• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.17 no.1
• /
• pp.88-98
• /
• 2016

The Korean summer is hot and humid, and air-conditioners consume considerable amounts of electricity. In such cases, the simultaneous use of indirect evaporative coolers may help reduce the sensible heat and save electricity. In this study, heat transfer and pressure drop characteristics of indirect or regenerative evaporative coolers made from plastic/paper are investigated. The results showed that heat and mass transfer model based on the ${\epsilon}-NTU$ method predicted the indirect evaporation efficiencies, cooling capacities and pressure drops adequately. Both for indirect or regenerative evaporative cooler, the indirect evaporation efficiency increased with increasing dry channel inlet temperature or relative humidity. The indirect evaporation efficiency of the regenerative evaporative cooler was larger than that of the indirect evaporative cooler.

• The Transactions of the Korea Information Processing Society
• /
• v.4 no.7
• /
• pp.1788-1796
• /
• 1997

In this paper, we propose a new connection admission control(CAC) algorithm for traffic control in ATM network in which traffic estimation is performed based on user-specified parameters at every moment of connection request or connection release by recursive formula which makes real-time calculation possible. And traffic estimation using cell flow measurement is carried out when the number of connectioned calls does not change during a measurement reflection period. Performance analysis of the proposed method is carried out using several aspects for homogeneous and heterogeneous bursty traffic. The results showed that the proposed CAC method revealed better performance, than conventional CAC method for burst model in both utilization and QoS point of view.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.12 no.2
• /
• pp.895-900
• /
• 2011

To improve drawback of existing modulation method, in this paper, we propose the channel forecast method using adaptive modulation which can improve throughput of channel. This method adaptively changes modulation method to the change of channel environments. In proposed method, channel's characteristics are measured in realtime to determine code rate to the changes of demanded channel's bit error rate. If bit error rate is increased, this method reduce code rate to maintain maximum throughput. We analysis performance of proposed method by Matlab.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2022.05a
• /
• pp.167-167
• /
• 2022

최근 빅데이터에 관련된 연구에 있어 데이터의 품질관리에 대한 논의가 꾸준히 이뤄져 오고 있다. 특히 이미지 처리 및 분석에 활용되어온 딥러닝 기술의 경우, 분류 작업 및 패턴인식 등으로부터 데이터의 특징을 추출함으로써 비지도학습(Unsupervised Learning)을 가능하게 한다는 장점이 있음에도 불구하고 빅데이터를 다루는 과정에 있어 용량, 다양성, 속도 및 신뢰성 측면에서의 한계가 있었다. 본 연구에서는 CCTV 영상을 활용한 강수량 산정 모델 개발에 있어 예측 정확도 향상 및 성능 개선을 도모할 수 있는 데이터 전처리 방법을 제안하였다. 서울 근린 AWS 4개소 지역(김포장기, 하남덕풍, 강동, 성남) 및 중앙대학교 지점 내 CCTV를 설치한 후, 최대 9개월의 영상을 확보하여 강수량 산정을 위한 딥러닝 모델을 개발하였다. 배경분리, 조도조정, 영역설정, 데이터증진, 이상데이터 분류 등이 가능한 알고리즘을 개발함으로써 데이터셋 자체에 대한 전처리 작업을 수행한 후, 이에 대한 결과를 기존 관측자료와 비교·분석하였다. 본 연구에서 제안한 전처리 방법들을 적용한 결과, 강수량 산정 모델의 예측 정확도를 평가하는 지표로 선정한 평균 제곱근 편차(Root Mean Square Error; RMSE)가 약 30% 감소함을 확인하였다. 본 연구의 결과로부터 CCTV 영상 데이터를 활용한 강수량 산정의 가능성을 확인할 수 있었으며 특히, 딥러닝 모델 개발시 필요한 적정 전처리 방법들에 대한 기준을 제시할 수 있을 것으로 판단된다.

• Journal of the Korean Institute of Telematics and Electronics C
• /
• v.36C no.7
• /
• pp.36-45
• /
• 1999

In order to control the flow of traffics in ATM networks and optimize the usage of network resources, an efficient control mechanism is necessary to cope with congestion and prevent the degradation of network performance caused by congestion. In this paper, Buffered Leaky Bucket which applies the same control scheme to a variety of traffics requiring the different QoS(Quality of Service) and Neural Networks lead to the effective buffer utilization and QoS enhancement in aspects of cell loss rate and mean transfer delay. And the cell scheduling algorithms such as DWRR and DWEDF for multiplexing the incoming traffics are enhanced to get the better fair delay. The network congestion information from cell scheduler is used to control the predicted traffic loss rate of Neural Leaky Bucket, and token generation rate and buffer threshold are changed by the predicted values. The prediction of traffic loss rate by neural networks can enhance efficiency in controlling the cell loss rate and cell transfer delay of next incoming cells and also be applied for other traffic controlling schemes. Computer simulation results performed for random cell generation and traffic prediction show that QoSs of the various kinds of traffcis are increased.

