The 10 th Asian International Conference on Fluid Machinery 21 st 23 rd October 2009, Kuala Lumpur Malaysia Paper ID: AICFM0131 Flow Characteristics in an Augmentation Channel of a Direct Drive Turbine for Wave Power Generation Deepak Prasad 1, Mohammed Asid Zullah 1, Young-Do Choi 2 and Young-Ho Lee 3 1 Graduate School, Korea Maritime University 1 Dongsam-dong Youngdo-ku, Busan, 606-791, Korea., dipz@pivlab.net, zullah@pivlab.net 2 School of Mechatronics, Changwon National University #9 Sarim-Dong, Changwon, 641-773, Korea., ydchoi@changwon.ac.kr 3 Division of Mechanical and Information Engineering, Korea Maritime University 1 Dongsam-dong Youngdo-ku, Busan, 606-791, Korea., lyh@hhu.ac.kr Abstract Cross flow turbine also known as Banki turbine, is a hydraulic turbine that may be classified as an impulse turbine. At present it has gained interest in small and low head establishments because of its simple structure, cost effectiveness and low maintenance. Therefore, the present paper expands on this idea and aims at implementing the Direct Drive Turbine (DDT) for wave power generation. Wave power has enormous amount of energy which is environmentally friendly, renewable and can be exploited to satisfy the energy needs. A Numerical Wave Tank (NWT) was used to simulate the sea conditions and after obtaining desired wave properties; the augmentation channel plus the front guide nozzle and rear chamber were integrated to the NWT. The augmentation channel consisted of a front nozzle, rear nozzle and an internal fluid region which represented the turbine housing. The front and rear nozzle were geometrically identical. Two different nozzle configurations were studied; spiral rear wall type and a straight rear wall type. In addition to this, the effect of front guide nozzle divergent angle was also studied. The general idea is to investigate how different augmentation channel geometry and front guide nozzle divergent angle affects the flow, the water horse power and the first stage (primary stage) energy conversion. The analysis was performed using a commercial CFD code of the ANSYS-CFX. The results of the flow in an augmentation channel of the Direct Drive Turbine in oscillating flow for all the cases are presented by means of pressure and velocity vectors. The water horse power (WHP) and first stage energy conversion for the models are also presented. Key Words : Wave power, augmentation channel, direct drive turbine, first stage energy conversion. 1
1. Introduction Power generation utilizing renewable sources has become a common practice recently, reflecting the major threats of climates change due to pollution, exhaustion of fossil fuels, and the environmental, social and political risks of fossil fuels Fortunately, renewable energy sources are available in many countries and this can be exploited to satisfy energy needs with having little or no impact on the environment. Hydro power has always been an important energy resource and wind power has its share of success. However, there exists another source which contains vast amount of energy; its ocean energy. Ocean contains energy in form of thermal energy and mechanical energy: thermal energy from solar radiation and mechanical energy from the waves and tides. However, power generation utilizing waves is presented in this paper. Ocean waves arise from the transfer of energy from the sun to wind and then water. Solar energy creates wind which blows over the ocean, converting wind energy to wave energy. This wave energy can travel thousands of miles with little energy loss. Most importantly, waves are a regular source of power with an intensity that can be accurately predicted several days before their arrival [1]. Wave is available 90% of the time compared to wind and solar resources which are available 30% of the time. In addition to this, wave energy provides somewhat 15 to 20 times more energy per square meter than wind or solar [2]. There is approximately 8,000-80,000 TWh/yr or 1-10 TW of wave energy in the entire ocean [2], and on average, each wave crest transmits 10-50 kw/m. Wave power refers to the energy of ocean surface waves and the capture of that energy to do useful work. There are many energy devices or energy converters available that can be used to extract power from ocean surface waves [3-4] but this study aims at using a Direct Drive Turbine (DDT) for such application. Direct Drive Turbine possesses many advantages; apart from cost effectiveness and ease of construction, it is also self cleaning, there is no problem of cavitation and its efficiency is much less dependent on the flow rate compared to other types of turbine [5]. It is vital to point out that only flow characteristics are studied in the current work without the turbine in a Numerical Wave Tank (NWT). The waves in the numerical wave tank were generated by a piston type wave maker which was located at the wave tank inlet. The inlet which was modeled as a plate wall moved sinusoidally with the general function, x a sin wt. The objective of the