Reducing underwater radiated noise from a ship is a critical issue for ensuring the survivability of the vessel. As high-speed signal processing and interlocking algorithms become more sophisticated, the heat intensity of shipboard equipment is increasing per unit volume. When designing shipboard equipment, it is necessary to consider the trade-off between heat dissipation and noise reduction. Following an analysis of the trade-offs, it was determined that the arrangement of Fan Ass'y A and B exhibited excellent noise and heat dissipation characteristics. Based on this, PWM control operating zones were derived. It was determined that the placement of Fan Ass'y A and B in the operating zone would increase the PWM duty cycle from 33% to 58% using a signal frequency of 25kHz band with guaranteed reliability. This would increase the noise by approximately 9dB(A) but reduce the internal board reference temperature by up to 15℃.

The noise of large and high-power machines was evaluated and the establishment of mitigation measures was studied. The noise level of large machinery and high-power machinery installed at domestic plant sites was investigated and compared with the noise disclosure regulations to see if they met the standards of the Occupational Safety and Health Act. This investigated the soundproofing design of large and high-power machines and the soundproofing design of complex noise of large machines installed in the plant, and prepared the design standards of the plant design company. In the future, we will compile a database of data to secure standards for research and plant design related to noise reduction, and propose noise improvement and management measures for large and high-power machines.

일상생활의 대화중에 포함되는 잡음, 특히 모든 주파수 대역에 포함되는 백색잡음에 의해 오염된 음성신호는 청각적으로 심한 불쾌감과 거부감을 주며 대화의 명료성을 저해시키는 요인으로 작용할 수 있다. 본 논문은 이러한 잡음환경 하에서 음성인식을 위하여 음성에 부가된 잡음을 제거하는 방범으로 프레임 단위로 웨이브렛 변환을 하여 웨이브렛 계수의 표준편차를 이용하여 시간 적응 임계값을 정하는 새로운 방법을 제안한다. 음성의 특성을 고려하기 위하여 고주파 성분을 많이 가지는 무성음의 경우는 cD1 신호에서, 저주파 성분을 많이 가지는 유성음의 경우는 cA3 신호의 표준편차를 이용하여 시간 적응 임계값을 설정하였다. 실험을 통해 제안한 방법이 일반적인 웨이브렛 변환과 웨이브렛 패킷 변환을 이용한 방법보다 SNR과 MSE 측면에서 향상됨을 확인할 수 있었다. 또한 웨이브렛 변환과 웨이브렛 패킷 변환에서는 파열음, 마찰음 및 파찰음 성분이 많이 제거되는 반면 제안한 방법은 본래 신호와 유사하게 복원됨을 실험 결과 확인할 수 있었다.