• Journal of the Korean Society of Marine Environment & Safety
• /
• v.22 no.1
• /
• pp.129-137
• /
• 2016

In the present study we verified performance of feed-forward control algorithm using short term prediction of ship motion information by taking advantage of developed numerical simulation model of FPSO motion. Up until now, various studies have been conducted about thrust control and allocation for dynamic positioning systems maintaining positions of ships or marine structures in diverse sea environmental conditions. In the existing studies, however, the dynamic positioning systems consist of only feedback control gains using a motion of vessel derived from environmental loads such as current, wind and wave. This study addresses dynamic positioning systems which have feedforward control gain derived from forecasted value of a motion of vessel occurred by current, wind and wave force. In this study, the future motion of vessel is forecasted via Brown's Exponential Smoothing after calculating the vessel motion via a selected mathematical model, and the control force for maintaining the position and heading angle of a vessel is decided by the feedback controller and the feedforward controller using PID theory and forecasted vessel motion respectively. For the allocation of thrusts, the Lagrange Multiplier Method is exploited. By constructing a simulation code for a dynamic positioning system of FPSO, the performance of feedforward control system which has feedback controller and feedforward controller was assessed. According to the result of this study, in case of using feedforward control system, it shows smaller maximum thrust power than using conventional feedback control system.

• Journal of the Computational Structural Engineering Institute of Korea
• /
• v.36 no.1
• /
• pp.9-18
• /
• 2023

Predicting the compressive strength of high-performance concrete (HPC) is challenging because of the use of additional cementitious materials; thus, the development of improved predictive models is essential. The purpose of this study was to develop an HPC compressive-strength prediction model using an ensemble machine-learning method of combined bagging and stacking techniques. The result is a new ensemble technique that integrates the existing ensemble methods of bagging and stacking to solve the problems of a single machine-learning model and improve the prediction performance of the model. The nonlinear regression, support vector machine, artificial neural network, and Gaussian process regression approaches were used as single machine-learning methods and bagging and stacking techniques as ensemble machine-learning methods. As a result, the model of the proposed method showed improved accuracy results compared with single machine-learning models, an individual bagging technique model, and a stacking technique model. This was confirmed through a comparison of four representative performance indicators, verifying the effectiveness of the method.

• Korean Journal of Remote Sensing
• /
• v.40 no.1
• /
• pp.71-80
• /
• 2024

In this study, a proposed ensemble learning technique aims to enhance the semantic segmentation performance of images captured by Unmanned Aerial Vehicles (UAVs). With the increasing use of UAVs in fields such as urban planning, there has been active development of techniques utilizing deep learning segmentation methods for land cover segmentation. The study suggests a method that utilizes prominent segmentation models, namely U-Net, DeepLabV3, and Fully Convolutional Network (FCN), to improve segmentation prediction performance. The proposed approach integrates training loss, validation accuracy, and class score of the three segmentation models to enhance overall prediction performance. The method was applied and evaluated on a land cover segmentation problem involving seven classes: buildings,roads, parking lots, fields, trees, empty spaces, and areas with unspecified labels, using images captured by UAVs. The performance of the ensemble model was evaluated by mean Intersection over Union (mIoU), and the results of comparing the proposed ensemble model with the three existing segmentation methods showed that mIoU performance was improved. Consequently, the study confirms that the proposed technique can enhance the performance of semantic segmentation models.

• Journal of the Institute of Convergence Signal Processing
• /
• v.14 no.2
• /
• pp.99-104
• /
• 2013

Replacement of sensor nodes for monitoring a wide range area such as mountains and forests needs a lot of time and cost. Using new and renewable energy around them can maximize the lifetime of wireless sensor networks, in which solar energy is infinite energy source that is available in 365 days. To design these sensor networks, solar energy model is essential and to estimate and analyze the overall photovoltaic energy. Using this, we can figure out important data such as the size and performance of solar panel needed. However, existing researches for solar energy harvesting consider parts of many factors to influence the quantity of solar energy gathered. In this paper, we suggest advanced solar energy harvesting model considering angular loss (solar cell panel), overheat loss (solar cell), rechargeable battery heat and cooling for each monthly properties. From our experimental results according to outdoor temperature, panel angle and the surface temperature of solar panel, we show these impact factors are correctly configured.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.27 no.4C
• /
• pp.371-379
• /
• 2002

In this paper, the MMIC SPST switches operating from DC to 3GHz were designed and implemented. Prior to the design of switches, the small and large-signal switch models were needed to predict switch performance accurately. The newly proposed small-signal switch model parameters were extracted from measured S-parameters using optimization technique with estimated initial values and boundary limits. In the extraction of large-signal switch model parameters, the current source was modeled by fitting empirical equations to measured DC data and the charge model was derived from extracted channel capacitances from measured S-parameters varying the drain-source voltage. To design basic series-shunt SPST switches and isolation-improved SPST switches, we applied this model to commercial microwave circuit simulator. The improved SPST switches exhibited 0.302dB insertion loss, 35.762dB isolation, 1.249 input VSWR, 1.254 output VSWR, and about 15.7dBm PldB with 0/-3V control voltages at 3GHz.