This action might not be possible to undo. Are you sure you want to continue?
Technische Universitaet Berlin, Carnotstr. 1A, D-10587, Germany e-mail: firstname.lastname@example.org
Computational ﬂuid dynamics (CFD) analysis has been used to solve the unsteady threedimensional viscous ﬂow in the entire impeller and volute casing of a centrifugal pump. The results of the calculations are used to predict the impeller/volute interaction and to obtain the unsteady pressure distribution in the impeller and volute casing. The calculated unsteady pressure distribution is used to determine the unsteady blade loading. The calculations at the design point and at two off-design points are carried out with a multiple frame of reference and a sliding mesh technique is applied to consider the impeller/volute interaction. The results obtained show that the ﬂow in the impeller and volute casing is periodically unsteady and conﬁrm the circumferential distortion of the pressure distribution at the impeller outlet and in the volute casing. Due to the interaction between impeller blades and the tongue of the volute casing the ﬂow is characterized by pressure ﬂuctuations, which are strong at the impeller outlet and in the vicinity of the tongue. These pressure ﬂuctuations are died away in the casing as the advancement angle increases. These reduced pressure ﬂuctuations are spread to the discharge nozzle; the pressure ﬂuctuations are also reﬂected to the impeller inlet and they affect the mass ﬂow rate through the blade passages. DOI: 10.1115/1.1776587
Single-stage centrifugal turbo-machines are mostly designed with a spiral volute casing. The asymmetric shape of the spiral volute and tongue results in a circumferential distortion of the ﬂow conditions at the outlet of the impeller. This distortion is especially pronounced at off-design points. The circumferential non-uniformity of the pressure ﬁeld causes unbalanced radial forces that must be considered by designing rotor-bearing systems. Due to the relative movement between impeller and volute casing the ﬂow at the outlet of the impeller is strongly interacting with the volute ﬂow. The unsteady interaction between these components generates pressure ﬂuctuations, which are responsible for unsteady dynamic forces. These unsteady dynamic forces give rise to vibration of the pump components and generate hydraulic noises. The physics of the hydrodynamic forces and the reasons of the vibration and noise generation are extensively reported in 1 . Considerable attention has already been focused to study the unsteady interaction in centrifugal turbomachines. Both experimental and numerical approaches have contributed to the understanding of the complex ﬂow ﬂuctuations due to the unsteady interaction. There are numerous examples of the experimental investigations of which Arndt et al. 2 , Kaupert and Staubli 3 , and Hagelstein et al. 4 are a representative sample. In addition, some numerical studies have been undertaken to capture the unsteady interaction and to predict the pressure ﬂuctuations. Some of the studies, e.g., Hillewaert and Van den Braembussche 5 , consider the ﬂow as inviscid and some authors, e.g., Longatte and Kueny 6 , use a two-dimensional model. In recent years, improved computational algorithms as well as hardware development have contributed to enhance CFD capability. It is now feasible to use CFD codes for a realistic prediction of the complex three-dimensional turbulent ﬂow in the entire pump and perform unsteady calcula´ tions see, for example, Zhang et al. 7 and Gonzalez et al. 8,9 . However, the knowledge about the unsteady pressure ﬂuctuations and the unsteady blade loading is still not satisfying. Furthermore,
Contributed by the International Gas Turbine Institute IGTI of THE AMERICAN SOCIETY OF MECHANICAL ENGINEERS for publication in the ASME JOURNAL OF TURBOMACHINERY. Paper presented at the International Gas Turbine and Aeroengine Congress and Exhibition, Vienna, Austria, June 13–17, 2004, Paper No. 2004-GT-54099. Manuscript received by IGTI, October 1, 2003; ﬁnal revision, March 1, 2004. IGTI Review Chair: A. J. Strazisar.
the design of the centrifugal pumps has already reached a level that only through a detailed understanding of the internal ﬂow an increase of the overall performance can be achieved. Due to the curved passages inside the impeller and the volute the ﬂow is to be considered as three-dimensional. Additionally, since the ﬂow following blade passages as well as the volute casing interacts with viscous boundary layers, secondary ﬂows are generated. Therefore, a correct simulation of the impeller/volute interaction requires the simultaneous solution of the three-dimensional unsteady Navier-Stokes equations in both the impeller and volute.
