• Proceedings of the KSR Conference
• /
• 2008.06a
• /
• pp.647-652
• /
• 2008

Railway signal system has been more complex, larger and required high reliability. So, maintenance by experience must be changed to optimize maintenance program or introduced systematic method for estabilish new maintenance program. In this paper, we introduced the maintenance period decision method which are Age based method and Block replacement method based on the failure distribution for the equipment. So, we allocated optimum maintenacne period for the railway signal equipment using block replacement method.

• Investigative Magnetic Resonance Imaging
• /
• v.21 no.3
• /
• pp.125-130
• /
• 2017

Purpose: To evaluate the clinical significance of T1 high signal intensity on the globus pallidus as a predictor of severe hepatic encephalopathy in patients with acute-on-chronic liver failure (ACLF), which is a distinct syndrome characterized by multi-organ dysfunction including cerebral failure. Materials and Methods: From January 2002 to April 2014, we retrospectively reviewed the magnetic resonance imaging (MRI) findings and clinical and magnetic resonance (MR) features of 74 consecutive patients (44 men and 30 women; mean age, 59.5 years) with liver cirrhosis. The chronic liver failure-sequential organ failure assessment score was used to diagnose ACLF. The pallidal index (PI), calculated by dividing the mean signal intensity of the globus pallidus by that of the subcortical frontal white matter were compared according to ACLF. The PI was compared with the Model for End-Stage Liver Disease (MELD) score in predicting the development of ACLF. Results: Fifteen patients who were diagnosed with ACLF had higher hepatic encephalopathy grades (initial, P = 0.024; follow-up, P = 0.002), MELD scores (P < 0.001), and PI (P = 0.048). In the ACLF group, the mean PI in patients with cerebral failure was significantly higher than that in the patients without cerebral failure (1.33 vs. 1.20, P = 0.039). In patients with ACLF, the area under the curve (AUC) for PI was 0.680 (95% confidence intervals [CI], 0.52-0.85), which was significantly lower than that for the MELD score (AUC, 0.88; 95% CI, 0.77-0.99) (P = 0.04). Conclusion: The PI can be an ancillary biomarker for predicting the development of ACLF and severe hepatic encephalopathy.

• Nuclear Engineering and Technology
• /
• v.54 no.4
• /
• pp.1230-1244
• /
• 2022

This paper proposes an algorithm for signal validation using unsupervised methods in emergency situations at nuclear power plants (NPPs) when signals are rapidly changing. The algorithm aims to determine the stuck failures of signals in real time based on a variational auto-encoder (VAE), which employs unsupervised learning, and long short-term memory (LSTM). The application of unsupervised learning enables the algorithm to detect a wide range of stuck failures, even those that are not trained. First, this paper discusses the potential failure modes of signals in NPPs and reviews previous studies conducted on signal validation. Then, an algorithm for detecting signal failures is proposed by applying LSTM and VAE. To overcome the typical problems of unsupervised learning processes, such as trainability and performance issues, several optimizations are carried out to select the inputs, determine the hyper-parameters of the network, and establish the thresholds to identify signal failures. Finally, the proposed algorithm is validated and demonstrated using a compact nuclear simulator.

• Journal of the Institute of Electronics and Information Engineers
• /
• v.50 no.9
• /
• pp.144-155
• /
• 2013

This paper presents a low-cost remote power-quality-failure monitoring system (RPMS) using Android App and TI MCU (micro-controller unit), which is appliable to a micro-grid. The designed RPMS testbed consists of smart nodes, a server, and Android APPs. Especially, the C2000-series MCU-based RPMS smart node that is low-cost compared to existing monitoring systems has both a signal processing function for power signal processing and a data transmission function for power-quality monitoring data transmission. The signal processing function implements both a wavelet-based power failure detection algorithm including sag, swell, and interruption, and a FFT-based power failure detection algorithm including harmonics such that reliable and real-time power quality monitoring is guaranteed. The data transmission function implements a low-complexity RPMS transmission protocol and defines a simple data format (msg_Diag) for power monitoring message transmission. We may watch the monitoring data in real time both at a server and Android phone Apps connected to the WiFi network (or WAN). We use RS-232 (or Bluetooth) as the wired (or wireless) communication media between a server and nodes. We program the RPMS power-quality-failure monitoring algorithm using C language in the CCS (Code Composer Studio) 3.3 environment.

