• Journal of Convergence for Information Technology
• /
• v.12 no.1
• /
• pp.31-38
• /
• 2022

This paper applies an expert independent unsupervised neural network learning-based multivariate time series data analysis model, MSCRED(Multi-Scale Convolutional Recurrent Encoder-Decoder), and to overcome the limitation, because the MCRED is based on Auto-encoder model, that train data must not to be contaminated, by using learning data sampling technique, called Subset Sampling Validation. By using the vibration data of power plant equipment that has been labeled, the classification performance of MSCRED is evaluated with the Anomaly Score in many cases, 1) the abnormal data is mixed with the training data 2) when the abnormal data is removed from the training data in case 1. Through this, this paper presents an expert-independent anomaly diagnosis framework that is strong against error data, and presents a concise and accurate solution in various fields of multivariate time series data.

• The Journal of the Acoustical Society of Korea
• /
• v.22 no.8
• /
• pp.660-669
• /
• 2003

This paper describes a theory to calculate sound source field from the spatial transform of sound field and the measured cross-power spectrum of sound pressure over a hologram plane close to a sound source, Calculating method is proposed to solve sound pressures from cross-power spectrums over a hologram plane, For this, Taylor series for the nonlinear equations is expanded, and it is calculated using Newton-Raphon method, Also, a wave number filter is used to reduce errors that is occurred on the backward propagation, and is performed numerical simulation of the circular piston sound source with infinite baffle in water to verify the proposed theory.

• The Transactions of the Korean Institute of Electrical Engineers P
• /
• v.64 no.2
• /
• pp.74-78
• /
• 2015

With the improvement of living standards and economic development, electricity consumption continues to grow. The electricity is a special energy which is hard to store, so its supply must be consistent with the demand. The objective of electricity demand forecasting is to make best use of electricity energy and provide balance between supply and demand. Hence, it is very important work to forecast electricity demand with higher precision. So, various forecasting methods have been developed. They can be divided into five broad categories such as time series models, regression based model, artificial intelligence techniques and fuzzy logic method without considering special-day effects. Electricity demand patterns on holidays can be often idiosyncratic and cause significant forecasting errors. Such effects are known as special-day effects and are recognized as an important issue in determining electricity demand data. In this research, we developed the power demand forecasting method using ELM(Extreme Learning Machine) for special day, particularly, lunar new year and Chuseok holiday.

• Current Optics and Photonics
• /
• v.6 no.5
• /
• pp.479-488
• /
• 2022

To account for the internal thermal effects of solid-state lasers, a method using a back propagation (BP) neural network integrated with a particle swarm optimization (PSO) algorithm is developed, which is a new wavefront distortion correction technique. In particular, by using a slab laser model, a series of fiber pumped sources are employed to form a controlled array to pump the gain medium, allowing the internal temperature field of the gain medium to be designed by altering the power of each pump source. Furthermore, the BP artificial neural network is employed to construct a nonlinear mapping relationship between the power matrix of the pump array and the thermally induced wavefront aberration. Lastly, the suppression of thermally induced wavefront distortion can be achieved by changing the power matrix of the pump array and obtaining the optimal pump light intensity distribution combined using the PSO algorithm. The minimal beam quality β can be obtained by optimally distributing the pumping light. Compared with the method of designing uniform pumping light into the gain medium, the theoretically computed single pass beam quality β value is optimized from 5.34 to 1.28. In this numerical analysis, experiments are conducted to validate the relationship between the thermally generated wavefront and certain pumping light distributions.

• 제어로봇시스템학회:학술대회논문집
• /
• 2003.10a
• /
• pp.2264-2269
• /
• 2003

This paper presents an improved method that constructs an equivalent T-S fuzzy model for nonlinear systems expressed by nonlinear differential equations including terms of power series. The method in this paper has fewer numbers of the rules than the previous methods as well as exactly expresses nonlinear systems. Moreover, this method can get wider feasible area satisfying the stability conditions than the previous methods. We show the improvement of modeling by comparing the proposed method with two previous methods through an inverted pendulum on a cart.