present study is to observe the flow field for the different augmentation channel geometry and front guide nozzle configuration. In other words, a free flow case is looked at with oscillating flow. The key focus is on the flow in the augmentation channel. Two different nozzle configurations were modeled; spiral rear wall type and a straight rear wall type. The latter type was studied to see the effect the shape of the rear wall has on the flow. It is stated that the water jet entering the turbine through the nozzle experiences uniform acceleration if the rear wall is spiral shaped [6]. In addition to this, effect of front guide nozzle divergent angle on water horse power and hence the first stage energy conversion is also part of this paper The entire model that is solved in a commercial CFD code of the ANSYS-CFX includes the turbine section (front guide nozzle, augmentation channel and the rear chamber) and the NWT. 2
2. Methodology 2.1 Modeling Three dimensional modeling was carried out using commercial software, UniGraphics NX 4. Fig. 1 shows the turbine section for the spiral wall augmentation channel. The total length of the augmentation channel for both spiral wall and straight wall is 700 mm. The width of the front guide nozzle, the augmentation channel and the rear chamber is also 700 mm in both cases. Figure 2 shows the dimensions for the Spiral wall augmentation channel and Straight wall augmentation channel. The length, height and the width of Numerical Wave Tank (NWT) is 15m, 1.5m and 1 m respectively as shown Fig. 3a. The height of the rear chamber is 1.5m. Figure 3b shows the two different front guide nozzles, the portion in bold line shows the front guide with a divergence angle, α = 7 and the thin dotted line for α = 0. Fig. 1 Turbine section showing various parts (a) and the 3D view (b). Fig. 2 Dimension for the spiral wall augmentation channel (a) and straight wall augmentation channel (b). Fig. 3 Dimension of the numerical wave tank and the turbine section (a) and front guide nozzle configuration (b). 3
2.2 Numerical Method Meshing for all the parts was done in ANSYS ICEM CFD software. For grid generation, hexahedral volume meshes were used. The hexahedral grids are used to ensure that the obtained results are of highest quality that is, high accuracy. The total number of nodes and elements for all the models were 500,000 and 462.000 respectively. Figure 4 shows the grid generation for the respective models. Fig. 4 Meshing for spiral wall turbine section (a), spiral wall augmentation channel (b) and straight wall augmentation channel (c) For the numerical analysis of the entire model (Numerical Wave Tank plus the Turbine Section), a commercial code of the ANSYS CFX was adopted. k Epsilon model was used as the turbulence model and for two phase flow calculations, a homogeneous model was adopted. Waves in the numerical wave tank were generated using a piston type wave maker which was located at the wave tank inlet. To accommodate for this a moving mesh section was employed as shown in Fig. 3. The inlet of the moving mesh section was modeled as a plate wall. The plate was assigned a specific displacement which was sinusoidal motion, as given by equation 1 where a and k are constants and t is the time. x a sin(4 k) (1) The side walls and the bottom walls for this section was assigned unspecified mesh motion. For the top of the moving mesh section the boundary condition was opening with unspecified mesh motion. The rest of the outside sections of the calculation domain were modeled as walls where no-slip boundary conditions are applied. The no-slip condition ensures that the fluid moving over the solid surface does not have a velocity relative to the surface at the point of contact. The boundary condition for the wave tank top and the rear chamber top was opening. At the opening, the relative pressure was set to zero, basically the opening were at atmospheric pressure. Lastly, appropriate interface regions were created. Table 1 shows the four different models that were studied and from now onwards they would be referred as case 1, 2, 3 and 4. Table 1 Model specification. Model Case 0 Straight Wall 1 0 Spiral Wall 2 7 Straight Wall 3 7 Spiral Wall 4 4
3. Results and Discussion The available Water Horse Power (WHP) is given by the equation 2: P WHP gq H (2) V A CS (2 Y ) 2A CS Y Q t t t (3) C f P WHP (4) P Wave ΔH in equation 2 is the head difference across the front nozzle and the rear nozzle. For the given period t, there are two oscillations in the rear chamber that is, the water level rises to a maximum and then falls to a minimum so displacing twice the volume and that is why the multiplying ΔY by 2. C f is the first stage energy conversion factor or the primary energy conversion factor. The water horse power, P WHP was non dimensionalized with the incoming wave power, P Wave. Power in waves was 66W. The key performance indicators for the four models are shown Table 2. Table 2 Performance parameters for the four models. Parameter 0 Divergence Angle 7 Divergence