Centrifugal Pump and Test Conditions
The impeller considered in this study is a commercial one. It is shrouded and has ﬁve backswept blades. The blade proﬁle varies between the hub and the shroud. The blade angle at the inlet varies from 18.5 deg from tangential at the shroud to 30.0 deg at the hub. The blade angle at the outlet is 23.5 deg. The single volute casing is unvaned. The shape of the single volute casing is designed according to the theory of a constant average velocity for all sections of the volute Stepanoff, 10 . The main dimensions and characteristic of the investigated pump and the test conditions for this study are presented in Table 1.
Numerical Model and Computational Methods
The numerical simulations have been carried out using the commercial code CFX-TASCﬂow. The ﬂow solver of the code employs for incompressible turbulent ﬂow the continuity equation and the three-dimensional time-averaged Navier-Stokes equations. In this study, the eddy-viscosity assumption is used to model the Reynolds stresses. The eddy viscosity is determined by means of the Standard k- turbulence model. The walls are modeled using a log-law wall function. The transport equations are discretized using an element based conservative ﬁnite volume method. The numerical calculations are carried out with a multiple frame of reference approach, whereby the impeller ﬂow ﬁeld is solved in a rotating frame and the casing in a ﬁxed one. The grid for these two frames of reference should be generated separately. The employed code requires provision of structured or blockstructured grids. Figure 1 shows the computational grids used to model the impeller and the volute. Both of the grids are block structured. The grid of the impeller models all impeller blades and passages. In order to enhance a fully developed ﬂow before enterAPRIL 2005, Vol. 127 Õ 363
Journal of Turbomachinery
Copyright © 2005 by ASME
that the absolute velocity vectors at the inlet are perpendicular to the inlet grid surface and point into axial direction. 9 and Longatte and Kueny 6 . i. and M for midspan). This time step is related to the rotational speed of the impeller and is chosen in such a way that one complete impeller revolution is performed after each 200 time steps. and has 356.. i. that the grids of the impeller and the casing are connected by means of a frozen-rotor interface.. at both off-design points—test conditions 2 and 3. 127. The time step of the unsteady calculations has been set to 2.68 m n 1482 rpm N s 68 min 1 Test Conditions: 1 Mass ﬂow rate 730.e. Gonzalez et al. For the steady-state calculations the following boundary conditions are assumed: At the inlet of the computational domain the mass ﬂow rate. some part of the suction nozzle is also modeled. S for suction side. the impeller blades. The grid for the impeller has 476. Therefore. one for the volute casing and one for the discharge nozzle.5 mm Volute Casing: d 3 523 mm b 3 94. It is assumed.e.34) ˙ ˙ ing the impeller. suggested that the assumption of a ﬁxed mass ﬂow rate at the inlet of the computational domain is physically unsuitable for unsteady calculations and in particular for considering the rotor/stator interaction. 2PM. The turbulent intensity was assumed to be 5%. 1 Grids of the computational domain 364 Õ Vol. and at the outlet the static pressure in a single grid face are speciﬁed. the casing walls and the walls of the suction nozzles. measured from the volute tongue.0243* 10 4 seconds. have been retained unchanged for all test conditions. APRIL 2005 . Therefore. for 1000 time steps. The calculations are carried out for four and half impeller revolutions. The nodal points 1VM and 2VM are located in volute midspan at different angular advancements . The points 1SM and 2SM are located at the suction side ( P stands for pressure side. In the blade passage the points 1PM.04 kg/s 3 Mass ﬂow rate 978. Both in the ﬁxed frame and in the rotating frame the solid walls. for example the time step and the number of iterations in each time step.568 nodes. i. The points 1DM and 2DM are located respectively at begin and end of the discharge nozzle at midspan Transactions of the ASME Fig. Furthermore. As already mentioned the numerical calculations are carried out with a multiple frame of reference.. the eddy length scale was assumed to be 10% of the diameter at the inlet plane. 