• Journal of Broadcast Engineering
• /
• v.27 no.6
• /
• pp.885-896
• /
• 2022

Motor failure in manufacturing plays an important role in future A/S and reliability. Motor failure is detected by measuring sound, current, and vibration. For the data used in this paper, the sound of the car's side mirror motor gear box was used. Motor sound consists of three classes. Sound data is input to the network model through a conversion process through MelSpectrogram. In this paper, various methods were applied, such as data augmentation to improve the performance of classifying fault motors and various methods according to class imbalance were applied resampling, reweighting adjustment, change of loss function and representation learning and classification into two stages. In addition, the curriculum learning method and self-space learning method were compared through a total of five network models such as Bidirectional LSTM Attention, Convolutional Recurrent Neural Network, Multi-Head Attention, Bidirectional Temporal Convolution Network, and Convolution Neural Network, and the optimal configuration was found for motor sound classification.

• Composites Research
• /
• v.17 no.1
• /
• pp.1-9
• /
• 2004

A new non-destructive evaluation method using tapping sound is proposed. This method, named Tapping Sound Analysis, is using the difference between tapping sound data of healthy structure and defective structure as the criteria of determination of internal defect of composite structure. For the characterization of tapping sound, a feature extraction method based on wavelet packet transform is proposed. And a feature index is defined for the decision of existence of internal defects. To prove the possibility of proposed method as a non-destructive evaluation method, experimental study is performed. The tapping sound data of healthy structure and defective structure are measured and compared based on the proposed decision method. The experimental results showed that the feature index is a good indicator for the determination of internal defects.

• International Journal of Automotive Technology
• /
• v.7 no.1
• /
• pp.43-49
• /
• 2006

In this paper, an index for the evaluation of a vehicle intake booming noise and intake sound quality were developed through a correlation analysis and a multiple factor regression analysis of objective measurement and subjective evaluation data. At first, an intake orifice noise was measured at the wide-open throttle test condition. And then, an acoustic transfer function between intake orifice noise and interior noise at the steady state condition was estimated. Simultaneously, subjective evaluation was carried out with a 10-scale score by 8 intake noise and vibration expert evaluators. Next, the correlation analysis between the psychoacoustic parameters derived from the measured data and the subjective evaluation was performed. The most critical factor was determined and the corresponding index for intake booming noise and sound quality are obtained from the multiple factor regression method. And, the optimal design of intake system was studied using the booming noise and the sound quality evaluation index for expectation performance of intake system. Conclusively, the optimal designing parameters of intake system from noise level and sound quality whose point of view were extracted by adapting comparative weighting between the booming noise and sound quality evaluation index, which optimized the process. These work could be represented guideline to system engineers, designers and test engineers about optimization procedure of system performance by considering both of noise level and sound quality.

• Proceedings of the Korean Society for Noise and Vibration Engineering Conference
• /
• 2010.05a
• /
• pp.43-43
• /
• 2010

Presently, there is the dilemma of uncertainty with respect to the evaluation of sound and vibration based on the fact that there is obviously no agreement upon appropriate methods to measure the "truth" concerning the acceptance of sound and vibration. To evaluate properly physical and perception data in sound and vibration it is necessary to implement new methods and innovative approaches to understand the input of human response in sound design. Fortunately, an elaborate dialogue of the usefulness and applicability of those approaches is in progress. Moreover, the need of using and combining perception and physical data in order to comprehend the process of human perception and evaluation sufficiently is widely accepted. However, still the question remains how the goal of an adequate combination can be achieved. Clearly, themultidimensional human perception cannot be easily reduced to singular numbers. Moreover, factors, among others the meaning of the sound, the composition of the diverse sound sources, the listener's attitude, expectations and experiences, are significant parameters which have to be considered to comprehend the different perceptions and evaluations with regard to specific stimuli. Taking under consideration the physical, psychological, and cognitive dimensions as well as the integration of aspects of design require partially various new approaches. While binaural measurement and analysis technologies and psycho-acoustics are well established as they are proved to be valuable auxiliary tools; it has not been achieved to develop generally acceptable measurement units concerning sound quality. Consequently, there is a need for new approaches and methods which make it possible to comprehend sufficiently the process of perception and evaluation. Going with people's mind will be one solution for the future; thisconcept will be introduced based on the development in sound design.