Angle Case 1 Case 2 Case 3 Case 4 H 0.23 0.23 0.23 0.23 t 2.5 2.5 2.5 2.5 Y 0.221 0.234 0.242 0.260 H 0.078 0.081 0.082 0.084 Q 0.0309 0.0327 0.0339 0.0364 P WHP 23.63 25.98 27.2 29.94 C f 0.36 0.39 0.41 0.45 Let s first look at the effect of the front and rear nozzle geometry on the water horse power and the first stage energy conversion. As expected the spiral rear wall nozzle model (Case 2 and Case 4) performed better than the straight rear wall model (Case 1 and Case 3) for the respective front guide nozzle divergent angle. Looking at the 0 divergent models, the 0 Spiral Wall model recorded 10% increase in the available water horse power which represented and increase of approximately 8.3% in the first stage energy conversion compared to the Straight Wall model (Case 1). For the 7 divergent models, the 7 Spiral Wall model recorded P WHP = 29.94W which corresponds to 10% and 9.8% increase in the water horse power and primary energy conversion respectively when compared to 7 Straight Wall model (Case 3). For the straight wall model, when water enters the channel, the fluid near the top wall of the front nozzle right before the internal fluid region seemed to collide with the upper wall (region 1) as shown in Fig. 5a. This results in the 5
severe flow modification and thus leads to lower water horse power. Region 2 also shows the flow colliding with the lower wall of the rear nozzle when water is entering the augmentation channel. When water is advancing re-circulating flow is observed in regions A and B while when water is flowing out of the augmentation channel vortices are observed in regions C and D. For the spiral rear wall model (Case 2 and Case 4) having the rear wall spiral ensures that the fluid enters smoothly and with uniform acceleration which results in better flow as shown in Fig. 5b. The result shows that having the rear spiral wall is advantageous and results in increase in the first stage energy conversion. In terms of the effect of front guide nozzle divergent angle on the water horse power, for both the Spiral Wall model and the Straight Wall model, the primary conversion was more for the front guide with 7 divergent angle. The increases were 13.8% and 15.4% respectively for the 7 Straight Wall model and 7 Spiral Wall model. Fig.5 Velocity vector in the augmentation channel for Case 1(a) and velocity vector in the spiral rear wall nozzle for Case 2 (b). Fig.6 Average velocity for the four cases. Average velocity in the front guide nozzle, the front and rear nozzle is shown in Fig. 6. The velocity right before the front guide inlet drops to 0.09m/s which corresponds to a 55% decrease in the mean velocity compared to the velocity recorded in the wave tank. The velocity then gradually increases in the front guide nozzle. The highest mean velocity was seen in Case 4 while the lowest was observed in Case 1. It is interesting to see that the velocities recorded in the front guide nozzle for 0 Spiral Wall (Case 2) and 7 6
Straight Wall (Case 3) is quite similar however due to moderately bigger flow passage the flow rate for the latter is slightly more and thus high water horse power as shown in Table 2. Fig.7 Average velocity in the front guide nozzle in the XZ plane at y=0. Since the front guide nozzle is symmetric about the x-axis, monitoring points were assigned to half the section as shown in Fig 7. z/w oi = 0 is the centre and z/w oi = 1 is a point on the side wall. Looking at section 3, higher velocity was recorded near the side walls. At section 2 the velocity drops from the centre but then gradually increases from z/w oi = 0.25 onwards to the side wall. At section 1 the variation in velocity from the centre to the side wall is moderate. Making the front guide nozzle divergent has two advantages. Firstly, for the 7 divergence angle front guide nozzle, there is a slight increase in the volume flow rate and also the change in the rear chamber water height. This is due to the slightly larger flow passage at the inlet of the front guide nozzle. Secondly, there is slight increase in the velocity recorded in the front guide nozzle for the respective 7 models than the 0 models. There is 10% increase in the mean velocity recorded in the front guide nozzle for 7 Spiral Wall model and 7 Straight Wall model compared to 0 Spiral Wall model and 0 Straight Wall model respectively. Even for the same divergent angles, the Spiral Wall models (Case 2 & Case 4) recorded higher velocity than the Straight Wall models (Case 1 & Case 3). For α = 0, the velocity recorded in the front guide nozzle for Case 2 is 5.5% higher than Case 1. On the other hand for α = 7, the velocity recorded in the front guide nozzle for Case 4 is 7.7% higher than Case 3. In addition to this, having the front guide nozzle divergent leads to an increase of about 6% and 9% in the mean velocity recorded in the front and rear nozzle when compared to the respective 0 models. 7
