Analysis of Secondary Flow Behavior in Low Solidity Cascade Diffuser of a Centrifugal Blower

Author(s):  
Masahiro Ishida ◽  
Tengen Murakami ◽  
Daisaku Sakaguchi ◽  
Hironobu Ueki ◽  
Hiroshi Hayami ◽  
...  

Aiming at reducing noise without deterioration of diffuser performance in a low solidity cascade diffuser (LSD) of a centrifugal blower, the authors have proposed to locate a shallow and short groove or a slot between the diffuser wall and the LSD blade tip limiting to near the blade leading edge. The effect of the LSD blade tip-groove on the blower characteristics and the noise characteristics were investigated experimentally as well as numerically. The mechanism being able to maintain the high LSD blade loading even at small flow rates was pursued in view points of the vortex formation and the induced secondary flow. In addition, the effect of the tip-groove length on the vortex formation in the shroud tip-groove and the secondary flow behavior in the LSD were analyzed numerically and an optimum tip-groove configuration was proposed. It is concluded that formations of the stable and intense vortex in the shroud tip-groove and the recirculating secondary flow along the shroud wall toward the impeller exit are the key factors for achieving a high LSD performance and reducing noise simultaneously at small flow rates.

2009 ◽  
Vol 2009.62 (0) ◽  
pp. 55-56
Author(s):  
Yu KOBA ◽  
Keiichi NAGOSHI ◽  
Tengen MURAKAMI ◽  
Masahiro ISHIDA ◽  
Daisaku SAKAGUCHI ◽  
...  

Author(s):  
Daisaku Sakaguchi ◽  
Masahiro Ishida ◽  
Tengen Murakami ◽  
Hiroshi Hayami ◽  
Yasutoshi Senoo ◽  
...  

A high efficiency and a wide operating range are required in recent centrifugal compressors and blowers. Low solidity circular cascade diffuser (LSD) was proposed in place of vaned diffuser and channel diffuser for achieving both higher pressure ratio and wider operating range. In the present study, aerodynamic performance and noise characteristics of LSD were investigated changing a radial location of LSD leading-edge. It is known that an interaction occurs between jet-wake flow discharged from the rotating impeller and the LSD leading edge, however jet-wake flow pattern is almost uniformalized until the radial position of R = 1.20 in the vaneless diffuser. Aiming at reducing the noise generated by the interaction mentioned above, the LSD leading edge was moved from RLSD = 1.10 to 1.15 and 1.20 in a centrifugal blower with low specific speed. The frequency spectra of noise were analyzed and the noise due to LSD were discussed from the view point of overall noise, discrete frequency noise and broadband noise. In order to clarify the relationship between the flow field and the noise, flow behavior in the impeller as well as in the diffuser was simulated using a Navier-Stokes solver ANSYS-CFX. According to the present experimental results, about 40% improvement in diffuser performance and about 11% improvement in unstable flow range were achieved in small flow rates by the LSD compared with the vaneless diffuser except for a small deterioration in diffuser performance at the large flow rate. On the other hand, the noise increased clearly in the case of RLSD = 1.10 at the large flow rate and the small flow rates as well. By locating the LSD blade leading edge downstream farther from the impeller exit, a remarkable reduction in overall noise was achieved without deterioration of diffuser performance. It is found that the noise increase due to LSD is mainly dependent on the broadband noise based on the jet-wake flow, the interaction between the reverse flow and the impeller blades results in the discrete frequency noise defined by the frequency of cross product of the rotational speed of the impeller and the number of LSD blade, and a significant reduction in noise is achieved by uniformalization of the jet-wake flow.


Author(s):  
Daisaku Sakaguchi ◽  
Hironobu Ueki ◽  
Masahiro Ishida ◽  
Hiroshi Hayami

Low solidity circular cascade diffuser abbreviated by LSD was proposed by Senoo et al. showing a high blade loading or a high lift coefficient without stall even under small flow rate conditions. These high performances were achieved by that the flow separation on the suction surface of the LSD blade was successfully suppressed by the secondary flow formed along the side walls. The higher performance of the LSD was achieved in both pressure recovery and operating range by adopting the tandem cascade because the front blade of the tandem cascade was designed suitably for small flow rates while the rear blade of the tandem cascade was designed suitably for large flow rates. In order to clarify the reason why the tandem cascade could achieve a high pressure recovery in a wide range of flow rate, the flow in the LSD with the tandem cascade is analyzed numerically in the present study by using the commercial CFD code of ANSYS-CFX 13.0. The behavior of the secondary flow is compared between the cases with the single cascade and the tandem one. It is found that the high blade loading of the front blade is achieved at the small flow rate by formation of the favorable secondary flow which suppresses the flow separation on suction surface of the front blade, and the flow separation on pressure surface of the front blade appeared at the design flow rate can be suppressed by the accelerated flow in the gap between the trailing edge of the front blade and the leading edge of the rear blade, resulting in the positive lift coefficient in spite of a large negative angle of attack.