• Proceedings of the Korean Society for Noise and Vibration Engineering Conference
• /
• 2007.05a
• /
• pp.701-705
• /
• 2007

The equipment noise of machine room that generates at the building where human-being is living, is becoming to an object of strong civil appeal, and because it influences to the residential space through the floor or wall, its damage is very serious level. Accordingly, while an efficient forming of the sound insulating measures is earnest, as most of the transmission loss data was the material measured in laboratory, in case when it was applied to the job site, due to the precision difference of constructing work and the influence of detoured transmitting sound, the case of disaccord is the most. Therefore, this thesis has intended to present a fundamental data for an efficient establishment of sound insulation measure, by means of comparison analysis with the existing transmission loss data, after measurement of the transmission loss on the object of various walls, at the construction equipment machinery room.

• Nuclear Engineering and Technology
• /
• v.55 no.2
• /
• pp.506-515
• /
• 2023

This study is to find a method to learn engine sound after the start-up of a diesel generator installed in nuclear power plant with an unsupervised deep learning algorithm (CNN autoencoder) and a new method to predict the failure of a diesel generator using it. In order to learn the sound of a diesel generator with a deep learning algorithm, sound data recorded before and after the start-up of two diesel generators was used. The sound data of 20 min and 2 h were cut into 7 s, and the split sound was converted into a spectrogram image. 1200 and 7200 spectrogram images were created from sound data of 20 min and 2 h, respectively. Using two different deep learning algorithms (CNN autoencoder and binary classification), it was investigated whether the diesel generator post-start sounds were learned as normal. It was possible to accurately determine the post-start sounds as normal and the pre-start sounds as abnormal. It was also confirmed that the deep learning algorithm could detect the virtual abnormal sounds created by mixing the unusual sounds with the post-start sounds. This study showed that the unsupervised anomaly detection algorithm has a good accuracy increased about 3% with comparing to the binary classification algorithm.

• 한국HCI학회:학술대회논문집
• /
• 2008.02a
• /
• pp.184-189
• /
• 2008

In this paper, the sound data either from the sampled or streamed source are utilized for generating a map in the video game play for the dynamiccal use of sound data and synesthesia to users. When users can generate sound in real-time or put the sampled source, it is analyzed and re-processed through Fourier transformation to show the 3D map in dynamic shape over time. We interpolate the generated data to enable the game agents and objects to move.

• The Journal of the Acoustical Society of Korea
• /
• v.35 no.2
• /
• pp.143-148
• /
• 2016

Deep neural network that effectively capture the characteristics of data has been widely used in various applications. However, the amount of sound database is often insufficient for learning the deep neural network properly, so resulting in overfitting problems. In this paper, we propose a transfer learning framework that can effectively train the deep neural network even with insufficient sound event data by employing rich speech or music data. A series of experimental results verify that proposed method performs significantly better than the baseline deep neural network that was trained only with small sound event data.

• Transactions of the Korean Society of Mechanical Engineers
• /
• v.17 no.10
• /
• pp.2426-2436
• /
• 1993

Vibration characteristics of a tire play an important role to judge a ride conformability and sound quality for a passenger car. In this study, the experimental investigation for the sound radiation of a radial tire has been examined. Based on the sound intensity techniques, the sound pressure field and the sound radiation are measured. It turns out that air pressure in tire, tread patterns, and aspect ratio of the tire govern the sound radiation characteristics. Then a numerical analysis for the tire element is conducted. During analysis, the tire element is modelled as an elastic ring. The comparison shows that the numerical output correlates to the experimental data.

• Proceedings of the Korean Society for Noise and Vibration Engineering Conference
• /
• 2002.05a
• /
• pp.719-723
• /
• 2002

The purpose of this study is to suggest the correlation of single-number ratings for sound insulation by floor impact. As a assessment method of impact sound insulation. we selected the IIC contour of ISO, A weighted sound level. Inverse A-weighting curve and L-Index of japanese industrial standard. And we estimated the single-number ratings by application the measured data of impact sound level to each method. The results showed that the coefficients of determination between each two single-number ratings were very high (more than 0.9169). And In the condition of same assessment method, the coefficient of determination for light-weight impact sound was higher than that for heavy-weight impact sound.