2 the nodal points are shown exaggeratedly big in order to distinguish them .664 nodes. the appropriate transformation occurs across the interface without any interface averaging.e. The grid of the impeller is generated in 16 blocks. whereby the points 1PM and 2PM are located at the blade pressure side at the inlet and outlet of the blade passage respectively.e. and 2SM are located at midspan of the passage. All parameter settings. The rotor/stator approach accounts for the interaction between two frames. Results and Discussions The results of the unsteady calculations are discussed for different nodal points which are shown in Fig. for 900 time steps. one block for each of the ﬁve impeller passages and each of the regions connecting blade passages with the inlet surface of the volute. For unsteady calculations in a previous study. The nodal points 1TM and 2TM are located directly at the tongue at midspan of the volute casing. This number of iterations is in most cases sufﬁcient to reduce the maximum residuals by three orders of magnitude. at the design point—test condition 1—as well as for ﬁve impeller revolutions. that they change their relative position through the calculation according to the angular velocity of the impeller. the same boundary conditions as for the steady state calculations were assumed. All of these sub grids are pinched H-grids...74) ˙ ˙ (m /m opt 1. in the present study for unsteady calculations the following boundary conditions are used: instead of a ﬁxed mass ﬂow rate at the inlet the total absolute pressure. The grid of the casing is generated in two blocks.. i. ﬁve blocks for the inlet region. the inlet surface of the impeller was extended upstream of the physical inlet region of the blades.e.2 kg/s Impeller outlet diameter Impeller outlet passage width Base circle diameter Volute width at the base circle Mass ﬂow rate Total head Rotational speed Speciﬁc speed (m /m opt 1) ˙ ˙ (m /m opt 0. The number of iterations in each time step has been set to 4.g. The chosen time step is small enough to get the necessary time resolution. therefore only 76% of the nodes are active nodes. The reference change occurs as the ﬂow crosses the interfaces. the eddy length scale and a reference pressure at one grid point are speciﬁed. Some authors. hub and shroud. The average values of residuals rms values reduce by four orders of magnitude. the turbulence intensity. The results are published in 11 . whereby the point 1TM is located at the impeller side of the tongue at 0) and the point 2TM is located at the discharge nozzle side of the tongue..0 kg/s 2 Mass ﬂow rate 541. The blades are deﬁned as solid bodies.e. i.Table 1 Main characteristics of the investigated pump Impeller: d 2 508 mm b 2 72. At the outlet for all variables with exception of pressure a zerogradient condition was assumed. one block for the suction nozzle. i. For unsteady calculations the grids are connected by means of a rotor/ stator interface.0 kg/s ˙ H 46.3 mm Design Point: m opt 730. The nodal points are selected in one blade passage passage 3 and in the volute casing and in the discharge nozzle of the casing. At each test condition at ﬁrst a steady-state calculation is carried out and the result is used to initialize the unsteady calculation at this test point. are modeled using a no-slip boundary condition. e. The two frames of reference are connected in such a way that for steady state calculations the relative position of the impeller and casing does not change through the calculations. 1SM.