• Proceedings of the KIEE Conference
• /
• 2011.07a
• /
• pp.1304-1305
• /
• 2011

In this study, tabbing cell samples in the soft touch method are compared samples according to soldering voltage conditions. Series resistance and power loss is measured by two measurement method. As a result, probe measurement method was found optimized soldering voltage optimized according to voltage condition and clamp measurement method was not difference. The purpose of this paper is to analysis the principle of measurement method.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.60 no.1
• /
• pp.8-13
• /
• 2011

A percentage current differential relay is widely used for transformer protection. Because many percentage current differential relays recently use modified methods instead of conventional methods for deciding the operating characteristics of the large current region, in this paper, the operating region of a percentage current differential relay is analyzed in input-output current domain instead of operating-restraint current domain. An effective method to set the operating region when a CT is saturated is proposed with a series of investigations comparing a conventional method with the proposed modified method. The performance of the proposed method is evaluated for internal and external faults of a power transformer having the voltage rating of 345/154kV. EMTP-RV is used for the relaying data collection.

• Progress in Superconductivity and Cryogenics
• /
• v.9 no.4
• /
• pp.32-36
• /
• 2007

We fabricated and tested a novel hybrid superconducting fault current limiter (SFCL) of three-phase $24kV_{rms}/630A_{rms}$ rating. In order to apply conventional resistive SFCLs into electric power systems, the urgent issues to be settled are as follows, such as initial installation price of SFCL, operation and maintenance cost due to ac loss of superconductor and the life of cryostat, and high voltage and high current problems. The ac loss and high cost of superconductor and cryostat system are main bottlenecks for real application. Furthermore in order to increase voltage and current ratings of SFCL, a lot of superconductor components should be connected in series and parallel which resulted in extreme high cost. In addition, the method to quench all components at the same instant needs very sophisticated skill and careful operation. Due to these problems, the practical applications of SFCL were pending. Therefore, in order to make practical SFCL, the price of SFCL should be lowered and should meet the demand of utilities. We designed novel hybrid SFCL which combines superconductor and conventional electric equipment including vacuum interrupter, power fuse and current limiting reactor. The main purpose of hybrid SFCL is to drastically reduce total usage of superconductor by adopting current commutation method by use of superconductor and high fast switch. Consequently, it was possible to get the satisfactory test results using this method, and further works for field tests are in the process.

• Nuclear Engineering and Technology
• /
• v.56 no.9
• /
• pp.3571-3584
• /
• 2024

Reactor core transient calculation is very important for the reactor safety analysis, in which the kernel is neutron kinetics calculation by simulating the variation of neutron density or thermal power over time. Compared with the point kinetics method, the time-space neutron kinetics calculation can provide accurate variation of neutron density in both space and time domain. But it consumes a lot of resources. It is necessary to develop a surrogate model that can quickly obtain the temporal and spatial variation information of neutron density or power with acceptable calculation accuracy. This paper uses the time-varying characteristics of power to construct a time function, parameterizes the time-varying characteristics which contains the information about the spatial change of power. Thereby, the amount of targets to predict in the space domain is compressed. A surrogate method using the machine learning is proposed in this paper. In the construction of a neural network, the input is processed by a convolutional layer, followed by a fully connected layer or a deconvolution layer. For the problem of time sequence disturbance, a structure combining convolutional neural network and recurrent neural network is used. It is verified in the tests of a series of 1D, 2D and 3D reactor models. The predicted values obtained using the constructed neural network models in these tests are in good agreement with the reference values, showing the powerful potential of the surrogate models.

• Journal of Advanced Marine Engineering and Technology
• /
• v.34 no.2
• /
• pp.332-338
• /
• 2010

Annually, electrical fires caused by arcing phenomena in power system rapidly increase as the use of more electric appliances, but there is no established method for the prevention of the accidents. With this background, this paper dealt with the experimental results on a series arc detection technique and a device for air conditioners. Series arcing phenomena that is generated in incomplete connection of air conditioners was simulated, and the frequency spectrum was analyzed. The Fast Fourier Transform (FFT) of the arc pulse showed that the dominant frequency components exist in ranges of 190 kHz~250 kHz and 900 kHz~1.6 MHz. An arc detection circuit with low cut off frequency of 170 kHz to attenuate 60 Hz by 170 dB and a signal discriminator were designed. Also, an algorithm which separate series arc signal from unwanted noises produced by switching operation, inverter, and surge was proposed. Application experiment was carried out on several types of air-conditioners by using the arc generator specified in UL1699, and the results showed the over 99 % accuracy.