• Journal of the Korean Society of Manufacturing Process Engineers
• /
• v.10 no.4
• /
• pp.70-75
• /
• 2011

Nowadays, modern gearboxes are characterized by high torque load demands, low running noise and compact design. Also durability of gearbox is specially a major issue for the industry. For the gearbox which used in wind turbine, gear transmission error(T.E.) is the excitation that leads the tonal noise known as gear whine, and radiated gear whine is also the dominant source of noise in the whole gearbox. In this paper, tooth modification for the high speed stage is used to compensate for the deformation of the teeth due to load and to ensure a proper meshing to achieve an optimized tooth contact pattern. The gearbox is firstly modeled in Romax software, and then the various combination analysis of the tooth modification is presented by using Windows LDP software, and the prediction of transmission error under the loaded torque for the helical gear pair is investigated, the transmission error, contact stress, root stress and load distribution are also calculated and compared before and after tooth modification under one torque condition. The simulation result shows that the transmission error and stress under the loads can be minimized by the appropriate tooth modification.

• The Journal of the Acoustical Society of Korea
• /
• v.33 no.1
• /
• pp.40-47
• /
• 2014

Interaction between the unsteady flow emitted from the blade of the centrifugal fan and the volute tongue region of fan duct is known as the main noise source of the centrifugal fan. In this paper, the relative contributions of the volute tongue region of the centrifugal fan is analyzed to utilize as the foundation data of low noise design. The internal hybrid CAA (Computational Aero-Acoustics) method is used to predict noise radiated from the main noise source. This method is the noise prediction technique using CFD (Computational Fluid Dynamics), Acoustic analogy, and BEM(Boundary Element Method). The relative contributions of the centrifugal fan volute tongue region using the hybrid CAA method show that the region between the cut-off and the scroll has high contribution than the region between the cut-off and the outlet and the hub region of blade has high contribution than the shroud region of blade. These results is utilized as the important data for the development of low noise centrifugal fan.

• The Journal of the Acoustical Society of Korea
• /
• v.31 no.6
• /
• pp.391-398
• /
• 2012

In this paper, prediction of centrifugal fan was conducted through combination the hybrid CAA method which was used to predict the fan noise with the FRPM technique which was used to generate the broadband noise source. Firstly, flow field surround the centrifugal fan was computed using the RANS equations and noise source region was deducted from the computed flow field. Then the FRPM technique was applied to the source region for generation of turbulence which satisfies the stochastic features. The noise source of the centrifugal fan was modeled by applying the acoustic analogy to the synthesized flow field from the computed and generated flow fields. Finally, the broadband noise of the centrifugal fan was predicted through combination the modeled noise source with the linear propagation which was realized using the boundary element method. It was confirmed that the proposed technique is efficient to predict the tonal and broadband noises of centrifugal fan through comparison with the measured data.

• Journal of Mechanical Science and Technology
• /
• v.17 no.4
• /
• pp.590-598
• /
• 2003

A computational aero-acoustic (CAA) method is used to predict the tonal noise generated from a cavity of automobile door seals or gaps at low flow Mach numbers (A$\_$$\infty$/=0.077 and 0.147) In the present method, the acoustically perturbed Euler equations are solved with the acoustic source term obtained from the unsteady incompressible Navier-Stokes calculations of the cavity flow in self-sustained oscillations. The aerodynamic and acoustic fields are computed for the Reynolds numbers based on the displacement thickness, Re$\_$$\delta$*/=850 and 1620 and their fundamental mode characteristics are investigated. The present method is also verified with the experimentally measured sound pressure level (SPL) spectra.

• The Journal of the Acoustical Society of Korea
• /
• v.5 no.1
• /
• pp.11-16
• /
• 1986

지금까지 알려진 여러 가지 음성부호화 방식들 중 4.8에서 9.6kbits/s 사이의 전송속도에서 제일 좋은 성능을 갖는 것은 Residual-Excited linear prediction 방식이다. RELP 부호화 방식은 전송속도가 낮을 때 합성음이 거칠거나 금속성의 잡음을 갖는 단점이 있다. 본 논문에서는 이러한 단점을 보완하여 음질을 개선하는 세가지의 방법들을 제안하며, 그들은 다음과 같다. 첫째는 여러개의 baseband를 이용한 spectral folding 방법이고, 둘째는 spectral folding 방법과 pulsed excitation 방법을 조합한 방법이며, 마 지막 방법은 여러개의 baseband를 사용한 spectral folding 방법과 pulsed excitation 방법을 조합한 방법 이다. 이 방법들을 사용하여 RELP vocoder의 음질을 많이 개선할 수 있으며, 9.6kbits/s 근처의 전송속 도에서 사용하기 위한 첫 번째 방법과 세 번째 방법은 spectral fording 이나 nonlinear distortion 방법 에서 문제가 되는 roughness 나 tonal noise를 거의 인지 할 수 없으며, 세 번째 방법이 첫 번째 방법보 다 우수하다. 두 번째 방법은 4.8 kbits/s 근처의 전송속도에 적합하며, 기존의 RELP 방식들에 비해 많 은 음질향상을 가져왔다. 제안한 세가지 방법들을 같은 조건에서 비교할 때 세 번째 방법이 가장 우수 하며, 이 경우 합성음은 원음과 거의 흡사하다.