i. Vol.74 ˙ ˙ has been computed.34 ˙ ˙ Journal of Turbomachinery APRIL 2005. By the set of boundary conditions used in the present approach.e. a time-average value of 734. The static pressure is normalized using a dynamic pressure based on the impeller outlet tip velocity: Cp p p ref 0.. initiated from steady-state solutions… of the casing. 2 shows the relative position of the impeller and casing for steady state calculations as well as at time 0 of unsteady calculations. For the results presented here.5 u 2 2 . 3 Unsteady pressure distribution at midspan of the blade passage at m Õ m optÉ1. 5 Unsteady pressure distribution at midspan of the blade passage at m Õ m optÉ1. the design mass ﬂow rate can only roughly be adjusted. 2 Location of the investigated nodal points. Furthermore.0 ˙ ˙ Fig. as well as the relative position of the impeller and volute casing at timeÄ0 „starting the unsteady calculations. the time histories of the pressure coefﬁcient Cp at the investigated nodal points inside the blade passage are shown. Fig. 127 Õ 365 . total pressure at the inlet of the computational domain and the static pressure at a single grid face at the outlet. Unsteadiness Inside the Impeller.005) ˙ ˙ Fig.Fig. 4 Unsteady pressure distribution at midspan of the blade passage at m Õ m optÉ0. 3. where the time-average of the computed mass ﬂow rate closely captures the design mass ﬂow rate.1 kg/s (m /m opt 1. (1) Fig. In Fig.
Furthermore.0 ˙ ˙ As it can be seen in Fig. A negative Cp refers to a pressure that is lower than the pressure at the reference point..e. Fig. which is actually typical for mass ﬂow rates in excess of the design mass ﬂow rate. the mass ﬂow rate exceeds marginally the design mass ﬂow rate. 3 the calculations are carried out for four and half impeller revolutions. a pressure surplus on the suction side nodal point 1SM compared to the corresponding values on pressure side nodal point 1PM can be observed.Fig.74 ˙ ˙ Fig. According to Fig.e. for 900 time steps. for 1000 time steps. 8 Unsteady pressure distribution in the volute casing at m Õ m optÉ1. 4 and 5. whereby each impeller revolution needs 200 time steps 40. at the inlet of the blade passage in the immediate vicinity of the leading edge. which is selected at the impeller eye. 9 Unsteady pressure distribution in the volute casing at m Õ m optÉ0. This result indicates a negative blade incidence.74 ˙ ˙ 366 Õ Vol. 7 Unsteady mass ﬂow rate through each blade passage at m Õ m optÉ1.3 ˙ ˙ Fig. And at both off-design points. 3 negative Cp values at the inlet of the blade passage can be observed. APRIL 2005 Transactions of the ASME . i. 3. At the design point Fig. Indeed. Figs.486 ms .. the ﬁnal periodic unsteady solution is achieved after one impeller revolution. 6 Unsteady mass ﬂow rate through each blade passage at m Õ m optÉ0. as already mentioned. for ﬁve impeller revolutions. i. 127.
5.0 ˙ ˙ Fig. it can be seen that the amplitude of the pressure ﬂuctuations within the impeller passage grows in magni- Fig. 11 Unsteady pressure distribution at the tongue of the casing at m Õ m optÉ1. 127 Õ 367 . 3—a nega˙ ˙ tive blade incidence—can also be observed in Fig.34 ˙ ˙ Journal of Turbomachinery APRIL 2005.74 and ˙ ˙ m /m opt 1. 10 Unsteady pressure distribution in the volute casing at m Õ m optÉ1. at the inlet of the blade passage the pressure at the nodal point 1PM is higher than at the suction side nodal point 1SM.34 ˙ ˙ Figures 4 and 5 show the time histories of the pressure coefﬁcients in the blade passage at off-design points m /m opt 0. the amplitude of the pressure ﬂuctuations at a mass ﬂow rate higher than the design Fig. representing a part load operation point. However. 3 with Figs. 4 and 5. but much more intensely. The phenomenon in Fig.34.74 ˙ ˙ tude as the trailing edge of the blade is approached. 3–5.Fig. Considering Figs. respectively. 13 Unsteady pressure distribution at the tongue of the casing at m Õ m optÉ1. 12 Unsteady pressure distribution at the tongue of the casing at m Õ m optÉ0. The Comparison of Fig. According to Fig. 4. shows that the amplitudes of the pressure ﬂuctuations at off-design points are considerably larger than the amplitudes at the design point. Vol.
