• Design & Manufacturing
• /
• v.15 no.1
• /
• pp.48-56
• /
• 2021

The quality of products produced by injection molding process is greatly influenced by the process variables set on the injection molding machine during manufacturing. It is very difficult to predict the quality of injection molded product considering the stochastic nature of manufacturing process, because the process variables complexly affect the quality of the injection molded product. In the present study we predicted the quality of injection molded product using Artificial Neural Network (ANN) method specifically from Multiple Input Single Output (MISO) and Multiple Input Multiple Output (MIMO) perspectives. In order to train the ANN model a systematic plan was prepared based on a combination of orthogonal sampling and random sampling methods to represent various and robust patterns with small number of experiments. According to the plan the injection molding experiments were conducted to generate data that was separated into training, validation and test data groups to optimize the parameters of the ANN model and evaluate predicting performance of 4 structures (MISO1-2, MIMO1-2). Based on the predicting performance test, it was confirmed that as the number of output variables were decreased, the predicting performance was improved. The results indicated that it is effective to use single output model when we need to predict the quality of injection molded product with high accuracy.

• Imaging Science in Dentistry
• /
• v.52 no.2
• /
• pp.219-224
• /
• 2022

Purpose: This study aimed to evaluate the performance of transfer learning in a deep convolutional neural network for classifying implant fixtures. Materials and Methods: Periapical radiographs of implant fixtures obtained using the Superline (Dentium Co. Ltd., Seoul, Korea), TS III(Osstem Implant Co. Ltd., Seoul, Korea), and Bone Level Implant(Institut Straumann AG, Basel, Switzerland) systems were selected from patients who underwent dental implant treatment. All 355 implant fixtures comprised the total dataset and were annotated with the name of the system. The total dataset was split into a training dataset and a test dataset at a ratio of 8 to 2, respectively. YOLOv3 (You Only Look Once version 3, available at https://pjreddie.com/darknet/yolo/), a deep convolutional neural network that has been pretrained with a large image dataset of objects, was used to train the model to classify fixtures in periapical images, in a process called transfer learning. This network was trained with the training dataset for 100, 200, and 300 epochs. Using the test dataset, the performance of the network was evaluated in terms of sensitivity, specificity, and accuracy. Results: When YOLOv3 was trained for 200 epochs, the sensitivity, specificity, accuracy, and confidence score were the highest for all systems, with overall results of 94.4%, 97.9%, 96.7%, and 0.75, respectively. The network showed the best performance in classifying Bone Level Implant fixtures, with 100.0% sensitivity, specificity, and accuracy. Conclusion: Through transfer learning, high performance could be achieved with YOLOv3, even using a small amount of data.

• Journal of Advanced Navigation Technology
• /
• v.28 no.3
• /
• pp.362-368
• /
• 2024

In this paper, we investigated the performance of directional and omnidirectional precoding schemes when transmitting to improve downlink performance in massive MIMO. Omnidirectional precoding was used to broadcast a common signal, such as a synchronization or control signal, to all users. The main purpose of omnidirectional precoding is to design the precoding matrix so that the signal transmitted in the downlink is the same in all directions and emitted with maximum energy. We propose a flexible omnidirectional precoding method for full-dimensional massive MIMO that can set the spatial coverage range to less than 120 degrees. The constraints of omnidirectionality of all antennas, equal transmit power, and maximum transmit rate are used to design the encoding matrix of the proposed method. The performance was evaluated in terms of spatial coverage by considering changing the spatial coverage of the antenna array by changing the distance between neighboring antennas in the antenna array.

• Journal of the Korean Institute of Illuminating and Electrical Installation Engineers
• /
• v.24 no.1
• /
• pp.152-159
• /
• 2010

High power density and power conversion efficiency have been required in the power converters according to the rapid growth of industry. In this context, the next generation High Speed Train(HST) requires power converter which has high-efficiency, high-performance and high-density. In this paper, the new control technique for battery charger used for the next generation HST is proposed. The phase shift ZVZCS converter is classified according to a resonant circuit which is located in the primary or secondary side. In this paper, The PWM switching technique using all-pass filter is proposed to control ZVZCS converter which has resonant circuit in the secondary side. ATmega_128 micro controller based in all-pass filter in substitute for phase shift IC is presented to have digital control. To verify the proposed topology, the simulation and experiment are performed by using PSIM software and 1[kW] experimental set-up.

