学位論文

Computational study on prediction and reduction of aerodynamic noise from fans

ファンから生じる空力騒音の予測と低減に関する計算的研究

eng

https://doi.org/10.15002/00014768

1

154

2018-03-24

Not Applicable (or Unknown)

32675乙第229号

2018-03-24

博士(工学)

主査 教授 御法川学, 副査 教授 塚本英明, 副査 教授 林茂

Turbomachinery is machinery device in which energy is delivered either to or from fluid that is continuously moving due to action of moving blades. Performance and flow noise are two major indices for evaluation of turbomachinery. In terms of energy transfer, researches on performance of turbomachinery have been conducted since long time ago; and these researches for performance improvement are still ongoing currently. In addition, flow noise produced by turbomachinery came to the fore as turbomachinery has been used in various fields and everyday life closely and frequently. Especially, consumers’ demand on improvement in affective quality has been increased and regulation on noise has been being reinforced due to damages and adverse effects caused by noise. Therefore, development of high performance and low noise turbomachinery is highly required. Meanwhile, experimental methods have been used to develop low noise turbomachinery; however, the experimental methods solely are not sufficient to achieve such aim since measuring in small turbomachinery is challenging. Hence, prediction technique, to which the numerical analysis method that yields complementary effects in combination with the experimental methods is applied, is required. This study was conducted with the aim of applying numerical analysis method for noise reduction in turbomachinery. Therefore, three-dimensional unsteady Navier-Stokes equations were solved to simulate the flow field. Turbulence models used to predict the flow field were SST k–ω model that provides outstanding simulation of separation and adverse pressure gradient in boundary layer and LES model that presents excellence in turbulence intensity modeling, respectively. Computational Aeroacoustics (CAA) used to predict the flow noise in this study was acoustic analogy that is one of the hybrid methods; and the acoustic analogy is the method analyzing unsteady flow field by using Computational Fluid Dynamics (CFD) and then predicting noise by using the information of unsteady flow field obtained from the results of CFD simulation. To conduct acoustic analogy, Lowson equation, which can be used to predict sound pressure for point force that is moving in a free field, was calculated. Despite of disadvantage that influence of an object including scattering, diffraction, and reflection within acoustic field is difficult to be considered, this method that directly reduces noise sources was able to be drawn since the locations of the noise source can be seen by numerical approach. Because predicting the location of the noise source is able to figure out the unsteady flow which causes the noise. As a result, the reduction method of flow-induced noise in this study is to find the way to reduce or remove the unsteady flow generating the noise, based on CAA and CFD. In order to indicate the location of the noise source, “Aeroacoustic source strength, Ast” was defined and compared with the location of the noise source measured by the acoustic camera to which beamforming technology is applied; and they were agreed qualitatively well each other. Due to miniaturization of electronics and maintenance of fan performance, whereas size of fan is getting smaller, the rotational speed of it getting higher. In this study following the current trend, three fans with each other different type were used for adopting numerical method to noise reduction; ⅰ) a small axial fan of rotor’s diameter D = 0.166 m and a rotation speed 2860 rpm with circular shroud, ⅱ) a small axial fan of rotor’s diameter D = 0.076 m and a rotation speed 7000 rpm with square-type shroud used in a rack mount server computer, ⅲ) a small centrifugal fan with rotor’s diameter D = 0.032 m and a rotation speed 10460 rpm used as a cooling fan in portable home electronics such as a small laptop computer. The noise of each type fan was predicted and compared with the measured noise. The predicted noise and measured noise presented agreement in tonal noise of the blade passing frequency (BPF) and its harmonic frequencies and in the broadband noise at low frequency. Although the broadband noise at high frequency was somewhat different due to random broadband noise, the shapes for noise reduction were able to be drawn effectively by predicting the location of the noise sources. Low noise models suggested for noise reduction provided the result of noise reduction from the prediction and specific noise level was used to evaluate the noise reduction considering the changes in fan performance. In case of the axial flow fan with circular shroud, the interaction between the rotating rotor blades and the flow separated from the inlet of the shroud was found to be the major cause of the noise through the analysis on the location of the noise sources and unsteady flow field. Consequently, reduction of the flow noise was predicted by correcting the shape of the shroud inlet. In the small axial flow fan installed in the rack mount server computer, the tonal noise occurring by irregular clearance between the blade tip and the shroud due to the square-shaped shroud was well predicted. In addition, coherence analysis was conducted to identify the relationship between the surface pressure fluctuation due to the flow and the sound pressure predicted from the microphone. As a result, the correlation for each frequency was well presented. For a centrifugal fan that is used as a cooling fan in home electronics such as a portable small laptop computer, the flow structure of the centrifugal fan was simulated by setting the condition to be analogous to the operating condition within the actual product. And then the reduction of the flow noise was predicted by correcting the tip of the impeller blades based on the location of the noise sources. This study aimed to apply the method of numerical analysis to the noise reduction in turbomachinery. For this, the unsteady flow field was analyzed, the result of noise prediction obtained from the flow filed information was compared and validated, and the location of the noise sources and the structure of the flow field causing the noise were understood; hence, the low noise design was able to be drawn effectively and properly. In this study, the reduction of the flow noise was successfully achieved by adopting the method of numerical analysis and the flow noise of the fan that were improved for noise reduction was predicted to be reduced by 0.8 and 3.7 dB, respectively.

Thesis