Fig. which results in a cyclic acceleration and deceleration of the ﬂuid ﬂow inside each blade passage. and secondary ﬂow „c… in the crosssectional plane of the volute casing at angular advancement Ä60 deg at three different time steps at off-design point „ m Õ m optÉ0. the relative position of the passage 1 to the tongue of the volute casing at time step zero as well as exemplary at the begin and at the end of one deceleration and one acceleration-phase are shown. At the upper part of Fig. †13‡ ˙ ˙ point is obviously more critical. 6 and 7. they are linked to the corresponding points of the graph for passage 1. Deviations of the exact repeatability in Figs. are reﬂected upstream to the impeller inlet see the pressure distribution for nodal points 1PM. Since the ﬂow is incompressible. 127. 7. 14 Pressure distribution at midspan of the pump „a…. 6 and 7 are partly numerical. According to these Figs. Comparing the pressure ﬂuctuations at the pressure side nodal points 1PM and 2PM to suction side nodal points 1SMand 2SM reveals that in all operating points the amplitude of the pressure ﬂuctuations on the pressure side of the blade passage is larger than on the suction side. However. also in a real pump an exact repeatability of mass ﬂow rates through Transactions of the ASME . APRIL 2005 through each blade passage varies with time according to the relative position of the blade passage to the tongue of the volute casing. 1SM . Thus. strong pressure ﬂuctuations at the impeller outlet caused by the interaction between the blades and tongue of the volute casing.74…. the mass ﬂow rate 368 Õ Vol. pressure ﬂuctuations at the impeller inlet inﬂuence the mass ﬂow rate through the blade passages. as it can be seen from Figs. pressure distribution „b…. The periodic pressure distribution at the impeller-inlet and outlet leads to a periodic ﬂow.
As each blade approaches the tongue. Contour Plots at Various Instances. These pressure ﬂuctuations are propagated in the volute and are spread to the discharge nozzle. nodal point 1TM.Fig. Unsteadiness Inside the Volute-Casing. 15 Pressure distribution „a… and secondary ﬂow „b… in the cross-sectional plane of the volute casing at angular advancement Ä260 deg at three different time steps at off-design point „ m Õ m optÉ0. By Considering Figs. According to this ﬁgure the tongue and the blades are interacting strongly. points 1VM. e. since the initial values at each impeller revolution vary slightly. very strong pressure ﬂuctuations in the vicinity of the tongue can be observed. Journal of Turbomachinery According to Figs. but is still signiﬁcant at the discharge nozzle 1DM . where both design and off-design points are taken into account. 8–10 it is evident. At the outlet of the discharge nozzle 2DM the pressure ﬂuctuations have almost vanished. At the discharge side of the tongue at nodal point 2TM these pressure ﬂuctuations have died away and thus. Vol. 11–13. The amplitude of these pressure ﬂuctuations die away with advancement of 2VM . the pressure ﬂuctuations at the impeller side of the tongue.74… ˙ ˙ each blade passage at every impeller revolution can not be achieved.. Figure 14 a shows the contour plots of the pressure coefﬁcient at the midspan of the impeller at three different time steps. 8–13. 127 Õ 369 . The unsteady calculations capture the interaction between the tongue and the blades.g. The ﬂuctuations are also reﬂected APRIL 2005. 1 TM. that pressure ﬂuctuations assume most pronounced intensities in points located closest to the tongue. The time histories of the pressure coefﬁcient Cp at the investigated nodal points in the volute casing and at the tongue of the casing are shown in Figs. the amplitudes of the ﬂuctuations are considerably smaller. are strong and have large amplitudes.
