Interaction of impeller and guide vane in a seriesdesigned

IOP Conference Series: Earth and Environmental Science Interaction of impeller and guide vane in a seriesdesigned axial-flow pump To cite this article: S Kim et al 212 IOP Conf. Ser.: Earth Environ. Sci. 1 3227 View the article online for updates and enhancements. Related content - Design optimization of a centrifugal pump impeller and volute using computational fluid dynamics J H Kim, K T Oh, K B Pyun et al. - Optimization and Inverse Design of Pump Impeller S Miyauchi, B Zhu, X Luo et al. - Numerical study of cavitation flows inside a tubular pumping station X L Tang, W Huang, F J Wang et al. This content was downloaded from IP address 46.3.196.244 on /2/218 at 6:47

IOP Conf. Series: Earth and Environmental Science 1 (212) 3227 doi:1.188/17-131/1/3/3227 Interaction of impeller and guide vane in a series-designed axial-flow pump S Kim 1, Y S Choi 1, K Y Lee 1 and J H Kim 2 1 Green Energy System Technology Center, Korea Institute of Industrial Technology, 3-3, Hongcheon-ri, Ipjang-myeon, Seobuk-gu, Cheonan-si, Chungnam, 331-82, Korea 2 R&D Center, HYOSUNG GOODSPRINGS, 43-1, Ungnam-Dong, Changwon-Si, Gyeongnam, 641-29, Korea E-mail: ks2928@kitech.re.kr Abstract. In this paper, the interaction of the impeller and guide vane in a series-designed axial-flow pump was examined through the implementation of a commercial CFD code. The impeller series design refers to the general design procedure of the base impeller shape which must satisfy the various flow rate and head requirements by changing the impeller setting angle and number of blades of the base impeller. An arc type meridional shape was used to keep the meridional shape of the hub and shroud with various impeller setting angles. The blade angle and the thickness distribution of the impeller were designed as an NACA airfoil type. In the design of the guide vane, it was necessary to consider the outlet flow condition of the impeller with the given setting angle. The meridional shape of the guide vane were designed taking into consideration the setting angle of the impeller, and the blade angle distribution of the guide vane was determined with a traditional design method using vane plane development. In order to achieve the optimum impeller design and guide vane, three-dimensional computational fluid dynamics and the DOE method were applied. The interaction between the impeller and guide vane with different combination set of impeller setting angles and number of impeller blades was addressed by analyzing the flow field of the computational results. 1. Introduction Axial-flow pumps or propeller pumps allow fluid to enter the impeller axially. The impeller can be driven directly by a sealed motor in the pipe or mounted to the pipe from the outside, or by a rightangle drive shaft that pierces the pipe. These pumps discharge fluid nearly axially, pumping the liquid in a direction that is parallel to the pump shaft. Axial-flow pumps are typically used in high-flow rate, low-head applications. For the series design of the impeller, the impeller meridional view was chosen as an arc type and the blade angle of the impeller was designed as an airfoil type. The guide vane of axial-flow pump was designed by analyzing the exit shape and flow-field of the optimally designed impeller. The meridional view of the guide vane was designed by analyzing the hub-tip ratio of the impeller and vane plane development was used in determining the blade angle. Published under licence by Ltd 1

IOP Conf. Series: Earth and Environmental Science 1 (212) 3227 doi:1.188/17-131/1/3/3227 Figure 1. Axial-flow pump3d geometry Figure 2. Meridional plane with various setting angles Figure 3. Meridional design variables DOE was used to analyze the sensitivity of the design variables with regard to the pump performance, as well as the optimum design of the impeller and guide vane. The tendency of performance change according shape was examined using the numerical analysis results, not through an experiment. The interaction between the optimally designed impeller and the guide vane was analyzed using the performance curve, which shows the performance of the axial-flow pump depending on the number of blades as well as the setting angle of the impeller. The performance curve map according to the change of setting angles and the number of blades was predicted in the optimum design model of the impeller. 2. Axial-flow pump design variable Figure 1 shows the three dimensional shape of axial-flow pump. Designing an axial-flow pump consists of designing the impeller and the guide vane. Series designing has been performed for the impeller. The impeller series design makes it possible to choose the ideal impeller shape that satisfies various flow rate and head requirements by changing the impeller setting angle and the blade number on the impeller. For the series designing of the impeller, the design variables of the impeller meridional plane were defined as arc type and the design variables of the blade angle of the impeller were defined as airfoil type. Blade design variables were defined by analyzing NACA4series. The guide vane of the axial-flow pump was designed by analyzing the exit shape and flow-field of the optimally designed impeller. The meridional plane of the guide vane was designed by analyzing the hub-tip ratio of the impeller, and vane plane development was used in defining the blade angle variables of the guide vane because it describes the blade angle distribution simply. 2.1. Impeller design variables 2.1.1. Design variables in a meridional plane. In the meridional plane design of the impeller series, an arc-type meridional shape was used to keep the meridional shape of the hub and shroud with various impeller setting angles, as shown in figure 2. Figure 3 shows the meridional design variables. M_s and M_h indicate the shroud and the hub meridional length respectively. M_m1 and M_m2 indicate the meridional length at 2% and % span 2

