• Geophysics and Geophysical Exploration
• /
• v.25 no.3
• /
• pp.140-161
• /
• 2022

Seismic data with missing traces are often obtained regularly or irregularly due to environmental and economic constraints in their acquisition. Accordingly, seismic data interpolation is an essential step in seismic data processing. Recently, research activity on machine learning-based seismic data interpolation has been flourishing. In particular, convolutional neural network (CNN) and generative adversarial network (GAN), which are widely used algorithms for super-resolution problem solving in the image processing field, are also used for seismic data interpolation. In this study, CNN-based algorithm, U-Net and GAN-based algorithm, and conditional Wasserstein GAN (cWGAN) were used as seismic data interpolation methods. The results and performances of the methods were evaluated thoroughly to find an optimal interpolation method, which reconstructs with high accuracy missing seismic data. The work process for model training and performance evaluation was divided into two cases (i.e., Cases I and II). In Case I, we trained the model using only the regularly sampled data with 50% missing traces. We evaluated the model performance by applying the trained model to a total of six different test datasets, which consisted of a combination of regular, irregular, and sampling ratios. In Case II, six different models were generated using the training datasets sampled in the same way as the six test datasets. The models were applied to the same test datasets used in Case I to compare the results. We found that cWGAN showed better prediction performance than U-Net with higher PSNR and SSIM. However, cWGAN generated additional noise to the prediction results; thus, an ensemble technique was performed to remove the noise and improve the accuracy. The cWGAN ensemble model removed successfully the noise and showed improved PSNR and SSIM compared with existing individual models.

• Economic and Environmental Geology
• /
• v.53 no.4
• /
• pp.479-489
• /
• 2020

A fault is a geological structure that can be a migration path or a cap rock of hydrocarbon such as oil and gas, formed from source rock. The fault is one of the main targets of seismic exploration to find reservoirs in which hydrocarbon have accumulated. However, conventional fault detection methods using lateral discontinuity in seismic data such as semblance, coherence, variance, gradient magnitude and fault likelihood, have problem that professional interpreters have to invest lots of time and computational costs. Therefore, many researchers are conducting various studies to save computational costs and time for fault interpretation, and machine learning technologies attracted attention recently. Among various machine learning technologies, many researchers are conducting fault interpretation studies using the support vector machine, multi-layer perceptron, deep neural networks and convolutional neural networks algorithms. Especially, researchers use not only their own convolution networks but also proven networks in image processing to predict fault locations and fault information such as strike and dip. In this paper, by investigating and analyzing these studies, we found that the convolutional neural networks based on the U-Net from image processing is the most effective one for fault detection and interpretation. Further studies can expect better results from fault detection and interpretation using the convolutional neural networks along with transfer learning and data augmentation.

• Journal of Institute of Control, Robotics and Systems
• /
• v.5 no.4
• /
• pp.454-461
• /
• 1999

A concrete floor trowel machine, developed in the U.S in 1990's, consists of only two rotary trowels, and doesn't need any other mechanism for motion such as wheels. When the machine flattens a concrete floor with its rotary trowels, the machine can move in any direction by utilizing the unbalanced friction forces occurring between the rotary wheels and the floor when the trowels are tilted in appropriate directions. In order to automate the trowels machine, this paper proposed the self-propulsive concrete finishing trowel robot which has twin trowels. For the control of the robot, this paper discussed the following. Firstly, the dynamics model of the driving frictional force applied on each trowel from the floor is derived. Secondly, the relationship between the driving force for the robot and the control variable of the robot is derived. Finally, the basic motion of the robot are realized by using the obtained relationship. This paper figures out how the concrete floor finishing robot with tow trowels moves and will contribute to realizing it.

• Proceedings of the Korean Society of Precision Engineering Conference
• /
• 2005.06a
• /
• pp.1311-1314
• /
• 2005

Recently drum washing machines are required to improve not only functions, but also endurance security. The spider is one of the major parts in a drum washing machine as a power transmission device. It is needed estimating for fatigue life because it rotates at high velocity when the drum washing machine works. In this study, we tried to estimate fatigue life of the Spider with CAE and verified the accuracy by comparing the CAE results with the experimental results. The estimation method of fatigue life for the spider with CAE will be applied to raise the efficiency of time and money in the design process of a new drum washing machine.