The large amplitudes of the pressure ﬂuctuations at impeller outlet and in tongue region die away in the casing as the advancement angle increases. it is obvious that pressure ﬂuctuations die away with increasing angular advancement the colors remain in all time steps almost unchanged . A. 15. 16. C. Nomenclature b2 Cp d2 F radial p u2 outlet passage width of the impeller pressure coefﬁcient outlet diameter of the impeller radial force blade loading pressure ˙ impeller outlet tip velocity density References ¨ ¨ 1 Gulich. 16 .. shown in part (c) of Fig. The secondary vectors are the projections of the calculated velocity vectors onto the investigated cross-sectional plane. 1999. Anlagenplanung und Betrieb. the coefﬁcient in Eq. It was conﬁrmed that due to the interaction between impeller and volute casing the ﬂow is characterized with pressure ﬂuctuations. F. It is to distinguish that at different time steps the pressure ﬂuctuates strongly colors vary strongly .. 112. and Staubli. d 2 is the impeller diameter. T. 1999. A. Since these pressure gradients do not vary with time. K. which is shown in Fig. 2 to different test conditions. To calculate the blade loading the static pressure was integrated over the entire surface of the blades. the secondary ﬂow. at all the time steps there are two pronounced minimum value regions in the pressure distribution of the crosssectional plane. the minimum and maximum values of the color scale of contours remain in all the time steps the same. Using the calculated pressure distribution in the impeller. Acosta. K. which are strong at impeller outlet and at the vicinity of the tongue. These reduced pressure ﬂuctuations are spread to the discharge nozzle and are reﬂected to the impeller inlet. In order to angle distinguish the pressure ﬂuctuations.. Stepanoff has suggested that the radial thrust due to due to the nonuniformity of the circumferential pressure distribution is a function of total head. but the pressure gradients in the cross-sectional plane remain in all time steps almost the same. In this cross-sectional plane only one deﬁcit region can be observed and correspondingly the secondary ﬂow has only one vortex. 16 Unsteady blade loading at different test conditions F radial – Stepanoff K r gHd 2 b 2 . T.3 and 0. J. 16. The blade loading is normalized using the following equation: 370 Õ Vol. they vary according to the shape of the cross-sectional plane and according to the local radius of the spiral volute see. 15.36••••••0. and an experimental coefﬁcient K r Fig. the secondary ﬂows are caused by pressure gradients perpendicular to the ﬂow direction. the pressure coefﬁcient as well as secondary ﬂow velocity vectors are depicted in a cross-sectional plane of the casing at an angular advancement of 260 deg. blade loading appears to assume—as expected—much larger values at off-design conditions 1. which are one of the most important reasons of vibration and hydraulic noises. According to Fig. 1990. unsteady blade loading.. The unsteady pressure distribution results in an unsteady blade loading. On the other hand. and Caghey. Applying Eq. Ein Handbuch fur Entwicklung. The comparison of the numerical results obtained in this study with the published experimental results show a qualitative good agreement of the results considering the behavior of the unsteady pressure ﬂuctuations.74 mass ﬂow ratios . at off design points—specially at 1. (2) where means the density of the ﬂuid. J. 2 Arndt. 14. 127.6 1 m ˙ m opt ˙ 2 . By considering Fig.3 mass ﬂow ratio—the amplitudes of the ﬂuctuations due to ﬂow unsteadiness are equal to the average net blade loading values. impeller diameter. ‘‘Experimental Investigation of Rotor-Stator Interaction in a Centrifugal Pump With Several Vaned Diffusers. E. Kreiselpumpen. 3 Kaupert. This unsteady blade loading gives rise to dynamic effects. Blade Loading Results. impeller width. N. b 2 . The centers of the vortices of the secondary ﬂow in part (c) correspond to the two pronounced minimum value regions in part (b) of the ﬁgure. H. ‘‘The Unsteady Pressure Field in a High Transactions of the ASME . and b 2 is the width of the impeller at the outlet. In Fig. In order to appraise the virtue of the present unsteady computational approach with semi-empirical correlation given by Stepanoff 10 may be used. As already mentioned. These pressure gradients are arising from the passage curvature centrifugal forces and they are in combination with boundary layers responsible for the secondary ﬂows in the volute casing. was calculated.. pp.. (4) Comparing normalized time-averaged loading values Fig. does not vary too. 4 is more or less uncertain in the setup considered here. numerical calculations appear to be less responsive to off-design conditions. The high velocity core in the cross section is driven by the impeller. APRIL 2005 Kr 0. 98 –107. It means contour level scales are consistent between all time steps.5 u 2 d 2 b 2 2 . Brennen.. Part (b) of this ﬁgure shows the distribution of the pressure coefﬁcient at the cross-sectional plane of the volute casing at the advancement 60 deg at the same time steps as in part (a). whereby the pressure gradients depend on the passage curvature. which is one of the most important reasons of vibration.’’ ASME J. the magnitude of blade loading can be estimated. Berlin. for example 12 . d 2 .F normalized F radial 0. Conclusion The results obtained show that the ﬂow ﬁeld in the impeller and volute casing of centrifugal pumps is periodically unsteady. Turbomach.. Springer. (3) The coefﬁcient K r depends on the operating point and reaches its maximum at shutoff: in the impeller and are spread to the inlet region. Using the unsteady pressure distribution inside the impeller.
