• Transactions of the Korean Society of Automotive Engineers
• /
• v.20 no.1
• /
• pp.95-105
• /
• 2012

The selective catalytic reduction (SCR) system is a highly-effective aftertreatment device for NOx reduction of diesel engines. Generally, the ammonia ($NH_3$) was generated from reaction mechanism of SCR in the SCR system using the liquid urea as the reluctant. Therefore, the precise urea dosing control is a very important key for NOx and $NH_3$ slip reduction in the SCR system. This paper investigated NOx and $NH_3$ emission characteristics of urea-SCR dosing system based on model-based control algorithm in order to reduce NOx. In the map-based control algorithm, target amount of urea solution was determined by mass flow rate of exhaust gas obtained from engine rpm, torque and $O_2$ for feed-back control NOx concentration should be measured by NOx sensor. Moreover, this algorithm can not estimate $NH_3$ absorbed on the catalyst. Hence, the urea injection can be too rich or too lean. In this study, the model-based control algorithm was developed and evaluated on the numerical model describing physical and chemical phenomena in SCR system. One channel thermo-fluid model coupled with finely tuned chemical reaction model was applied to this control algorithm. The vehicle test was carried out by using map-based and model-based control algorithms in the NEDC mode in order to evaluate the performance of the model based control algorithm.

• Smart Structures and Systems
• /
• v.26 no.2
• /
• pp.175-184
• /
• 2020

Conventional Monte Carlo simulation-based methods for seismic risk assessment of water networks often require excessive computational time costs due to the hydraulic analysis. In this study, an Artificial Neural Network-based surrogate model was proposed to efficiently evaluate the flow-based system reliability of water distribution networks. The surrogate model was constructed with appropriate training parameters through trial-and-error procedures. Furthermore, a deep neural network with hidden layers and neurons was composed for the high-dimensional network. For network training, the input of the neural network was defined as the damage states of the k-dimensional network facilities, and the output was defined as the network system performance. To generate training data, random sampling was performed between earthquake magnitudes of 5.0 and 7.5, and hydraulic analyses were conducted to evaluate network performance. For a hydraulic simulation, EPANET-based MATLAB code was developed, and a pressure-driven analysis approach was adopted to represent an unsteady-state network. To demonstrate the constructed surrogate model, the actual water distribution network of A-city, South Korea, was adopted, and the network map was reconstructed from the geographic information system data. The surrogate model was able to predict network performance within a 3% relative error at trained epicenters in drastically reduced time. In addition, the accuracy of the surrogate model was estimated to within 3% relative error (5% for network performance lower than 0.2) at different epicenters to verify the robustness of the epicenter location. Therefore, it is concluded that ANN-based surrogate model can be utilized as an alternative model for efficient seismic risk assessment to within 5% of relative error.

• Proceeding of Spring/Autumn Annual Conference of KHA
• /
• 2006.11a
• /
• pp.196-201
• /
• 2006

Existing model houses have played an important role in allow the customers to choose apartment. As the information technology has been advanced (e.g. a high-speed internet available in unit), customers' personality and preference to the design of apartment and the purchasing pattern have been changed. Construction firms have introduced VR(Virtual Reality) model house (e.g. Quick Time Virtual Reality) to meet the customers' expectation and need. The reality-based QTVR model house does not provide enough quality to satisfy the customers' expectation. To complement the shortcoming of the QTVR model house, this study presents a web-based cyber model house developed by using Turntool and Javascript. The cyber model house allows to communicate between supplier and customer over the internet.

• Transactions of the Korean Society of Automotive Engineers
• /
• v.18 no.3
• /
• pp.26-36
• /
• 2010

This paper provides a description of the combustion model to obtain an accurate dynamic engine phenomena that satisfies real-time simulation for model-based engine control. The combustion chamber is modeled as a storage device for mass and energy. The combustion process is modeled in terms of a two-zone model for the burned and unburned gas fractions. The mass fraction burnt is modeled in terms of a Wiebe function. The instantaneous net engine torque is calculated from the engine speed and the instantaneous piston work. The modeling accuracy has been tested with a cylinder pressure data on a test bench and also the ability of real-time simulation has been checked. The results show that combustion model yields sufficiently good performance for the model-based control logic design. However the influence factors effected on model accuracy are some room for improvement.

• Journal of the Korean Data and Information Science Society
• /
• v.16 no.4
• /
• pp.1147-1157
• /
• 2005

Bonus-Malus system is generally constructed based on claim frequency and Bayesian credibility model is used to represent claim frequency distribution. However, there is a problem with traditionally used credibility model for the purpose of constructing bonus-malus system. In traditional Bonus-Malus system adopted credibility model, individual estimates of premium rates for insureds are determined based solely on the total number of claim frequency without considering when those claims occurred. In this paper, a new model which is a modification of structural time series model applicable to counting time series data are suggested. Based on the suggested model relatively higher premium rates are charged to insured with more claim records.

