• Proceedings of the KIEE Conference
• /
• 2004.11b
• /
• pp.17-20
• /
• 2004

This paper presents a modeling and simulation of a grid-connected wind turbine generation system with respect to wind variations and three-phase fault in the system. It describes the modeling of the wind turbine system including the drive train model, induction generator model, and grid-interface model on MATLAB/Simulink. Case studies demonstrate that the pitch angle control is carried out to achieve maximum power extraction for wind speed variations and the duration of a fault on the system influences on the output of the wind turbine generator.

• Transactions of Materials Processing
• /
• v.23 no.3
• /
• pp.139-144
• /
• 2014

In the current study, we present a new model for predicting width spread of a slab with a dog-bone shaped cross section during rolling in the roughing train of a hot strip mill. The approach is based on the extremum principle for a rigid plastic material and a three dimensional admissible velocity field. The upper bound theorem is used for calculating the width spread of the slab. The prediction accuracy of the proposed model is examined through comparison with the predictions from 3-D finite element (FE) process simulations.

• Annual Conference of KIPS
• /
• 2019.05a
• /
• pp.455-457
• /
• 2019

In this paper we study the effect of decomposed spatiotemporal convolutions for action recognition in videos. Our motivation emerges from the empirical observation that spatial convolution applied on solo frames of the video provide good performance in action recognition. In this research we empirically show the accuracy of factorized convolution on individual frames of video for action classification. We take 3D ResNet-18 as base line model for our experiment, factorize its 3D convolution to 2D (Spatial) and 1D (Temporal) convolution. We train the model from scratch using Kinetics video dataset. We then fine-tune the model on UCF-101 dataset and evaluate the performance. Our results show good accuracy similar to that of the state of the art algorithms on Kinetics and UCF-101 datasets.

• Proceedings of the KSR Conference
• /
• 2008.11b
• /
• pp.1602-1606
• /
• 2008

In the Rail Safety Regulations article 16, deceleration rate in the survival spaces should be limited as far as is practicable to 5g, and shall not be more than 7.5g. As it is impractical to evaluate complete train behaviour by testing, the achievement of the objectives shall be validated by dynamic simulations corresponding to the reference collisions scenarios. But initial design and evaluation procedure, impact dynamics model which classified 1D and 2D is more useful than full scale model. This paper presents acceleration response characteristics between 1D and 2D dynamics model under head-on collision in standard collision scenarios.

• Nuclear Engineering and Technology
• /
• v.53 no.12
• /
• pp.4060-4066
• /
• 2021

Flow-accelerated corrosion (FAC) of carbon steel piping is a significant problem in nuclear power plants. The basic process of FAC is currently understood relatively well; however, the accuracy of prediction models of the wall-thinning rate under an FAC environment is not reliable. Herein, we propose a methodology to construct pipe wall-thinning rate prediction models using artificial neural networks and a convolutional neural network, which is confined to a straight pipe without geometric changes. Furthermore, a methodology to generate training data is proposed to efficiently train the neural network for the development of a machine learning-based FAC prediction model. Consequently, it is concluded that machine learning can be used to construct pipe wall thinning rate prediction models and optimize the number of training datasets for training the machine learning algorithm. The proposed methodology can be applied to efficiently generate a large dataset from an FAC test to develop a wall thinning rate prediction model for a real situation.

• Journal of the Korean Society of Industry Convergence
• /
• v.22 no.1
• /
• pp.29-37
• /
• 2019

A new permanent magnet biased hybrid maglev actuator is developed. Compared to the classical hybrid maglev actuators, the new maglev has unique flux paths such that bias fluxes are separated with control flux paths. The control flux paths have minimum reluctances only developed by air gaps, so the currents to produce control fluxes can be minimized. The consumed power to operate this maglev system can also be minimized. The gravity load can be compensated with the static magnetic forces developed by the permanent magnet bias fluxes while external disturbances are controlled with the bidirectional AC magnetic forces developed by control fluxes by currents. 1-D circuit model is developed for this model such that the flux densities and magnetic forces are extensively analyzed. 3-D finite element model is also developed to analyze the performances of the maglev actuator.

• Journal of Korea Multimedia Society
• /
• v.25 no.10
• /
• pp.1524-1534
• /
• 2022

This study proposes the classification model of Minhwa genre based on object detection of deep learning. To detect unique Korean traditional objects in Minhwa, we construct custom datasets by labeling images using object keywords in Minhwa DB. We train YOLOv5 models with custom datasets, and classify images using predicted object labels result, the output of model training. The algorithm consists of two classification steps: 1) according to the painting technique and 2) genre of Minhwa. Through classifying paintings using this algorithm on the Internet, it is expected that the correct information of Minhwa can be built and provided to users forward.

• ETRI Journal
• /
• v.45 no.5
• /
• pp.862-873
• /
• 2023

Determining whether an autonomous self-driving agent is in the middle of an intersection can be extremely difficult when relying on visual input taken from a single camera. In such a problem setting, a wider range of views is essential, which drives us to use three cameras positioned in the front, left, and right of an agent for better intersection recognition. However, collecting adequate training data with three cameras poses several practical difficulties; hence, we propose using data collected from one camera to train a three-camera model, which would enable us to more easily compile a variety of training data to endow our model with improved generalizability. In this work, we provide three separate fusion methods (feature, early, and late) of combining the information from three cameras. Extensive pedestrian-view intersection classification experiments show that our feature fusion model provides an area under the curve and F1-score of 82.00 and 46.48, respectively, which considerably outperforms contemporary three- and one-camera models.

• Journal of information and communication convergence engineering
• /
• v.22 no.2
• /
• pp.133-138
• /
• 2024

Dissolved oxygen (DO) is an important factor in ecosystems. However, the analysis of DO is frequently rather complicated because of the nonlinear phenomenon of the river system. Therefore, a convenient model-free algorithm for DO variable is required. In this study, a data-driven algorithm for predicting DO was developed by combining XGBoost and an artificial neural network (ANN), called ANN-XGB. To train the model, two years of ecosystem data were collected in Anyang, Seoul using the Troll 9500 model. One advantage of the proposed algorithm is its ability to capture abrupt changes in climate-related features that arise from sudden events. Moreover, our algorithm can provide a feature importance analysis owing to the use of XGBoost. The results obtained using the ANN-XGB algorithm were compared with those obtained using the ANN algorithm in the Results Section. The predictions made by ANN-XGB were mostly in closer agreement with the measured DO values in the river than those made by the ANN.

• Smart Media Journal
• /
• v.13 no.4
• /
• pp.9-15
• /
• 2024

Tree-based algorithms have been the dominant methods used build a prediction model for tabular data. This also includes personal credit data. However, they are limited to compatibility with categorical and numerical data only, and also do not capture information of the relationship between other features. In this work, we proposed an ensemble model using the Transformer architecture that includes text features and harness the self-attention mechanism to tackle the feature relationships limitation. We describe a text formatter module, that converts the original tabular data into sentence data that is fed into FinBERT along with other text features. Furthermore, we employed FT-Transformer that train with the original tabular data. We evaluate this multi-modal approach with two popular tree-based algorithms known as, Random Forest and Extreme Gradient Boosting, XGBoost and TabTransformer. Our proposed method shows superior Default Recall, F1 score and AUC results across two public data sets. Our results are significant for financial institutions to reduce the risk of financial loss regarding defaulters.