IOP Conf. Series: Earth and Environmental Science 1 (212) 3227 doi:1.188/17-131/1/3/3227 (hub: %, shroud: 1%), respectively. The hub-tip ratio, which is defined at the maximum radius, was defined as h_ratio. Figure 4. NACA airfoil variables Figure. Impeller setting angle variables Figure 6. Meridional design variables Figure 7. Vane plane development design variables 2.1.2. Design variables in an airfoil. The NACA airfoils are airfoil shapes for aircraft wings developed by the National Advisory Committee for Aeronautics (NACA). The shape of the NACA airfoils is described using a series of digits following the word "NACA". The parameters in the numerical code can be entered into equations to precisely generate the cross-section of the airfoil and calculate its properties. The NACA four-digit wing sections define the profile by the following parameters[1~2]: First digit describes the maximum camber as a percentage of the chord. Second digit describes the distance of the maximum camber from the leading edge of the airfoil in tens of percents of the chord. Last two digits describe the maximum thickness of the airfoil as a percentage of the chord. The airfoil design variables are defined in figure 4 and. In figure 4, Camber_H means the maximum camber, Camber_D means the location of the maximum camber, and Chord_(h, s) represents the chord length of the hub and the shroud. Set_ in figure indicates the setting angle. The specific speed (rpm, m3/min, m) of the pump is 216, the design-flow rate is 8176 CMH, the total head is 6.1m, and the number of impeller blades is 2EA. 2.2. Guide vane design variables 2.2.1. Design variables in a meridional plane. Figure 6 shows the meridional design variables and the designed meridional plane of the guide vane. GV_z indicates the length of the axial direction of the guide vane. The GV_ratio represents the hub-tip ratio of the guide vane. The hub-tip ratio of guide vane, however, was defined and designed according to the impeller exit, which changes depending on the impeller setting angle, considering the impeller series design. The axial direction length of the guide vane was fixed by analyzing the existing shape. 2.2.2. Design variables in a vane plane development. Figure 7 presents the vane plane development, showing the angles and lengths of the blade. On the figure, "h" means hub, "m" means mid-span, and "s" means shroud. (R*d )_(h, m, s) indicates the total length of the arc at each radius from the front view. dm_(h, m, s) represents the total values of the blade length in the meridional plane. The terms % 1_(h, m, s), and % 2_(h, m, s) show the portion of the blade having the same blade angle at the leading edge and trailing edge respectively, and they are presented as a percentage of length out of the whole length of the y-axis. 1_(h, m, s) means the inlet angle of the blade from the guide vane, 3

IOP Conf. Series: Earth and Environmental Science 1 (212) 3227 doi:1.188/17-131/1/3/3227 while 2_(h, m, s) means the outlet angle of the blade from the guide vane. R2*d was defined as the difference of the sweep angles of the exit between the hub and the shroud[3]. Figure 8. Boundary conditions for the impeller calculation Figure 9. Boundary conditions for the axial-flow pump calculation 3. Numerical analysis method The three-dimensional shape of the impeller and guide vane was generated using the ANSYS CFX- BladeGen program. The structured grid system was generated using ANSYS CFX-TurboGrid, which is a fluid machinery grid generation program[4]. The impeller has 2- blades and the guide vane has 7 blades. By using periodic conditions, flow passages with one impeller blade and one guide vane were combined for the numerical analysis of the axial-flow pump. To analyze the linked interface of the impeller and guide vane, stage-average interface conditions were used. When the numerical analysis was performed for the impeller only, the inlet and outlet part of the impeller was simplified as a straight pipe. Figures 8 and 9 show the boundary conditions for the impeller only and axial-flow pump(impeller and guide vane) calculation. We set the atmospheric pressure on the inlet section of the impeller, and the mass flow rate on the exit section as a boundary condition. The rotational speed of the impeller was 227 rpm. After the mesh test,, grids were used for the impeller and 3, were used for the guide vane. The ANSYS CFX-13, which is a commercial CFD code, was used for the numerical analysis. For the turbulent model, the shear stress transport k- model, which is appropriate for the prediction of flow separation, was used to analyze turbulent flow through the impeller and guide vane. We used water as a working fluid. Roughness and the tip clearance effect are not included in this calculation. 4. Design of experiments Design of experiments is based on a modern analysis of statistics, which aids in the selection of the main cause of abnormal fluctuations from many possible causes. In this study, 2 k factorial designs and the RSM of the design of the experiment were used as numerical optimization methods to optimize the design. Minitab14, a commercial program, was used for the analysis of DOE. 4.1. 2 k factorial designs Generally, 2 k factorial designs are represented as n k, the DOE in which the number of factors is k and the number of levels is n. The advantage of the factorial experiment is that we can assume the main effect (the sole effect of the factor) and interaction effect (the effect between factors) of all factors. It is a convenient screening method that can be used to find the core factor when there are many factors involved at the beginning of the experiment. Response surface analysis is used to determine the changes around the optimum numerical value. In this study, considering the number of factors involved and possible experiments, expenses, and time, we used fractional factorial designs in which the number of experiments is reduced by deleting less meaningful interactions. 4.2. Response surface method We used the RSM to examine the relationship between one or more response factors and quantitative experimental factors or factor groups. The purpose of this analysis is to determine the optimal 4

