• Journal of Magnetics
• /
• v.22 no.1
• /
• pp.34-42
• /
• 2017

Material defects in the Steam Generator Tubes (SGT) of sodium cooled fast breeder reactor (PFBR) can lead to leakage of water into sodium. The water and sodium reaction will lead to major accidents. Therefore, the examination of steam generator tubes for the early detection of defects is an important requirement for safety and economic considerations. In this work, the Magnetic Flux Leakage (MFL) based Non Destructive Testing (NDT) technique is used to perform the defect detection process. The rectangular notch defects on the outer surface of steam generator tubes are modeled using COMSOL multiphysics 4.3a software. The obtained MFL images are de-noised to improve the integrity of flaw related information. Grey Level Co-occurrence Matrix (GLCM) features are extracted from MFL images and taken as input parameter to train the neural network. A comparative study on characterization have been carried out using feed-forward back propagation (FFBP) and cascade-forward back propagation (CFBP) algorithms. The results of both algorithms are evaluated with Mean Square Error (MSE) as a prediction performance measure. The average percentage error for length, depth and width are also computed. The result shows that the feed-forward back propagation network model performs better in characterizing the defects.

• Journal of the Korean Geotechnical Society
• /
• v.18 no.1
• /
• pp.71-78
• /
• 2002

This paper presents a back-calculation technique leer the prediction of the behavior of earth wall inserted in multi-layered soil deposit. The soil properties are back-calculated from the measured displacement at each construction stage and the behavior of earth wall far the next construction stage is predicted using back-calculated soil properties. For multi-layered soil deposit, the back-calculation would be very difficult due to the increase in the number of variables. In this study, to solve this difficulty, the back-calculation was performed successively from the lowest layer to the upper layers. An efficient elasto-plastic beam-column analysis was used for forward analysis to minimize the computation time of iterative back-calculation procedure. The coefficients of subgrade reaction and lateral earth pressure necessary for the formation of p-y curve were selected as back calculation variables, and to minimize the effect of abnormal behavior of the wall which might be caused by any unexpected action during construction, the difference between measured displacement increment and computed displacement increment at each construction stages is used as the objective function of optimization. The constrained sequential linear programming was used for the optimization technique to found values of variables minimizing the objective function. The proposed method in this study was verified using numerically generated data and measured field data.

• Journal of Intelligence and Information Systems
• /
• v.16 no.3
• /
• pp.1-14
• /
• 2010

The back-propagation neural network (BPN) has long been successfully applied in bankruptcy prediction problems. Despite its wide application, some major issues must be considered before its use, such as the network topology, learning parameters and normalization methods for the input and output vectors. Previous studies on bankruptcy prediction with BPN have shown that many researchers are interested in how to optimize the network topology and learning parameters to improve the prediction performance. In many cases, however, the benefits of data normalization are often overlooked. In this study, a genetic algorithm (GA)-based normalization transform, which is defined as a linearly weighted combination of several different normalization transforms, will be proposed. GA is used to extract the optimal weight for the generalization. From the results of an experiment, the proposed method was evaluated and compared with other methods to demonstrate the advantage of the proposed method.

• 제어로봇시스템학회:학술대회논문집
• /
• 1999.10a
• /
• pp.226-229
• /
• 1999

We have been developing a method to build models for time series using Genetic Programming. The proposed method has been applied to various kinds of time series e.g. computer-generated chaos, natural phenomena, and financial market indices etc. Now we apply the prediction method to time histories of seismic ground motion i.e. one-step-ahead prediction of seismographic amplitude. Waves of earthquakes are composed of P-waves and S-waves. They propagate in different speeds and have different characteristics. It is believed that P-waves arrive firstly and S-waves arrive secondly. Simulations were performed based on real data of Hyuganada earthquake which broke out at southern part of Kyushuu Island in Japan. To our surprise, prediction model built using the earthquake waves in early time can enough precisely predict main huge waves in later time. Lots of experiments lead us to conclude that every slice of data involves P-wave and S-wave. The simulation results suggest the GP-based prediction method can be utilized in alarm systems or dispatch systems in an emergency.

