• Proceedings of the Korean Society of Computer Information Conference
• /
• 2017.07a
• /
• pp.9-10
• /
• 2017

본 논문에서 소음, 진동을 이용한 딥러닝 기반 기계 고장진단 임베디드 시스템을 제안하였다. 제안된 시스템은 기계로부터 취득된 소리와 진동을 바탕으로 학습한 DNN모델을 통해 실시간으로 기계 고장을 진단한다. 딥러닝 기술을 사용하여 학습에 따라 적용대상이 변경될 수 있도록 함으로써 특정 기계에 종속적이지 않고 가변적으로 다양한 기계에 대해 고장 예지 및 건전성 관리를 제공하도록 설계하였으며, 이를 증명하기 위해 액추에이터를 환풍기로 설정하여 정상상태와 4가지 비정상상태의 5가지상태를 학습하여 실험한 결과 93%의 정확도를 얻었다.

• Journal of the Korean Society of Manufacturing Process Engineers
• /
• v.19 no.9
• /
• pp.47-52
• /
• 2020

With the 4th industrial revolution, condition monitoring using machine learning techniques has become popular among researchers. An overload due to complex operations causes several irregularities in MOSFETs. This study investigated the acquired voltage to analyze the overcurrent effects on MOSFETs using a failure mode effect analysis (FMEA). The results indicated that the voltage pattern changes greatly when the current is beyond the threshold value. Several features were extracted from the collected voltage signals that indicate the health state of a switched-mode power supply (SMPS). Then, the data were reduced to a smaller sample space by using a principal component analysis (PCA). A robust machine learning algorithm, the support vector machine (SVM), was used to classify different health states of an SMPS, and the classification results are presented for different parameters. An SVM approach assisted by a PCA algorithm provides a strong fault diagnosis framework for an SMPS.

• Journal of the Computational Structural Engineering Institute of Korea
• /
• v.31 no.1
• /
• pp.63-69
• /
• 2018

Bearing is an essential component in many rotary machineries. To prevent its unpredicted failures and undesired downtime cost, many researches have been made in the field of Prognostics and Health Management(PHM), in which the key issue is to establish a proper feature reflecting its current health state properly at the early stage. However, conventional features have shown some limitations that make them less useful for early diagnostics and prognostics because it tends to increase abruptly at the end of life. This paper proposes a new feature extraction method using the envelope analysis and weighted sum with correlation coefficient. The developed method is demonstrated using the IMS bearing data given by NASA Ames Prognostics Data Repository. Results by the proposed feature are compared with those by conventional approach.

• Journal of the Korean Society of Propulsion Engineers
• /
• v.18 no.6
• /
• pp.50-58
• /
• 2014

This paper is about a short survey on the recent research activities regarding health monitoring and management for liquid rocket engines. For this, we investigate the precedent techniques developed in advanced space-industry countries which are USA, EU, Russia, Japan and China. Particularly, we focus on the technologies applied in China, a recently joined to the advanced space-industry countries in this field. Then we discuss some important points to be considered to apply to the development of the Korea Space Launch Vehicle KSLV-II and other related projects.

• Journal of the Korea Institute of Military Science and Technology
• /
• v.27 no.2
• /
• pp.294-303
• /
• 2024

The objective of condition-based maintenance plus(CBM+) is to improve the availability and maintenance efficiency of missiles, bolstering national defense capabilities. This study proposes an application of CBM+ to enhance the reliability and the safety of missiles, which are the devices typically stored for long durations. CBM+ CBM+ does not only contribute to defense capabilities, but it also aims to reduce maintenance costs. This study focuses particularly on the dormant stage of the missile life-cycle, in which various failure modes and environmental impacts on failure mechanisms are investigated. The effectiveness of maintenance strategies and the implementation of CBM+ is evaluated using simulation data.

• Proceedings of the Korean Society for Noise and Vibration Engineering Conference
• /
• 2014.10a
• /
• pp.533-538
• /
• 2014

Power generator is one of the most important component of electricity generation system to convert mechanical energy to electrical energy. I t designed robustly to maintain high system reliability during operation time. But unexpected failure of the power generator could happen and it cause huge amount of economic and social loss. To keep it from unexpected failure, health prognostics should be carried out In this research, We developed a health prognostic method of stator windings in power generator with statistical data analysis and degradation modeling against water absorption. We divided whole 42 windings into two groups, absorption suspected group and normal group. We built a degradation model of absorption suspected winding using Fick's second law to predict upcoming absorption data. Through the analysis of data of normal group, we could figure out the distribution of data of normal windings. After that, we can properly predict absorption data of normal windings. With data prediction of two groups, we derived upcoming Directional Mahalanobis Distance (DMD) of absorption suspected winding and time vs DMD curve. Finally we drew the probability distribution of Remaining Useful Life of absorption suspected windings.