’’ ASME J. Keiper. 122. 121. 2002.’’ Proceedings of the 10th International Symposium on Transport Phenomena and Dynamics of Rotating Machinery. Paper ISROMAC10-2004-032.’’ ASME J. 597– 606. H. 1999.. Turbomach. ‘‘Unsteady Radial Thrust of a Centrifugal Pump due to the Impeller/Volute Interaction..’’ ASME J. ´ ´ Gonzalez. ‘‘Experimental and Numerical Investigation of the Flow in a Centrifugal Compressor Volute.. Blanco. J. H.. Engeda. ‘‘Numerical Simulation of the Dynamic Effects Due to Impeller-Volute Interaction in a Centrifugal Pump.. R. Vol. 621– 629. Van den Braembussche...’’ ASME Paper FEDSM2002-31181.’’ ASME J. H.. 2004.. K.. M. 127 Õ 371 . Hagelstein.’’ ASME Paper FEDSM99-6866. Rotating Machinery. J. D. Hillewaert.. K. 1999. ‘‘Numerical Calculation of Impeller/Volute Interaction in a Centrifugal Pump.. 1957. and Fernandez. 22–31. FL. J. A. Journal of Turbomachinery APRIL 2005. pp. R. 348 –355.. and Siekmann. pp. 245–252.. Fluids Eng.4 5 6 7 8 Speciﬁc Speed Centrifugal Pump Impeller—Part I: Inﬂuence of the Volute. ‘‘Numerical Analysis of Unsteady Hydrodynamic forces on a Diffuser Pump Impeller due to Rotor-Stator Interaction. 2000.’’ Proceedings of the 5th European Conference on Turbomachinery Fluid Dynamics and Thermodynamics. 10 Stepanoff. pp. pp. 2000. pp. 13 Majidi. M. E. 2002. Longatte.. E. R. pp. 6 4 . Krieger Publishing Company.... Paciﬁc Center of Thermal-Fluids Engineering. and Van den Braembussche. and Rautenberg.. ‘‘Numerical Simulation of Impeller-Volute Interaction in Centrifugal Compressors.. and Kueny. ‘‘Unsteady Flow Structure on a Centrifugal Pump: Experimental and Numerical Approaches. and Tsukamoto.. A. Fluids Eng. E. C. 121.. ‘‘Numerical Calculation of Secondary Flow in Pump Volute and Circular Casing Using 3D Viscous Flow Techniques. K. J. 2003.. 603– 608. Malabar. L. and Santolaria.’’ Int. 12 Majidi. 2002. 124. J. Centrifugal and Axial Flow Pumps. J. ‘‘Analysis of Rotor-Stator-Circuit Interactions in a Centrifugal Pump. Turbomach. J. K. Wang. ´ ´ 9 Gonzalez. A... Santolaria. F. 11 Majidi... Blanco... K. C. Zhang. Hillewaert.’’ ASME Paper FEDSM2002-31182. A.. Fernandez.
This action might not be possible to undo. Are you sure you want to continue?
We've moved you to where you read on your other device.
Get the full title to continue reading from where you left off, or restart the preview.