2008 ◽  
Vol 2008 (0) ◽  
pp. 203-204
Author(s):  
Tengen MURAKAMI ◽  
Yu KOBA ◽  
Daisaku SAKAGUCHI ◽  
Masahiro ISHIDA ◽  
Yasutoshi SENOO

Author(s):  
Daisaku Sakaguchi ◽  
Masahiro Ishida ◽  
Hironobu Ueki ◽  
Hiroshi Hayami ◽  
Yasutoshi Senoo

This paper deals with the effect of the blade leading edge location (RLSD) of a low solidity cascade diffuser (LSD) on noise and diffuser performance in a centrifugal blower. The noise of the LSD was measured and analyzed comparing with that of vaneless diffuser (VLD) in view points of overall noise, discrete frequency noise and broadband noise. The numerical flow analysis was conducted in the impeller and the diffuser by using a Navier-Stokes solver. The noise of the VLD varied little in a wide flow rate range, on the other hand, that of the LSD increased remarkably in the small flow rate by about 7 dB. The noise of the LSD did not increase near the design flow and was almost equal to that of the VLD. It was found that the increase in noise due to LSD is dependent mainly on the broadband noise between 600∼1000Hz, which was closely correlated to the lift force of the LSD blade. The two kinds of discrete frequency noise appeared due to an interaction between the rotating impeller and the LSD blade and another interaction between the rotating impeller blades and the reverse flow toward the impeller exit, but their influence on the overall noise were relatively small. By shifting the LSD blade leading edge location downstream from RLSD = 1.1 to 1.2, the noise was reduced by about 3 dB at the maximum without deterioration of the diffuser performance. The maximum lift coefficient of the LSD blade was achieved as high as 1.5 at the high attack angle of 17 degrees even in the case of RLSD = 1.2, resulting in improvement of the diffuser performance by about 40% and in reduction of the unstable flow range by about 11%.


Author(s):  
Huang Chen ◽  
Yuanchao Li ◽  
Joseph Katz

Previous studies have shown that axial casing grooves (ACGs) are effective in delaying the onset of stall, but degrade the performance of axial turbomachines around the best efficiency point (BEP). Our recent experimental study [1] in the JHU refractive index-matched liquid facility have examined the effects of ACGs on delaying stall of a one and half stage compressor. The semicircular ACGs based on Müller et al. [2] reduce the stall flow rate by 40% with a slight decrease in pressure rise at higher flow rates. Stereo-PIV (SPIV) measurements at a flow rate corresponding to the pre-stall condition of the untreated machine have identified three flow features that contribute to the delay in stall. Efficiency measurements conducted as part of the present study show that the ACGs cause a 2.4% peak efficiency loss. They are followed by detailed characterizations of the impact of the ACGs on the flow structure and turbulence in the tip region at high flow rates away from stall. Comparisons with the flow structure without casing grooves and at low flow rate are aimed at exploring relevant flow features that might be associated with the reduced efficiency. The SPIV measurements in several meridional and radial planes show that the periodic inflow into the groove peaks when the rotor blade pressure side (PS) overlaps with the downstream end of the groove, but diminishes when this end faces the blade suction side (SS). The inflow velocity magnitude is substantially lower than that occurring at a flow rate corresponding to the pre-stall conditions of the untreated machine. Yet, entrainment of the PS boundary layer and its vorticity during the inflow phase generates counter-rotating radial vortices at the entrance to the groove, and a “discontinuity” in the appearance of the tip leakage vortex (TLV). While being exposed to the blade SS, the backward tip leakage flow causes flow separation and formation of a counter-rotating vortex at the downstream corner of the groove, which migrates towards the passage with increasing flow rate. Interactions of this corner vortex with the TLV cause fragmentation of the latter, creating a broad area with secondary flows and elevated turbulence level. Consequently, the vorticity shed from the blade tip remains scattered from the groove corner to the blade tip long after the blade clears this groove. The turbulence peaks around the corner vortex, the TLV, and the shear layer connecting it to the SS corner. During periods of inflow, there is a weak outflow from the upstream end of the groove. At other phases, most of the high secondary flows are confined to the downstream corner, leaving only weak internal circulation in the rest of the groove, but with a growing shear layer with elevated (but weak) turbulence originating from the upstream corner. Compared to a smooth endwall, the groove also increases the flow angle near the blade tip leading edge (LE) and varies it periodically. Accordingly, the magnitude of circulation shed from the blade tip and leakage flow increase near the leading edge. The insight from these observations might guide the development of ACGs that take advantage of the effective stall suppression by the ACGs but alleviate the adverse effects at high flowrates.


2006 ◽  
Vol 129 (3) ◽  
pp. 608-618 ◽  
Author(s):  
Hans-Jürgen Rehder ◽  
Axel Dannhauer

Within a European research project, the tip endwall region of low pressure turbine guide vanes with leakage ejection was investigated at DLR in Göttingen. For this purpose a new cascade wind tunnel with three large profiles in the test section and a contoured endwall was designed and built, representing 50% height of a real low pressure turbine stator and simulating the casing flow field of shrouded vanes. The effect of tip leakage flow was simulated by blowing air through a small leakage gap in the endwall just upstream of the vane leading edges. Engine relevant turbulence intensities were adjusted by an active turbulence generator mounted in the test section inlet plane. The experiments were performed with tangential and perpendicular leakage ejection and varying leakage mass flow rates up to 2%. Aerodynamic and thermodynamic measurement techniques were employed. Pressure distribution measurements provided information about the endwall and vane surface pressure field and its variation with leakage flow. Additionally streamline patterns (local shear stress directions) on the walls were detected by oil flow visualization. Downstream traverses with five-hole pyramid type probes allow a survey of the secondary flow behavior in the cascade exit plane. The flow field in the near endwall area downstream of the leakage gap and around the vane leading edges was investigated using a 2D particle image velocimetry system. In order to determine endwall heat transfer distributions, the wall temperatures were measured by an infrared camera system, while heat fluxes at the surfaces were generated with electric operating heating foils. It turned out from the experiments that distinct changes in the secondary flow behavior and endwall heat transfer occur mainly when the leakage mass flow rate is increased from 1% to 2%. Leakage ejection perpendicular to the main flow direction amplifies the secondary flow, in particular the horseshoe vortex, whereas tangential leakage ejection causes a significant reduction of this vortex system. For high leakage mass flow rates the boundary layer flow at the endwall is strongly affected and seems to be highly turbulent, resulting in entirely different heat transfer distributions.


Sign in / Sign up

Export Citation Format

Share Document