IOP Conf. Series: Earth and Environmental Science 1 (212) 3227 doi:1.188/17-131/1/3/3227 conditions of a factor for optimizing the response factor. By applying a central composite to the main factor, which greatly influences the response variable, we generated an experimental set and applied RSM. Main Effects Plot (data means) for Ht Main Effects Plot (data means) for t 7. Set_ _h Set_ _m2 Set_ _s Point Type Corner 94. Set_ _h Set_ _m2 Set_ _s Point Type Corner 6. Center 94. Center 6. 93. Mean of Ht 7. 6. 6. -3 3 3-3 -3 3 M_h M_m1 M_m2 Mean of t 94. 94. 93. -3 3 3-3 -3 3 M_h M_m1 M_m2 7. -1 M_s 1-1 1-1 1 h_ratio 94. -1 M_s 1-1 1-1 1 h_ratio 6. 94. 6. 93. -1 1.47..3-1 1.47..3 Figure 1. Main effects plot for Ht Figure 11. Main effects plot for t. Impeller design Figure 12. Plot for response optimization.1. Effect of impeller design variables The selected variables from the 2 k factorial were the ones with a great influence on the axial-flow pump performance among the impeller design variables. M_h, M_m1, M_m2, and M_s, which are the meridional lengths at each radius, and h_ratio, which is the hub-tip ratio, were selected as design variables for the meridional plane. For the airfoil design variables, Set_ _h, Set_ _m2, and Set_ _s, which are the setting angles at the hub, mid-span, and shroud respectively, were selected. The variation range of the design variables M_h, M_m1, M_m2, and M_s were set to be ±1% from each base design value. The range of Set_ _h, Set_ _m2, and Set_ _h were given as ±3. Camber_H was fixed as and Camber_D as 4 according to the 4-digit NACA series definition. In addition, the blade thickness of the hub and shroud was fixed by design restrictive conditions. The influence of the eight impeller design variables on the Ht and t was analyzed using a main effects plot, and is shown in figures 1 and 11. Ht is the total head, and t is the total efficiency. In figure 1, as the design variables Set_ _m2, Set_ _s, M_h, M_m2 and M_s increase, the Ht increases as well. However, M_m1 shows a reverse tendency. M_m1 and h_ratio have very little influence on the Ht. The effects on the t show generally the same tendency as those on Ht, as shown in figure 11. However, the variables M_h and M_m2 have the opposite tendency..2. Optimal impeller design The design variables Set_ _s, Set_ _m2, and M_m2 were selected as the main variables for RSM through a screening procedure using the 2 k factorial. Fifteen numerical analysis experiment conditions were created through the central composite method using these three variables. The other design variables were fixed considering the results of the 2 k factorial design. In RSM, considering the loss of the guide vane, the model with the highest efficiency was selected to be the target of the design in the impeller design-flow rate. The response optimization method was used to specify the model-satisfying target value, and the graph of response optimization was drawn as