• The Journal of the Acoustical Society of Korea
• /
• v.14 no.6
• /
• pp.48-53
• /
• 1995

A considerable number of methods are available for predicting traffic noise levels of road networks where sound is freely propagating. But surrounding buildings reflect back sound to the road and sound energy is increased by these reflectors. Therefore, this study was focussed on the establishment of the prediction method of road traffic noise under reflective conditions. This prediction method was developed by establishing prediction formulas of noise level such as $L_{10},\;L_{50},\;L_{90}\;and\;L_{eq}$. The sound energy density was employed to establish prediction formulas in terms of independent variables. The validity of the proposed prediction formulas was been confirmed by applying them to actually measured parameters of road traffic noise and noise level data. On the whole, the agreement between measured and predicted noise levels appeared to be satisfactory. The conclusion might be reached that the method developed in this study could be used to predict road traffic noise under reflective conditions.

• Transactions of Materials Processing
• /
• v.3 no.1
• /
• pp.84-96
• /
• 1994

Simulation of 3 dimensional large irregularly shaped stamping process by a dynamic approach, based on an explicit time integration scheme, has been shown to be highly efficient and robust in comparison to traditional, implicit, quasi-static ones. The objective of the work is to evaluate the results from explicit code in application to deep drawing of rectangular cup and stamping of automotive front fender, in which deformation, force, thickness distribution are calculated. The method of predicting spring back and formability by and explicit code are suggested and applied to the processes.

• Proceedings of the KIEE Conference
• /
• 1994.07a
• /
• pp.475-478
• /
• 1994

In this paper two methods of torque control for brushless DC motor with non-ideal trapezoidal back EMF are presented. One is the method of modulating the reference current so as to give a constant torque since the torque is given by the back EMF and the. phase currents. And the current control loop includes the feedforward control of back EMF and of the neutral voltage between the neutral points of the inverter and of the machine. The other is a direct voltage calculation algorithm for a given reference torque. In the two methods, the time delay due to the calculation is compensated by one sampling current prediction. The simulation results are presented to verify the proposed methods.

• Proceedings of the KIEE Conference
• /
• 1997.11a
• /
• pp.32-34
• /
• 1997

For a given brushless DC linear motor, we suggest the numerical prediction methods to analyze it's thrust characteristics. First, we calculate the magnetic flux density by the finite element method, and we then compute the maximum thrust with three computational methods - a Lorentz equation, a Maxwell stress method and a virtual work method. To confirm the accuracy of the computational methods, we measure the thrust of the linear motor made by our laboratory with a force-torque sensor. Also, we calculate the thrust by the measured back electromotive force. To choose the appropriate method for a specified application, we compare the maximum thrusts of the computational method and the calculation by the back electromotive force with the measured one. We conclude that the Maxwell stress method is turned out the best because it has the most accurate results among three computational methods and it is more convenient than the calculation method by the back electromotive force.

• Proceedings of the KIEE Conference
• /
• 2008.07a
• /
• pp.752-753
• /
• 2008

The analytical expressions for magnetic field distributions considering slotting effects, cogging torque and back-emf considering skew effects are established. On the basis of magnetic field solutions, electrical parameters such as back-emf constant and winding inductance are obtained. The predicted results are validated extensively by non-linear finite element (FE) analyses. In particular, test results such as back-emf, cogging torque, inductance and resistance measurements are given to confirm the analyses. Finally, generating performances are investigated by applying estimated parameters to equivalent circuit (EC) of the permanent magnet generator (PMG) and validated extensively by FE calculations and measurements.

• Proceedings of the KSME Conference
• /
• 2001.06a
• /
• pp.740-745
• /
• 2001

The plastic deformation behavior of formed CICC fur the superconducting Tokamac fusion device was examined and appropriate manufacturing information was provided. A relation between travel of the bending roller and spring back displacement was obtained via virtual manufacturing. The radius of CICC after forming was expressed as a function of the bend-roll travel. The maximum von Mises stress after spring back was also monitored fur the SAGBO prediction. Next, the variation of the CICC cross-sectional area was examined during the first turn and during conduit bending with the largest curvature. Finally, the coil radius was measured and compared with the data generated from the virtual manufacturing. The measured data showed similar pattern as predicted one. Using the mapping function found to match with the real data, the data from the virtual manufacturing may facilitate accurate manufacturing.