• Journal of the Korean Society for Railway
• /
• v.16 no.5
• /
• pp.402-409
• /
• 2013

In metropolitan railway systems, the average commuting time keeps increasing as the scheduled speed increases, and this leads to a decline of rail service usage and competitiveness. Therefore, effective express operation for urban trains is required to improve the scheduled speed. In this study, based on the obtainable time shortening efficiency and economic viability, several express operations are suggested for urban railways and these suggestions are compared by considering high performance trains with acceleration/deceleration and maximum speed improvement. As a result, the optimum express system, which can minimize the cost for sidetrack construction, is suggested.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.18 no.1
• /
• pp.609-617
• /
• 2017

Railways are one of the safest and most important transportation systems in the world. On the other hand, due to the increasing complexity of the railway system and the running distance of rail vehicles, railway accidents occur continuously every year. In particular, in the case of high-speed trains and freight trains, if the function of the axle bearing is lost due to abnormal overheating of the axle box bearing, the load on the axle becomes uneven. Therefore, abnormal overheating in the train axle box bearings can cause serious accidents or derailments. For this purpose, a Hot Box Detector (HBD) was installed in the track side of a high speed line to detect abnormal overheating. This paper proposes an EWMA technique-based axle temperature monitoring method to detect abnormal overheating quickly and efficiently. A statistical design of the proposed method was also performed. The proposed method has better performance compared to the current method in the case of abnormal overheating and the performance is improved by approximately 170% at the maximum.

• Convergence Security Journal
• /
• v.23 no.1
• /
• pp.117-123
• /
• 2023

To detect advanced ransomware attacks with machine learning-based models, the classification model must train learning data with high-dimensional feature space. And in this case, a 'curse of dimension' phenomenon is likely to occur. Therefore, dimensionality reduction of features must be preceded in order to increase the accuracy of the learning model and improve the execution speed while avoiding the 'curse of dimension' phenomenon. In this paper, we conducted classification of ransomware by applying three machine learning models and two feature extraction techniques to two datasets with extremely different dimensions of feature space. As a result of the experiment, the feature dimensionality reduction techniques did not significantly affect the performance improvement in binary classification, and it was the same even when the dimension of featurespace was small in multi-class clasification. However, when the dataset had high-dimensional feature space, LDA(Linear Discriminant Analysis) showed quite excellent performance.

• Journal of The Korean Astronomical Society
• /
• v.56 no.2
• /
• pp.213-224
• /
• 2023

Type II solar radio bursts show frequency drifts from high to low over time. They have been known as a signature of coronal shock associated with Coronal Mass Ejections (CMEs) and/or flares, which cause an abrupt change in the space environment near the Earth (space weather). Therefore, early detection of type II bursts is important for forecasting of space weather. In this study, we develop a deep-learning (DL) model for the automatic detection of type II bursts. For this purpose, we adopted a 1-D Convolution Neutral Network (CNN) as it is well-suited for processing spatiotemporal information within the applied data set. We utilized a total of 286 radio burst spectrum images obtained by Hiraiso Radio Spectrograph (HiRAS) from 1991 and 2012, along with 231 spectrum images without the bursts from 2009 to 2015, to recognizes type II bursts. The burst types were labeled manually according to their spectra features in an answer table. Subsequently, we applied the 1-D CNN technique to the spectrum images using two filter windows with different size along time axis. To develop the DL model, we randomly selected 412 spectrum images (80%) for training and validation. The train history shows that both train and validation losses drop rapidly, while train and validation accuracies increased within approximately 100 epoches. For evaluation of the model's performance, we used 105 test images (20%) and employed a contingence table. It is found that false alarm ratio (FAR) and critical success index (CSI) were 0.14 and 0.83, respectively. Furthermore, we confirmed above result by adopting five-fold cross-validation method, in which we re-sampled five groups randomly. The estimated mean FAR and CSI of the five groups were 0.05 and 0.87, respectively. For experimental purposes, we applied our proposed model to 85 HiRAS type II radio bursts listed in the NGDC catalogue from 2009 to 2016 and 184 quiet (no bursts) spectrum images before and after the type II bursts. As a result, our model successfully detected 79 events (93%) of type II events. This results demonstrates, for the first time, that the 1-D CNN algorithm is useful for detecting type II bursts.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.40 no.9
• /
• pp.1699-1709
• /
• 2015

Moving terminals like high-speed-train undergo high Doppler frequency shift, and this leads to carrier frequency offsets that have to be compensated to avoid degradation of communication performance. In multiple access mechanism like OFDMA, base-stations need complex hardware to compensate the uplink frequency offset. In this paper, we propose a method, which can reduce burden of the base-station and makes frequency offset estimation and compensation simple. This method contains transmitting new synchronization signal, estimating frequency offsets in base-station, transmitting feedback information to terminal, and compensating the offset in uplink transmission. Simulation results show the proposed method operates well in high Doppler frequency shift conditions of 500 km/h which is the requirements of 5G mobile communication.

• Journal of Advanced Marine Engineering and Technology
• /
• v.38 no.3
• /
• pp.262-268
• /
• 2014

This research introduces the structural design and the validation results of the flexible high speed coupling for 2MW class wind turbine which transmit and cut off torque between gear box and generator. The high speed coupling requires electrical insulation to prevent electrical surface damages on gear box. Therefore glass fiber reinforced plastics is applied to absorb the vibration and deformation of power train and to transmit required torque. Finite element analysis was performed to optimize the thickness and accumulation number of glass fiber reinforced plastics. Torque limiter which cut off the abnormal torque is designed in frictional disc type. The design of the coupling was validated with the performance test of prototype.