IOP Conf. Series: Earth and Environmental Science 1 (212) 3227 doi:1.188/17-131/1/3/3227 shown in figure 12. As a result of response optimization, the Ht was predicted to be 7.23 m and the t was predicted to be 94.63% when Set_ _s is 3, Set_ _m2 is, and M_m2 is 1%. The chosen optimization design was numerically analyzed and the Ht and t are predicted to be 7.32m and 94.7%. Main Effects Plot (data means) for Ht Main Effects Plot (data means) for t 9.84 9.82 i 1_h i 1_s 2_h Point Type Corner Center 92. 92.2 i 1_h i 1_s 2_h Point Type Corner Center 9.8 92. -3 3-3 3 - -3 3-3 3 - Mean of Ht 9.84 9.82 9.8 2_s R2*d % 1 Mean of t 92. 92.2 92. 2_s R2*d % 1 - - 1 - - 1 % 2 % 2 9.84 9.82 9.8 92. 92.2 92. 1 1 Figure 13. Main effects plot for Ht Figure 14. Main effects plot for t Figure 1. Plot for response optimization 6. Guide vane design The number of impeller blades from the previous study was 2. However, it was changed to 3 here because this design was considered to have more advantage in the market. Since the design-flow rate does not change when the number of blades is 3, the guide vane was designed according to the designflow rate that was used in the impeller design 6.1. Effect of guide vane design variables The variables of the vane plane development of the guide vane were chosen through a 2 k factorial experiment. They were defined as the following. Variables i 1_h and i 1_s were the incidence angle, 2_h and 2_s were exit angles, % 1 and % 2 were the identical blade angles from the inlet/outlet of the guide vane. R2*d was defined as the difference of the sweep angles of the exit between the hub and the shroud. For the influence of the design variables of the guide vane on the performance, a main effects plot was used to analyze the influence of these 7 variables as seen in figures 13 and 14. In figure 13, we see the variables % 1, % 2, and i 1_h influence on the Ht in the order of most to least, and the design variables from the inlet part affect the Ht rather than the exit part. We also see that the Ht increases as % 1, % 2, and i 1_h decrease, and i 1_h, i 1_s, % 1, and 2_s increase. In figures 13 and 14, the relationships between t and the variables are expected to be identical with the Ht tendency. In the guide vane design, the impeller has the same tendency, because the same optimally designed impeller model from the previous study is used. By putting together all the analyzed results from the main effects plot, we can determine how much influence the guide vane designing variables have on performance through the 2 k factorial. In addition, it is found that the design variables from the inlet part have substantial influence on the Ht and the t rather than the variables from the exit part. 6.2. Optimal guide vane design Through the previous 2 k factorial, guide vane variables i 1_h, % 1, and 2_h, which have substantial influence on the head and efficiency, were screened, and these 3 variables generated 1 experimental conditions of numerical analysis, including a central value using a central composition. 6

IOP Conf. Series: Earth and Environmental Science 1 (212) 3227 doi:1.188/17-131/1/3/3227 Apart from the 3 main variables, the variables that have good design values in the 2 k factorial were selected and fixed. Because the loss of pressure should be minimized in the guide vane design, the model with the highest efficiency was set as the target of design. The response optimization method was used to specify the shape that satisfies the target value, and the response optimization graph in figure 1 is the result of it. The efficiency was predicted to be 92.8% when i 1_h is 1, % 1 is 11%, and 2_h is. The chosen optimization design was numerically analyzed and the Ht and t are predicted to be 9.9m and 92.82%. Total head [m] Figure 16. Impeller with number of blades 2 imp_12_2ea_gv_18 imp_18_2ea_gv_18 imp_24_2ea_gv_18 imp_3_2ea_gv_18 1 1.6.8 1 1.2 1.4 1.6 1.8 Total efficiency [%] 1 9 9 8 8 7 7 imp_12_2ea_gv_18 6 imp_18_2ea_gv_18 imp_24_2ea_gv_18 imp_3_2ea_gv_18 6.6.8 1 1.2 1.4 1.6 1.8 Total head [m] 1 Figure 17. Impeller with setting angles 2 imp_12_3ea_gv_18 imp_18_3ea_gv_18 imp_24_3ea_gv_18 imp_3_3ea_gv_18 1.6.8 1 1.2 1.4 1.6 1.8 Total efficiency [%] 1 9 9 8 8 7 7 imp_12_3ea_gv_18 6 imp_18_3ea_gv_18 imp_24_3ea_gv_18 imp_3_3ea_gv_18 6.6.8 1 1.2 1.4 1.6 1.8 2 (a) 2EA 1 2 (b) 3EA 1 9 9 Total head [m] 1 1 imp_12_4ea_gv_18 imp_18_4ea_gv_18 imp_24_4ea_gv_18 imp_3_4ea_gv_18.6.8 1 1.2 1.4 1.6 1.8 Total efficiency [%] 9 8 8 7 7 6 6 imp_12_4ea_gv_18 imp_18_4ea_gv_18 imp_24_4ea_gv_18 imp_3_4ea_gv_18.6.8 1 1.2 1.4 1.6 1.8 Total head [m] 1 1 imp_12_ea_gv_18 imp_18_ea_gv_18 imp_24_ea_gv_18 imp_3_ea_gv_18.6.8 1 1.2 1.4 1.6 1.8 Total efficiency [%] 9 8 8 7 7 6 6 imp_12_ea_gv_18 imp_18_ea_gv_18 imp_24_ea_gv_18 imp_3_ea_gv_18.6.8 1 1.2 1.4 1.6 1.8 (c) 4EA Figure 18. Ht and t curve with number of blades (d) EA 7. Interaction between the impeller and guide vane The setting angle (shroud) of the optimally designed impeller is 18.The guide vane designed at the setting angle (18 ) of the optimally designed impeller was studied as well as the tendency of Ht and t, depending on the setting angle and the number of blades from the optimally designed impeller. The Ht and t curve were drawn using the numerical analysis result of the low-flow rate and highflow rate from the design-flow rate. Figure 16 shows the three dimensional shape of the impeller when the number of the blade is 2 to, and figure 17 shows the three dimensional shape when the setting angle of the impeller is 12 to 3. Figure 18, created from the guide vane shape when the impeller setting angle is 18, shows the Ht and 7