Fig.8 Velocity recorded at the periphery for Case 3 and Case 4. Figure 8 shows the velocity recorded at the periphery of the 7 Spiral Wall model and 7 Straight Wall model. The portion in dotted lines is for the straight wall model. It is very clear that the shape of the front nozzle, that is the geometry of the rear wall has a significant effect on the flow. Higher velocity is observed for the 7 Spiral Wall model than the 7 Straight Wall model. The increase is about 8% and for the 0 Spiral Wall model the velocity is 4% higher than 0 Straight Wall model. It is important to note that the velocity presented is the resultant velocity and not components. Apart from this the performance of the two different augmentation channels can also be evaluated with respect to the mean velocity recorded in the numerical wave tank. For the spiral rear wall model the mean velocity in the front and rear nozzle is 70% higher and for the straight rear wall model its 60% higher than that recorded in the wave tank. Again this indicates the benefits of spiral rear wall. The formation of waves in the numerical wave tank is shown by help of volume fraction in Fig. 9. It shows the air/water free surface and the standing wave at the rear wave tank wall. Figure 10 shows the wave height profile in the numerical wave tank for a time period of approximately 25 seconds at a point in the middle of the wave tank and 0.54m below the free surface. Corresponding to this point the mean velocity was 0.2m/s. Fig.9 Volume fraction in the wave tank showing the free surface and the standing wave. 0.2 0.15 0.1 Height ( m ) 0.05 0-0.05-0.1-0.15 0 5 10 15 20 25 Time ( s) 8
Fig. 10 Water wave height in the numerical wave tank. The total pressure in the front nozzle and the rear nozzle for Case 3 and Case 4 is shown in Fig. 11. The pressure variation in the front nozzle is more than the pressure variation in the rear nozzle, approximately 2500Pa and 1300 Pa respectively. The pressure recorded for Case 3 is lower than Case 4 and there is a pressure decrease of about 500Pa in both the front and rear nozzle. In addition to this, 7 Spiral Wall model and 7 Straight Wall model recorded slightly higher pressure compared to 0 Spiral Wall model and 0 Straight Wall model respectively. Pressure ( Pa ) 6750 6500 6250 6000 5750 5500 5250 5000 4750 4500 4250 4000 3750 Case 3 Front Nozzle (P3) Case 3 Rear Nozzle (P4) Case 4 Front Nozzle (P3) Case 4 Rear Nozzle (P4) 0 5 10 15 20 Fig. 11 Total pressure in the front and rear nozzle for Case 3 and Case 4. 4. Conclusion The effect of different augmentation channel geometry and front guide nozzle divergent angle on flow, the water horse power and the first stage (primary stage) energy conversion was successfully investigated. From the present study it is seen that the Spiral Wall models (Case 2 and Case 4) performed better than the Straight Wall models (Case 1 and Case 3) for respective front guide nozzle divergent angles. For α = 0, there is a 10% in the water horse power for Case 2 than Case 1 and for α = 7, Case 4 also recorded an increase of 10% in the water horse power compared to Case 3. In terms of the effect of front guide nozzle divergent angle on primary energy conversion, for both the Spiral Wall model and the Straight wall model, the primary conversion was more for the front guide with 7 divergent angle. The increases were 13.8% and 15.4% respectively. In addition to this, having the rear wall spiral was advantages as it ensured better and smooth flow. For the spiral rear wall model the mean velocity in the front and rear nozzle is 70% higher and for the straight rear wall model its 60% higher than that recorded in the wave tank. Furthermore, the mean velocity recorded in the front guide nozzle for 7 Spiral Wall model and 7 Straight Wall model is 10 % 9
higher than 0 Spiral Wall model and 0 Straight Wall model respectively. Finally, having the front guide nozzle divergent leads to an increase of about 6% and 9% in the mean velocity recorded in the front and rear nozzle when compared to the respective 0 models. A cs C f g H ΔH P Wave P WHP Cross-section area [m 2 ] First stage energy conversion factor Acceleration due to gravity [9.81 m 2 /s] Wave height [m] Head difference [m] Wave power [W] Water horse power [W] Nomenclature Q t V W oi ΔY ρ Volume flow rate [m 3 /s] Period [s] Volume [m 3 ] cross sectional width at section i [m] Rear chamber water level difference [m] Water density [kg/m 3 ] References [1] NOAA Library s Oceanic and Atmospheric Sciences, 2009, Waves and Swells, http//www.lib.noaa.gov/docs/windandsea6.html#waves [2] Wavemill Energy Corp., 2009, Electric Power from Ocean Waves, http//www.wavemill.com [3] Isaacs, J. D., Castel, D. and Wick, G. L., 1976, Utilization of the Energy in Ocean Waves, Ocean Eng, Vol. 3, pp. 175-187. [4] McCormick, M. E., 1974, Analysis of a Wave Energy Conversion Body, J. Hydronaut, Vol. 8, pp 77-82. [5] Olgun, H., 1998, Investigation of the Performance of a Cross Flow Turbine, Int. J. Energy Res. Vol. 22, pp 953-964. [6] Aziz, N. M. and Desai, V. R., 1994, An Experimental Investigation of Cross Flow Turbine Efficiency, J. Fluid Eng., Vol. 116, pp 545-550. 10