• Proceedings of the KSME Conference
• /
• 2004.04a
• /
• pp.1029-1034
• /
• 2004

Recent engineering process requires fast development and manufacturing of the products. This paper mainly discusses the process of rapid product development (RPD) from the reverse engineering to the optimal design. A laser scanning system scans a product and the efficient data processing method reduces the scanned point data. The reduced (scanned) points model is transformed to a finite element model without the construction of a CAD model. Since CAD modeling is a time-consuming work, skipping this step can save much time. This FE model is updated from the result based on the structural characteristics from modal test of the real model. For FE model updating, Response Surface Method is adopted. Finally, the updated FE model is optimized using the reliability-based topology optimization, which is developed recently. All these processes are applied to the design of an upper part model of a cellular phone.

• Journal of the Korean Society of Radiology
• /
• v.18 no.5
• /
• pp.541-549
• /
• 2024

The various structures of artificial neural networks, such as convolutional neural networks (CNNs) and recurrent neural networks (RNNs), have been extensively studied and served as the backbone of numerous models. Among these, a transformer architecture has demonstrated its potential for natural language processing and become a subject of in-depth research. Currently, the techniques can be adapted for image processing through the modifications of its internal structure, leading to the development of Vision transformer (ViT) models. The ViTs have shown high accuracy and performance with large data-sets. This study aims to develop a ViT-based model for detecting pneumonia using chest X-ray images and quantitatively evaluate its performance. The various architectures of the ViT-based model were constructed by varying the number of encoder blocks, and different patch sizes were applied for network training. Also, the performance of the ViT-based model was compared to the CNN-based models, such as VGGNet, GoogLeNet, and ResNet. The results showed that the traninig efficiency and accuracy of the ViT-based model depended on the number of encoder blocks and the patch size, and the F1 scores of the ViT-based model ranged from 0.875 to 0.919. The training effeciency of the ViT-based model with a large patch size was superior to the CNN-based models, and the pneumonia detection accuracy of the ViT-based model was higher than that of the VGGNet. In conclusion, the ViT-based model can be potentially used for pneumonia detection using chest X-ray images, and the clinical availability of the ViT-based model would be improved by this study.

• Journal of The Korean Association of Information Education
• /
• v.6 no.2
• /
• pp.187-200
• /
• 2002

In this study, we developed Internet Based Problem Solving Learning Model based on difficulty of internet based problem solving teaching and learning. For these, firstly we developed 7 Steps Problem Solving Model applying ICT-EUS into proper step. Next, we developed Internet Based Problem Solving Learning Model reflecting 7 Steps Problem Solving Model and searched an outline and characteristics of support system to apply the model into instruction. The teaching and learning model is composed of four steps of (1) design (2) preparation (3) teaching and learning execution (4) management, and 7 Steps Problem Solving Model is the core of teaching and learning execution step. The 7 Steps Problem Solving Model and Internet Based Problem Solving Learning Model are not for functional use but for general use. In other words they can be used commonly in all subjects.

• Journal of Intelligence and Information Systems
• /
• v.6 no.1
• /
• pp.35-50
• /
• 2000

Diverse requirements of users on web-based model management force a system agent to develop user-adaptive building a model in reality and providing an adequate solution method of the model. The relationship between models is important knowledge for the agent to effectively build a new model to adaptively adjust an existing model under a problem and to efficiently connect the new model into an adequate solution method. Since the generating process of the inter-model relationship is more difficult than the building a new model however the process mostly depends on the knowledge of operation research experts. Without the adequate scheme of the inter-model relationship the burden of the management for the agent increases rapidly and the quality of the services may worsen. This study shows that meta-knowledge generated from relationship between models is important for the user to build a model in reality and to acquire the solver appropriate to the model. The relationship that consists of common and exclusive objects between models can be represented by frames. The system under development to implement the idea includes user-adaptive ability which identifies a model through forward chaining method and searches the solver appropriate to the model by using the meta knowledge. We illustrate the meta knowledge with an applied delivery system in supply chain management.

• Journal of information and communication convergence engineering
• /
• v.9 no.6
• /
• pp.666-670
• /
• 2011

In this paper, we propose a stable active contour based tracking method which utilizes the bimodal segmentation technique to obtain a background color diminished image frame. The proposed method overcomes the drawback of the Mansouri model which is liable to fall into a local minimum state when colors appear in the background that are similar to the target colors. The Mansouri model has been a foundation for active contour based tracking methods, since it is derived from a probability based interpretation. By stabilizing the model with the proposed speed function, the proposed model opens the way to extend probability based active contour tracking for practical applications.