• Proceedings of the Korean Society of Propulsion Engineers Conference
• /
• 2003.10a
• /
• pp.267-270
• /
• 2003

An intelligent performance diagnostic computer program of a gas turbine using the NN(Neural Network) was developed. Recently on-condition performance monitoring of major gas path components using the GPA(Gas Path Analysis) method has been performed in analyzing of engine faults. However because the types and severities of engine faults are various and complex, it is not easy that all fault conditions of the engine would be monitored only by the GPA approach. Therefore in order to solve this problem, application of using the NNs for learning and diagnosis would be required. Among then, a BPN (Back Propagation Neural Network) with one hidden layer, which can use an updating learning rate, was proposed for diagnostics of PT6A-62 turboprop engine in this work.

• Proceedings of the Korea Institute of Fire Science and Engineering Conference
• /
• 1997.11a
• /
• pp.252-259
• /
• 1997

A method to automate hazard analysis of chemical plants is proposed in this paper. The proposed system is composed of three knowledge bases - unit knowledge base, organizational knowledge base and material knowledge base, and three hazard analysis algorithms - deviation, malfunction and accident analysis algorithm. Hazard analysis inference procedure is developed based on the actual hazard analysis procedures and accident development sequence. The proposed algorithm can perform hazard analysis in two methods and represent all conceivable types of accidents using accident analysis algorithm. In addition, it provides intermediate steps in the accident propagation, and enables the analysis result to give a useful information to hazard assessment. The proposed method is successfully demonstrated by being applied to diammonium phosphate manufacturing process. A system to automate hazard analysis is developed by using the suggested method. The developed system is expected to be useful in finding the propagation path of a fault or the cause of a malfunction as it is capable to approach causes of faults and malfunctions simultaneously.

• Journal of Korean Society of Coastal and Ocean Engineers
• /
• v.18 no.4
• /
• pp.269-282
• /
• 2006

We numerically studied tsunami propagation characteristics through Korean Straits based on nonlinear shallow water equation, a robust wave driver of the near field tsunamis. Tsunamis are presumed to be generated by the earthquake in Tsuhima-Koto fault line. The magnitude of earthquake is chosen to be 7.5 on Richter scale, which corresponds to most plausible one around Korean peninsula. It turns out that it takes only 60 minutes for leading waves to cross Korean straits, which supports recently raised concerns at warning system might be malfunctioned due to the lack of evacuation time. We also numerically obtained the probability of tsunami inundation of various levels, usually referred as tsunami hazard, along southern coastal area of Korean Peninsula based on simple seismological and Kajiura (1963)'s hydrodynamic model due to tsunami-generative earthquake in Tsuhima-Koto fault line. Using observed data at Akita and Fukaura during Okushiri tsunami in 1993, we verified probabilistic model of tsunami height proposed in this study. We believe this inundation probability of various levels to give valuable information for the amendment of current building code of coastal disaster prevention system to tame tsunami attack.

• Journal of Electrical Engineering and Technology
• /
• v.13 no.6
• /
• pp.2392-2401
• /
• 2018

Creepage discharge faults in air on solid insulating material play a vital role in degradation and ageing of material which ultimately leads to breakdown of power equipment. And electric discharge decompose air in to its by-products such as Ozone and $NO_x$ gases. By analyzing air decomposition gases is a potential method for fault diagnostic in air. In this paper, experimental research has been conducted to study the effect of creepage discharge on rate of generation of air decomposition by-products using different insulating materials such as RTV, epoxy and fiberglass laminated sheet. Moreover XRF analysis has been done to analyze creepage discharge effect on these insulating materials. All experiments have been done in an open air test cell under constant temperature and pressure conditions. While analysis has been made for low and high humidity conditions. The results show that the overall concentration of air decomposition by-products under creepage discharge in low humidity is 4% higher than concentration measured in high humidity. Based on this study a mathematical relationship is also proposed for the rate of generation of air decomposition by-products under creepage discharge fault. This study leads to indirect way for diagnostic of creepage discharge propagation in air.