• Journal of KSNVE
• /
• v.9 no.5
• /
• pp.991-997
• /
• 1999

It is getting higher the demand on a good sound in a car. Like some other sensory qualities, achieving a good sound is still treated as a black-box followed by the individual taste of tuning engineers. This study aims to define the attributes of a good car audio sound and the assessment factors affecting the sound quality. With a variety of surveys on professional listening groups and car audio installers, 26 most frequently used vocabularies are selected as typical presentations of car audio sound quality. After repetitions of listening tests with deliberately selected test CD, pre-recorded 15 different digital car-recordings, headphones replay system, and three different listening groups, it is proved that there are at least 5 different assessment factors which significantly affect the car audio sound quality. They are high fidelity, spectral balance, tonal accuracy, sound stage and acoustic fault.

• Proceedings of the Acoustical Society of Korea Conference
• /
• 1994.06a
• /
• pp.650-655
• /
• 1994

The persence of distinct tones in environmental noises leads to an increasing anaoyance especially if the tones are of long duration. Before 1992 in Germany the tonal content of noises had to be subjectively estimated by the acoustic consultants and up to six dB could be added to the measured Leq depending on the tonality content of the noise under consideration. In order to give an objective basis for tonality estimated a DIN norm proposal for tonality evaluations was introduced in 1992. This proposal is compared with two different procedures : the prominece-ratio calculation which was proposed by Bienvenue and Nobile /4/ and the proposal of Ares /5/. The output of the three models is contrasted with subjective tonality judgements. It turns out that the prominence ratio model yields the best agreement with the subjective tonality assessments od the sound set chosen.

• Proceedings of the Acoustical Society of Korea Conference
• /
• spring
• /
• pp.541-544
• /
• 2002

피치 (pitch)는 주파수와 관련된 인간이 실제 느끼는 음의 높이로 라우드니스 (loudness), 음색 (timbre)과 함께 소리의 음질을 결정하는 중요한 요소로 알려져 있다. 이러한 피치는 음성 해석 및 분리를 위해 많은 연구가 이루어진 반면 소음 분석 및 음질 향상을 위한 방향으로의 연구는 부족한 상황이다. 본 연구에서는 기저막 (basilar membrane)의 위치에 따른 주파수 분리 이론인 위치이론 (place theory)을 기본으로 한 기존의 가상 피치 (virtual pitch) 분석 알고리즘을 소음에 적용하기 위해서 수정하고, 절러가지 소음에 적용하였다. 본 연구에서의 알고리즘은 소음의 주파수 특성에 의존한 방법이기 때문에, 토널 (tonal) 성분이 존재하는 소음의 적용에는 적합한 결과를 나타냈지만, 그 이외의 소음에 대해서는 정확한 분석이 어렵다. 따라서 기본 주파수 (fundamental frequency)와 이와 관련된 고조파음(harmonics)이 음질에 중요한 영향을 미치는 소음의 음질 해석 린 음질 향상을 위해 본 연구의 알고리즘에 의한 피치 분석과 기존의 음질 인자를 적용하면 보다 효율적인 결과를 얻을 수 있을 것이다. 이런 소음의 예로는 엔진의 부밍 소음이나 기어 whine 소음 등이다.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2020.07a
• /
• pp.343-345
• /
• 2020

본 논문에서는 선택적 노이즈 캔슬링을 위한 환경 인지 기술을 제안한다. 기존의 노이즈 캔슬링은 모든 소리를 구분 없이 차단하여 여러 가지 문제를 유발할 수 있으며 공통된 노이즈 캔슬링 동작으로 각 소음에 최적화된 성능을 보장할 수 없다. 이러한 문제를 해결하기 위해 제안하는 방법은 대표적 오디오 특성인 멜-스펙트로그램과 스펙트로그램 기반의 시간적 특성 벡터를 사용하여 환경 인지를 진행한다. 본 논문에서는 attack, rotation, sawing으로 구성된 3가지 소음과 speech, tonal로 구성된 2가지 비 소음으로 총 5가지 클래스를 분류한다. 제안하는 방법에서 특성 벡터로 멜-스펙트로그램만을 사용했을 때 87.5%의 분류 성능을 보였으며, 스펙트로그램 기반의 시간적 특성을 추가했을 때 분류 성능이 91.2%로 향상되었다.

• The Journal of the Acoustical Society of Korea
• /
• v.35 no.3
• /
• pp.216-222
• /
• 2016

Generation of a quiet zone in noisy environment is undoubtedly of considerable realistic significance. This paper describes development and implementation of a multichannel real-time active noise control (ANC) system for 3 dimensional noisy environment to enhance the quiet zone performance in terms of size and noise cancellation gain. The proposed ANC system employes a multichannel delay-compensated filtered-X least mean square (FXLMS) algorithm; its real-time implementation is designed in TMS320C6713 digital signal processor (DSP) board. The system is evaluated for cancelling various tonal frequency noises in the range from 100 to 500 Hz, and the performance is then illustrated by measuring the quiet zone in terms of sound pressure level (SPL) attenuation. Experiment results show that a quiet zone of quiet with satisfactory size and maximum 24 dB noise attenuation is successfully generated.