2017]. Modelica Association Modelica – a unified object-oriented language for systems modeling language specification version 3.4, https://www.modelica.org/documents/ModelicaSpec34 .pdf/;; 2017 [Accessed 11 October 2017]. Tiller M. Introduction to physical modeling with modelica. 1st ed. Massachusetts: Kluwer Academic Publishers Press; 2001. Magnani G, Rocco P. Mechatronic analysis of a complex transmission chain for performance optimization in a machine tool. Mechatronics 2010;20(1):85–101. Enge-Rosenblatt O, Schneider P. Modelica wind turbine models with structural changes related to different operating modes. In: Proceedings of the 6th international modelica conference, March 3–4, Bielefeld, Germany; 2008. Strobel M, Vorpahl F, Hillmann C, Gu X, Zuga A, Wihlfahrt U. The onwind modelica library for offshore wind turbines-implementation and first results. In: Proceedings of the 8th international modelica conference, March 20-–22, 2011, Dresden, Germany. Petersson J, Isaksson P, Tummescheit H, Ylikiiskila J. Modeling and simulation of a vertical wind power plant in Dymola/Modelica. In: Proceedings of the 9th international modelica conference, September 3-–5, 2012, Munich, Germany. Eberhart P, Chung TS, Haumer A, Kral C. Open source library for the simulation of wind power plants. In: Proceedings of the 11th international modelica conference, September 21-–23, 2015, Versailles, France. Bolin K, Bluhm G, Eriksson G, Nilsson ME. Infrasound and low frequency noise from wind turbines: exposure and health effects. Environ Res Lett 2011;6(3). Moller H, Pedersen CS. Low-frequency noise from large wind turbines. J Acoust Soc Am 2011;129(6):3727–44. Madsen HA. Low frequency noise from wind turbines mechanisms of generation and its modelling. J Low Freq Noise 2010;29(4):239–51. Wang T, Coton FN. A high resolution tower shadow model for downwind wind turbines. J Wind Eng Ind Aerodyn 2001;89(10):873–92. Khemili I, Romdhane L. Dynamic analysis of a flexible slider-crank mechanism with clearance. Eur J Mech A-Solids 2008;27(5):882–98. Akbari S, Fallahi F, Pirbodaghi T. Dynamic analysis and controller design for a slider-crank mechanism with piezoelectric actuators. J Comput Des Eng 2016;3(4):312–21. Liu Z, Zhan J, Fard M, Davy JL. Acoustic properties of a porous polycarbonate material produced by additive manufacturing. Mater Lett 2016;181:296–9. Schubel PJ, Crossley RJ. Wind turbine blade design. Energies 2012;5(9):3425–49. Benini E, Toffolo A. Optimal design of horizontal-axis wind turbines using blade-element theory and evolutionary computation. J Solar Energy Eng-Trans ASME 2002;124(4):357–63. Shen WZ, Hansem MO, Sørensen JN. Determination of the angle of attack on rotor blades. Wind Energy 2009;12(1):91–8. Montgomery DC. Design and analysis of experiments. 9th ed. New Jersey: John Wiley and Sons Press; 2017. Bettini P, Airoldi A, Sala G, Landro LD, Ruzzene M, Spadoni A. Composite chiral structures for morphing airfoils: numerical analyses and development of a manufacturing process. Compos Part B: Eng 2010;41(2):133–47. Yeom KS, Huh JS, Kwak BM. Development of a structural optimal design code using response surface method implemented on a CAD platform. Trans Korean Soc Mech Eng 2001;1(3):580–5. ANSYS Inc. ANSYS Fluent user’s guide version 12.0. https://users.ugent.be/∼mvbelleg/flug-12-0.pdf; [accessed Accessed 11 October 2017]. Bai Y, Sun D, Lin J, Kennedy D, Williams F. Numerical aerodynamic simulations of a NACA airfoil using CFD with block-iterative coupling and turbulence modeling. Int J Comput Fluid Dyn 2012;26(2):119–32. Walker C, Manera A, Niceno B, Simiano M, Prasser HM. Steady-state RANS-simulations of the mixing in a T-junction. Nucl Eng Des 2010;240(9):2107–15. Frank T, Lifante C, Prasser HM, Menter F. Simulation of turbulent and thermal mixing in T-junctions using URANS and scale-resolving turbulence models in ANSYS CFX. Nucl Eng Des 2010;240(9):2313–28. Elmqvist H, Mattsson SE, Otter M. Modelica - The the new object oriented modeling language. In: The 12th European simulation multiconference, June 16-–19, 1998, Manchester, United Kingdom. Fritzson P. Principles of object-oriented modeling and simulation with modelica 2.1. 1st ed. Linköping: John Wiley and Sons Press; 2004. Mattsson SE, Elmqvist H, Otter M. Physical system modeling with Modelica. Control Eng Pract 1998;6(4):501–10. Otter M, Elmqvist H, Mattsson SE. The new Modelica multiBody library. In: Proceedings of the 3rd international modelica conference, November 3-–4, 2003, Linköping, Sweden. Pujana-Arrese A, Mendizabal A, Arenas J, Prestamero R, Landaluze J. Modelling in Modelica and position control of a 1-DoF set-up powered by pneumatic muscles. Mechatronics 2010;20(5):535–52. Donida F, Ferreti G, Savaresi SM, Tanelli M, Schiavo F. Motorcycle dynamics library in Modelica. In: Proceedings of the 5th international modelica conference, September 4-–5, 2006, Vienna, Austia. Marzouk OA. Characteristic of the flow-induced vibration and forces with 1- and 2-dof vibrations and limiting solid-to-fluid density ratios. J Vib Acoust 2010;132(4). Gabbai RD, Benaroya H. An overview of modeling and experiments of vortex-induce vibration of circular cylinders. J Sound Vib 2005;282(3). Razak AA. Overview of wind turbine modeling in Modelica language. Int J Eng Technol 2012;4(5):551–3. Lim CW. Dynamic response of a 2.75 MW wind turbine applying torque control method based on torque-mode. Korean Soc Fluid Mach J Fluid Mach 2013;16(6):5–11. Lim CW. Design and manufacture of small-scale wind turbine simulator to emulate torque response of MW wind turbine. Int J Precis Eng Manuf-Green Technol 2017;4(4):409–18. Yeongmin Yoo is an Integrated M.S. and Ph.D. student in Mechanical Engineering at Yonsei University. His research interests are on the field of structural analysis, finite element method and integrated system simulation using Modelica.

Soyoung Lee is a M.S. student in Mechanical Engineering at Yonsei University. Her research interests are on the field of mechanical vibration and aeroacoustics analysis using finite element method.

11

Y. Yoo et al.

Mechatronics 55 (2018) 1–12 Jaehyun Yoon is a Ph.D. student in Mechanical Engineering at Yonsei University. His research interests are on the field of aerodynamics, dynamics, control and optimization in robust design of wind power system and multi-rotor air vehicles.

Jongsoo Lee received B.S. and M.S. in Mechanical Engineering at Yonsei University, Seoul, Korea in 1988 and 1990, respectively and Ph.D. in Mechanical Engineering at Rensselaer Polytechnic Institute, Troy, NY in 1996. After a research associate at Rensselaer Rotorcraft Technology Center, he has been a professor of Mechanical Engineering at Yonsei University. His research interests include multi-disciplinary design optimization (MDO), reliability-based robust engineering design, prognostics and health management (PHM) and artificial intelligence & machine learning with applications to structures, fatigue/durability, lifetime prediction, flow induced noise and vibration problems.

12