IOP Conf. Series: Earth and Environmental Science 1 (212) 3227 doi:1.188/17-131/1/3/3227 t curve according to the change in the number of blades and setting angle of the impeller. In figure 18 (a), the Ht and t curve of 2 impeller blades are shown, (b) is for 3 blades, (c) is for 4 blades, and (d) is for impeller blades. The range of the setting angle change of each case is 12 to 3, and changed by 6 at a time. In the Ht and t curve, the maximum efficiency point moved to the high- flow rate as the setting angle increased, and to the low-flow rate as the setting angle decreased. The maximum efficiency point moved to the low-flow rate as the number of the impeller blade decreased, and it moved to the high-flow rate when the number of blades increased. From the Ht curve, we can determine that the Ht increases, and the inclination also increases as the number of the impeller blades increases. Figure 19. Htcurve map with number of blades and setting angles of the impeller Figure 19 shows the Ht curve in a range of -% from the maximum efficiency point according to the setting angle and the number of the blades of the impeller. Here, the alteration range of the impeller setting angle is 12 to 3. The flow rate ranges of the Ht curve maps become more similar as the number of the impeller blades increases because the range of impeller setting angle is the same. However, we can see that the range of the Ht curve map is increasing as the number of the impellers blade increases. When we put figure 19 together, we can see that the change of the impeller setting angle has an effect on the design-flow rate, and the number of the impeller blades is affects Ht. From the Ht curve map in figure 19, various Ht can be selected from a broad range of flow rates. This shows the possibility of selecting various Ht from a broad flow rate range using the alteration of the setting angle and the number of blades of the impeller. 8. Conclusions The influence of the design variables of the axial-flow pump impeller and guide vane on performance was studied. The optimization model could be induced using the design variables affecting performance. The performance map in accordance with the setting angle and the number of the blades of the impeller also could be described. (1) Our own axial-flow pump could be designed using well-formulated design variables of the impeller and guide vane. (2) The importance of the design variables that affect the performance could be understood using a 2 k factorial. (3) The optimization process was performed to the variables that have a large influence on the total head and the total efficiency using the response surface method. (4) The interaction between the optimally designed impeller and guide vane could be analyzed, and the tendency of the head and the efficiency curve in accordance with the alteration of the setting angle and the number of the blades of the impeller could be analyzed as well. () By describing the performance map in accordance with the setting angle and the number of the blades of the impeller, the possibility of series designing the axial-flow pump could be confirmed. 8

IOP Conf. Series: Earth and Environmental Science 1 (212) 3227 doi:1.188/17-131/1/3/3227 References [1] Stepanoff J 197 Centrifugal and Axial Flow Pumps (John Wiley & Sons).pp138-16 [2] Imaichi K,Murakami Y, Tsurusaki. and Cho K R 22 The Basis of Pump Design, Daiyoungsa. pp 28-26 [3] Kim S, Choi Y S, Lee K Y and Kim J H, 211 J. of Fluid Machinery and Systems 4(1) 14-24 [4] Choi Y S, Kim J H, Lee K Y and Yang S H 21 J. of Fluid Machinery and Systems 3(1) 39-49 9