• The Transactions of the Korean Institute of Electrical Engineers B
• /
• v.54 no.5
• /
• pp.253-259
• /
• 2005

Magnetic encoder with relatively low pulse per rotation is generally used for detecting speed of the high-speed rotating machine. It is due to the fact of the mechanical problems of vibration and bearing stiffness and also the limit of maximum output pulse of the mounted encoder. The magnetic encoder is divided into two types, that is, toothed gear-wheel method and magnetic wheel method according to the shape of the rotation disk. In case of detecting speed by the tooth gear-wheel, the encoder itself can be acted as the additional inertia where the number of tooth determining the output pulse and the width of the wheel detecting the change of the magnetic flux density are relatively enough large considering the volume of the rotating machine. While the magnetic wheel method has the limit of the magnetizing number of the ring magnet, there is relatively few, if nv, the influence of inertia on the machine. In this paper, it is proposed a simple magnetic wheel encoder suited for the high speed rotating machine and the method of signal processing and the output characteristics are examined through the V/F operation of max 48,000(rpm) and 2.4(KW) spindle motor.

• Journal of the Korean Society for Precision Engineering
• /
• v.29 no.5
• /
• pp.500-505
• /
• 2012

Recently, a high speed spindle is an essential part of machine tools to satisfy latest demand of high precision product and machining of hard materials. But, there are many disadvantages such as heat generation of built-in-motor, bearing friction, noise, vibration and displacement because of the high speed. Many researches on spindle systems have been conducted for solving these problems. In this study, technical trend of machine tools spindle systems are analyzed with patent PSM, mapping and grouping. The analysis is carried out for the applied patent during January 2000 and December 2009 in Korea, Japan, EU and U.S.A. And development of the direction, strategy and promising technologies of the spindle system are suggested.

• Journal of Biosystems Engineering
• /
• v.36 no.5
• /
• pp.361-368
• /
• 2011

The amount of export of paprika has been increased rapidly in recent years. Therefore, its cultivation area has greatly increased in Korea according to current consumer's attraction. Moreover, it becomes one of the major exporting products while it recorded 53 million dollars, in 2009 resulting in 40% of the total vegetables export. Most of the products are exported to Japan, but it is necessary to prolong the quality preservation periods to export paprika to nations like U.S.A. or EU. However, to encourage an export to many countries, washing and disinfection became more important to deal with longer transportation and medical inspection. The non-chemical use is very important due to stronger regulation of safety to agricultural production. Accordingly, this study was performed to determine the optimum conditions and develop a prototype washing machine, hot water washing of paprika. The results were as follows : The working performance of the prototype was 938 kg/hr, and which was 1.5 times higher than the conventional air gun type washing machine. The operation cost of prototype was 30 won/kg, and 56% of the cost was reduced when compared with air gun type washing machine.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.15 no.1
• /
• pp.241-248
• /
• 2011

In this paper we proposed and implemented a diagnosis system to control nuclear fuel handling machine. The proposed system consists of data acquisition system, diagnosis algorithm and faults simulator. Since the test on real operation of the fuel handling machine is impossible, we evaluated the proposed system by diagnosis experiments using the faults simulator, with which test signals on abnormal states of the bearing ball and the inner race of the bearing are generated. The experiments showed that resulting diagnosis analysis are consistent with the theoretical expectations.

• Journal of the Korean Society for Precision Engineering
• /
• v.18 no.11
• /
• pp.74-79
• /
• 2001

In this two-part paper, a virtual machine tool (VMT) is presented. In part 1, the analytical foundation of a virtual machining system, envisioned as the foundation for a comprehensive simulation environment capable of predicting the outcome of cutting processes, is developed. The VMT system purposes to experience the pseudo-real machining before real cutting with a CNC machine tool, to provide the proper cutting conditions for process planners, and to compensate or control the machining process in terms of the productivity and attributes of products. The attributes can be characterized with the machined surface error, dimensional accuracy, roughness, integrity and so forth. The main components of the VMT are cutting process, application, thermal behavior and feed drive modules. In part 1, the cutting process module is presented. The proposed models were verified experimentally and gave significantly better prediction results than any other method. The thermal behavior and feed drive modules are developed in part 2 paper. The developed models are integrated as a comprehensive software environment in part 2 paper.

• IEIE Transactions on Smart Processing and Computing
• /
• v.2 no.3
• /
• pp.117-129
• /
• 2013

The increasing demand for monitoring wildfires and their impact on the land surface have prompted studies of burned area extraction and analysis. To differentiate burned and unburned area, the earlier method of the Moderate Resolution Imaging Spectro-radiometer (MODIS) Burned Area Detection Algorithm was proposed to estimate the change in land surface based on the reflectance energy. The energy, whose wavelengths are sensitive to burning, was selected to calculate the change parameter $Z_{score}$. This method was applied using the MODIS images to produce a MODIS Burned Area product. The approach was to simplify this algorithm to make it compatible with the Landsat 7 ETM+ SLC-off images. To extract the refined version of burned regions, post-processing was carried out by applying a median filter, dilation morphology algorithm, and finally a gap filling method. The experimental results showed that the detailed burned areas extracted from the proposed method exhibited more spatial details than those of the MODIS Burned products in the large U.S areas. The results also revealed the discontinuous distribution of burned regions in Vietnam forests.