• 제어로봇시스템학회:학술대회논문집
• /
• 2003.10a
• /
• pp.142-145
• /
• 2003

Many control applications using Neural Network need a priori information about the objective system. But it is impossible to get exact information about the objective system in real world. To solve this problem, several control methods were proposed. Reinforcement learning control using neural network is one of them. Basically reinforcement learning control doesn't need a priori information of objective system. This method uses reinforcement signal from interaction of objective system and environment and observable states of objective system as input data. But many methods take too much time to apply to real-world. So we focus on faster learning to apply reinforcement learning control to real-world. Two data types are used for reinforcement learning. One is reinforcement signal data. It has only two fixed scalar values that are assigned for each success and fail state. The other is observable state data. There are infinitive states in real-world system. So the number of observable state data is also infinitive. This requires too much learning time for applying to real-world. So we try to reduce the number of observable states by classification of states with Self-Organizing Map. We also use neural dynamic programming for controller design. An inverted pendulum on the cart system is simulated. Failure signal is used for reinforcement signal. The failure signal occurs when the pendulum angle or cart position deviate from the defined control range. The control objective is to maintain the balanced pole and centered cart. And four states that is, position and velocity of cart, angle and angular velocity of pole are used for state signal. Learning controller is composed of serial connection of Self-Organizing Map and two Multi-layer Feed-Forward Neural Networks.

• Journal of the Korean Society of Manufacturing Technology Engineers
• /
• v.24 no.2
• /
• pp.203-210
• /
• 2015

This paper investigates the failure of a car with a 2.5-liter CRDi engine of the Hyundai Company. The failure is caused by intermittent poor acceleration while driving. To analyze the cause, we investigated the air intake volume, the fuel injection, and the air-fuel ratio, which were determined to be normal. The brake switch signal error was discovered while analyzing the function that limits the output of the engine. While investigating the cause, we discovered the corrosion of the pins on the connector of the brake switch. We determined that it was generated by soapy water flowing in the solar film. Therefore, the cause of the failure was the brake switch signal errors. Additionally, we determined that ECM was the normal fail-safe mode that implemented the override device for safety during normal acceleration. Based on these results, further solar film experiments must be conducted to fully elucidate the causes.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.17 no.8
• /
• pp.805-815
• /
• 1992

In this paper, robust decoupling control scheme of miftivarlable systems Is studied. Design methods for Input-Output decoupling systems with robustness against signal failure In arbitrary feedback loop or actuator loop Is suggested based on the Riccati type matrix equation and state feedback, and is simulated In Turbo-Generator systems with B-Input, 2 output. The results of simulation represents the decoupled and stable response against the failure of signal In sensor or actuator loop. However, the system designed by conventional ,it ate feedback shows the unstable response. This method Is applied for robust decoupling control of the complicated multivariable systems.

• Geomechanics and Engineering
• /
• v.34 no.1
• /
• pp.1-15
• /
• 2023

This study conducts uniaxial compression tests on intact and single crack-contained rocks to investigate the time-frequency domain characteristics of acoustic emission (AE) signals monitored during the deformation failure process. A processing approach, short-time Fourier transform (STFT), is performed to obtain the evolution characteristics of time-frequency domain of AE signals. The AE signal modes at different deformation stages of rocks are different. Five modes of AE signal are observed during the cracking process of rocks. The evolution characteristics of time-frequency domain of AE signals processed by STFT can be utilized to evaluate the damage process of rocks. The difference of time-frequency domain characteristics between intact and cracked rocks is comparatively analyzed. The distribution characteristics of frequency changing from a single band-shaped cluster to multiple band-shaped clusters can be regarded as an early warning information of damage and failure of rocks. Meanwhile, the attenuation of frequency enables the exploration of rock failure trends.

• Proceedings of the Korean Society of Tribologists and Lubrication Engineers Conference
• /
• 2002.10b
• /
• pp.191-192
• /
• 2002

The scuffing failure is a critical problem in modern machine components, especially for the requirement of high efficiency and small size. In this study. scuffing experiments are conducted using Acoustic Emission(AE) measurement by an indirect sensing approach to detect scuffing failure. Using AE signals we con get and indication about the state of the friction processes, about the quality of solid and liquid layers on the contacting surface in real time. The FFT(Fast Fourier Transform)analyses of the AE signal are used to understand the interfacial interaction and the relationship between the AE signal and the state of contact is presented.

• Proceedings of the Korean Society of Precision Engineering Conference
• /
• 2002.10a
• /
• pp.883-887
• /
• 2002

We should maintain the maximum operation capacity for production facilities and find properly out the fault of each equipment rapidly in order to decrease a loss caused by its failure. The acoustic and vibration signals of a machine always carry the dynamic information of the machine. These signals are very useful fur the feature extraction and fault diagnosis. We performed a fundamental study which develops a system of fault diagnosis for a pump. We experimented vibrations by acceleration sensors and noises by microphones, compared and analysed for normal products, artificially deformed products. We tried to search a change of the dynamic signals according to machine malfunctions and analyse the type of deformation or failure. The results showed that acoustic signals as well as vibration signals can be used as a simple method lot a detection of machine malfunction or fault diagnosis.