• Journal of the Korean Society of Manufacturing Process Engineers
• /
• v.20 no.9
• /
• pp.28-34
• /
• 2021

The Fourth Industrial Revolution has led to the development of drones for commercial and private applications. Therefore, the malfunction of drones has become a prominent problem. Failure mode and effect analysis was used in this study to analyze the primary cause of drone failure, and blade breakage was observed to have the highest frequency of failure. This was tested using a vibration sensor placed on drones along the breakage length of the blades. The data exhibited a significant increase in vibration within the drone body for blade fracture length. Principal component analysis was used to reduce the data dimension and classify the state with machine learning algorithms such as support vector machine, k-nearest neighbor, Gaussian naive Bayes, and random forest. The performance of machine learning was higher than 0.95 for the four algorithms in terms of accuracy, precision, recall, and f1-score. A follow-up study on failure prediction will be conducted based on the results of fault diagnosis.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2023.05a
• /
• pp.402-402
• /
• 2023

상수관망은 충분한 양질의 수돗물을 공급하기 위한 사회기반 시설물이다. 상수관로의 노후화, 누수 등은 수도 사고 발생의 가능성을 증가시키고, 수돗물 안전성에 대한 신뢰도를 감소시킨다. 수돗물 공급 전 과정을 인공지능(AI), 정보통신기술(ICT)과 결합한 지능형 상수도관 예측 및 관리 시스템을 구축하여, 상수도 수질 사고를 조기에 감지하고 사전에 취약지점을 예측할 필요가 있다. 이를 위해서는 상수관로의 위험도를 평가하기 위한 체계적인 데이터와 기준이 필요하다. 본 연구에서는 상수관로의 위험도 예측모델을 개발하기 위해 상수관로 위험도와 관련된 평가 인자를 선정하고 분류하였으며, 각 인자의 명확한 기준을 제시하였다. 국내·외 상수도 위험도 평가 항목에 대한 자료를 비교 및 분석하였고, 전문가 자문을 통해 인자를 정립하여 상수관로 위험도 평가 지표를 개발하였다. 개발된 평가 지표의 현장 적용성과 실효성 검증을 위해 정량적인 데이터 확보가 가능하고 상태를 평가할 수 있는 대상 지역을 선정하였다. 문헌 자료의 평가항목들과 전문가 의견을 바탕으로 상수관로 위험도 평가 인자를 31개의 직접 인자와 5개의 간접 인자로 구분하였고, 인자별 평가 기준을 제시하였다. 직접 인자는 노후화 정도를 파악할 수 있는 노후도 평가 항목, 지역 특성을 반영한 토양 부식성 항목, 실시간으로 측정하여 결과를 제공하는 실시간 계측 항목, 직접적인 수질 결과를 제공하는 정수장 수질 항목, 상수관로의 건전성을 평가하는 자산관리 항목으로 분류하였다. 추후, 위험도 평가 운용을 위한 알고리즘을 개발하면 상수도 사고 위험에 대한 예방 및 대응 전략을 수립할 수 있을 것으로 기대된다.

• Journal of the Korea Institute of Military Science and Technology
• /
• v.27 no.1
• /
• pp.1-7
• /
• 2024

Weather is one of the main causes of aircraft accidents, and among the phenomena caused by weather, icing is a phenomenon in which an ice layer is formed when an object exposed to an atmosphere below a freezing temperature collides with supercooled water droplets. If this phenomenon occurs in the rotor blades, it causes defects such as severe vibration in the airframe and eventually leads to loss of control and an accident. Therefore, it is necessary to foresee the icing situation so that it can ascend and descend at an altitude without a freezing point. In this study, vibration data in normal and faulty conditions was acquired, data features were extracted, and vibration was predicted through deep learning-based algorithms such as CNN, LSTM, CNN-LSTM, Transformer, and TCN, and performance was compared to evaluate blade icing. A method for minimizing operating loss is suggested.

• Journal of the Korean Society of Manufacturing Process Engineers
• /
• v.19 no.12
• /
• pp.21-27
• /
• 2020

Lithium-ion batteries are the heart of energy-storing devices and electric vehicles. Owing to their superior qualities, such as high capacity and energy efficiency, they have become quite popular, resulting in an increased demand for failure/damage prevention and useable life maximization. To prevent failure in Lithium-ion batteries, improve their reliability, and ensure productivity, prognosticative measures such as condition monitoring through sensors, condition assessment for failure detection, and remaining useful life prediction through data-driven prognostics and health management approaches have become important topics for research. In this study, the residual useful life of Lithium-ion batteries was predicted using two efficient artificial recurrent neural networks-ong short-term memory (LSTM) and gated recurrent unit (GRU). The proposed approaches were compared for prognostics accuracy and cost-efficiency. It was determined that LSTM showed slightly higher accuracy, whereas GRUs have a computational advantage.