이상진단에서 상세 정보 획득을 위한 공기중 초음파 신호 특성 검토 = Investigation of Signal Characteristics of Airborne Ultrasounds to Obtain Detailed Information Regarding Abnormality Detection

국문 초록 (Abstract)

공기중 초음파 기반 이상 유무 탐지는, 비접촉 방식으로 휴대용 장비를 이용하여 비교적 원거리 측정이 가능하다는 장점이 있어 최근 그 활용도가 증가하고 있다. 과거에는 신호 대 잡음비를 향상시키기 위해 공진형 변환기를 적용하는 것이 일반적이었으나, 근래에는 음원위치 추적기술이 결합되어 발생 위치 추정 및공간 필터링을 통한 신호 대 잡음비 개선도 가능해짐에 따라 마이크로폰과 같은 광대역 센서가 사용되고 있다. 이는 넓은 주파수 대역의 신호 특성을 진단에 활용할 수 있으며, 단순히 이상 유무를 판별하는 것에서 이상의 종류 및 정도를 판별하는 것까지 확장할 수 있다는 것을 의미한다. 본 연구에서는 가스 누출 및 부분 방전으로 인해 발생하는 음향 신호를 분석하기 위하여, 이를 시뮬레이션 하는 장치를 만들어 그 특성을 관찰하였다. 또한, 이상 유형을 분류하기 위해 스펙트로그램을 활용한 머신러닝 기반 모델에서 높은 정확도로 진단이 가능한 것을 확인하였다. 이러한 결과로부터, 공기중 초음파 대역의 광대역 측정 결과를 통해 자세한 이상상태와 관련된 보다 다양한 정보를 판단할 수 있을 것으로 기대된다.

다국어 초록 (Multilingual Abstract)

Ultrasound-based abnormality detection is a non-contact method that has the advantage of being able to measure from a relatively distant location with portable equipment. Therefore, their use has been increasing lately. Conventionally, resonant type transducers have been applied to enhance the signal to noise ratio. Recently, this diagnostic technology has been combined with array-based sound source localization techniques, making it possible to not only estimate the location of occurrence but also improve the signal through spatial filtering and to ensure a sufficient signal-to-noise ratio even with a broadband sensor such as a microphone. This implies that the signal characteristics of a wide frequency band can be used for diagnosis and it can be expanded from simply identifying the presence or absence of an abnormality to identifying the type and degree of the abnormality. In this study, the acoustic signals induced by gas leakages and partial discharges were measured and the detailed time and frequency characteristics were observed. For this purpose, simulating devices to create the leakage and partial discharge were designed. It was observed that the signal in the frequency band of interest increased dramatically through spatial and high pass filtering. In addition, it was confirmed that a good decision is made in the machine learning-based model using a spectrogram based on the short-time Fourier transform (STFT) to classify the types of abnormality. Therefore, it is expected that various information related to the detailed conditions of abnormality can be determined using the broadband measurement results of the airborne ultrasonic band.

참고문헌 (Reference)

1 A. G. Edwards, "Vacuum Research and the Mass Spectrometer" 5 : 93-108, 1955

2 D. Wang, "The Ultrasonic Characteristics Study of Weak Gas Leakage" 2015

3 "Standard Practice for Leaks Using Ultrasonics"

4 G. L. Gore, "Radioactive Tracer Techniques" 8 (8): 12-17, 1956

5 "Non-destructive Testing — Acoustic Emission Testing (AT) — Leak Detection by Means of Acoustic Emission"

6 A. Howard, "MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications"

7 M. Florkowski, "Influence of Harmonics on Partial Discharge Measurements and Interpretation of Phase-resolved Patterns" 196 : 111198-, 2022

8 "IEEE Guide for Field Testing and Evaluation of the Insulation of Shielded Power Cable Systems Rated 5 kV and Above"

9 "IEC 61094-4:1995, Measurement Microphones - Part 4: Specifications for Working Standard Microphones"

10 "High-voltage Test Techniques - Partial Discharge Measurements"

11 H. Ilkhechi, "Generation of Acoustic Phase-resolved Partial Discharge Patterns by Utilizing UHF Signals" 113 : 906-915, 2019

12 T. Wang, "Gas Leak Location Detection Based on Data Fusion with Time Difference of Arrival and Energy Decay Using an Ultrasonic Sensor Array" 18 (18): 2985-, 2018

13 T. Wang, "Gas Leak Localization and Detection Method based on a Multi-point Ultrasonic Sensor Array with TDOA Algorithm" 26 (26): 095002-, 2015

14 N. Akashah, "A review : Partial Discharge Detection Using Acoustic Sensor on High Voltage Transformer" 1432 : 012004-, 2020

15 P. Murvay, "A Survey on Gas Leak Detection and Localization Techniques" 25 (25): 966-973, 2012