• Journal of Korean Society of Coastal and Ocean Engineers
• /
• v.31 no.6
• /
• pp.368-378
• /
• 2019

Logic trees for probabilistic tsunami hazard assessment include numerous variables to take various uncertainty on earthquake generation into consideration. Results from the hazard assessment vary in different way as more variables are considered in the logic tree. This study is conducted to estimate the effects of various scaling laws and fault parameters on tsunami hazard at the nearshore of Busan. Active fault parameters, such as strike angle, dip angle and asperity, are adjusted in the modelling of tsunami propagation, and the numerical results are used in the sensitivity analysis. The influence of strike angle to tsunami hazard is not as much significant as it is expected, instead, dip angle and asperity show a considerable impact to tsunami hazard assessment. It is shown that the dip angle and the asperity which determine the initial wave form are more important than the strike angle for the assessment of tsunami hazard in the East Sea.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.36 no.4
• /
• pp.373-378
• /
• 2012

The objective of this research is to examine the probabilistic approach to evaluating turbine ejection frequency considering common-cause failure. This paper identifies basic turbine ejection mechanisms under high and low speeds and presents a detailed probabilistic methodology (fault tree) for assessing ejection frequency. The alpha factor methodology is applied to common-cause failure evaluations. The frequencies under different test schemes are compared and the propagation of uncertainty through the fault tree model is evaluated. The following conclusions were reached: (1) the turbine blade ejection frequency due to ductile failure under high speed is around 8.005E-7/yr; (2) if common-cause failure is considered, the frequency will be increased by 11% and 33% depending on the test scheme; and (3) if the parameter uncertainties are considered, the frequency is estimated to be in the range of 9.35E-7 to 1.13E 6, with 90% confidence.

• Journal of the Korean Institute of Electrical and Electronic Material Engineers
• /
• v.20 no.1
• /
• pp.65-73
• /
• 2007

Pattern classification of single and multiple discharge sources was applied using a wavelet image signal method in which a feature extraction was applied using a hidden sub-image. A feature extracting method that used vertical and horizontal images using an MSD method was applied to an averaging process for the scale of pulses for the phase. A feature extracting process for the preprocessing of the input of a neural network was performed using an inverse transformation of the horizontal, vertical, and diagonal sub-images. A back propagation algorithm in a neural network was used to classify defective signals. An algorithm for wavelet image processing was developed. In addition, the defective signal was classified using the extracted value that was quantified for the input of a neural network.

• Journal of KSNVE
• /
• v.10 no.2
• /
• pp.331-337
• /
• 2000

The necessity of diagnosis of the rotating machinery which is widely used in the industry is increasing. Many research has been conducted to manipulate field vibration signal data for diagnosing the fault of designated machinery. As the pattern recognition tool of that signal, neural network which use usually back-propagation algorithm was used in the diagnosis of rotating machinery. In this paper, self-organizing feature map(SOFM) which is unsupervised learning algorithm is used in the abnormal defect diagnosis of rotating machinery and then learning vector quantization(LVQ) which is supervised learning algorithm is used to improve the quality of the classifier decision regions.

• Geomechanics and Engineering
• /
• v.16 no.1
• /
• pp.11-19
• /
• 2018

Propagation of the generalized Rayleigh waves in an elastic half-space covered by an elastic layer for different initial stress combinations and imperfect contact conditions is investigated. Three-dimensional linearized theory of elastic waves in initially stressed bodies in plane-strain state is employed, the corresponding dispersion equation is derived and an algorithm is developed for numerical solution to this equation. Numerical results on the influence of the initial stress patterns and on the influence of the contact conditions are presented and discussed. The case where the external forces are "follower forces" is considered as well. These investigations provide some theoretical foundations for the study of the near-surface waves propagating in layered mechanical systems and can be successfully used for estimation of the degree of the bonded defects between layers, fault characteristics and study of the behavior of seismic surface waves propagating under the bottom of the oceans.

• Journal of Information Technology Services
• /
• v.10 no.2
• /
• pp.235-245
• /
• 2011

Early criticality prediction models that determine whether a design entity is fault-prone or not are becoming more and more important as software development projects are getting larger. Effective predictions can reduce the system development cost and improve software quality by identifying trouble-spots at early phases and proper allocation of effort and resources. Many prediction models have been proposed using statistical and machine learning methods. This paper builds a prediction model using Support Vector Machine(SVM) which is one of the most popular modern classification methods and compares its prediction performance with a well-known prediction model, BackPropagation neural network Model(BPM). SVM is known to generalize well even in high dimensional spaces under small training data conditions. In prediction performance evaluation experiments, dimensionality reduction techniques for data set are not used because the dimension of input data is too small. Experimental results show that the prediction performance of SVM model is slightly better than that of BPM and polynomial kernel function achieves better performance than other SVM kernel functions.