• 대한설비공학회:학술대회논문집
• /
• 대한설비공학회 2008년도 동계학술발표대회 논문집
• /
• pp.487-492
• /
• 2008

The cooling of data centers has emerged as a significant challenge as the density of IT server increases. Server installations, along with the shrinking physical size of servers and storage systems, has resulted in high power density and high heat density. The introduction of high density enclosures into a data center creates the potential for "hot spots" within the room that the cooling system may not be able to address, since traditional designs assume relatively uniform cooling patterns within a data center. The cooling system for data center consists of a CRAC or CRAH unit and the associated air distribution system. It is the configuration of the distribution system that primarily distinguishes the different types of data center cooling systems, this is the main subject of this paper.

• 한국컴퓨터정보학회논문지
• /
• 제29권7호
• /
• pp.149-158
• /
• 2024

본 연구는 다양한 실내 환경에서 공기질을 모니터링 할 수 있는 다중 센서 시스템인 AirDeep-Room을 설계 및 개발하였다. 이 시스템은 CO2, TVOC, 미세먼지, 온도, 습도 등을 실시간으로 측정하며, 멀티 센서를 통해 다양한 환경에서의 공기질 변화를 분석할 수 있는 장점을 제공한다. AirDeep-Room 시스템은 low data 형식으로 데이터를 실시간으로 확인할 수 있으며, 실내 환경에서의 공기질에 대한 분석을 진행하였다. 설계한 센서 시스템을 상이한 환경인 서버실과 강의실에 설치해 실험을 수행한 결과, 서버실에서의 온도와 습도는 -0.24의 음의 상관관계를, 강의실에서는 0.43의 양의 상관관계를 보였다. 강의실에서의 실험 결과는 학생 수와 CO2 및 TVOC 농도가 r=0.69로 높은 상관관계를 보였다. 이는 학생 수가 실내 공기질 변화에 중요한 영향을 미칠 수 있음을 보여준다. 이러한 결과는 AirDeep-Room 시스템이 다양한 환경에서 공기질을 보다 효과적으로 관리할 수 있는 기반을 제공하며, 인구 밀도가 높은 공간에서 공기질 개선을 위한 중요한 역할을 할 수 있을 것으로 기대된다.

• 융합정보논문지
• /
• 제11권11호
• /
• pp.57-65
• /
• 2021

본 논문은 딥 러닝(Deep Learning)을 이용하여 대기오염측정망 데이터 중 특정 증상이 나타나는 이상 데이터를 탐지하는 방법을 제시한다. 기존 방법들은 일반적으로 시계열 데이터 내에서 기존과는 다른 특이한 패턴이 나타나는 데이터를 탐지하여 이상치로 분류하며, 이는 특정 증상만을 탐지하기에는 적합하지 않다. 본 논문에서는 주로 이미지의 전경 분리(Sementic Segmentation)에 사용되는 DeepLab V3+ 모델의 2차원 합성곱 신경망 구조를 1차원 구조로 변형하여 이미지 대신 여러 센서의 시계열 측정값을 입력받고 특정 증상이 나타나는 데이터를 탐지하도록 하는 방법을 제시한다. 또한, 데이터에 '조각별 집계 근사법(Piecewise Aggregate Approximation)'을 적용하여 잡음이 많은 대기오염측정망 데이터의 복잡도를 줄임으로써 성능을 높인다. 실험 결과를 통해 준수한 성능으로 이상치 탐지를 수행할 수 있음을 확인할 수 있다.

• 한국컴퓨터정보학회논문지
• /
• 제23권4호
• /
• pp.139-145
• /
• 2018

In this paper, we propose the methodology and system to show weather information to spatial system. When using the spatial information system, it is easy and convenient to show information such as target location, mission contents, enemy threats and so on. However, drawing 1-dimensional weather information on 3-dimensional space in spatial information system is hard task. To fuse data, we need to add a spatial layer including weather information to spatial layers and perform space modeling for showing weather information as spatial data in a virtual space. The virtual space is shown by receiving meteorological data and then changing in real time through weather database linkage.

• 한국멀티미디어학회논문지
• /
• 제20권2호
• /
• pp.363-370
• /
• 2017

Data assimilation is an initializing method for air quality forecasting such as PM10. It is very important to enhance the forecasting accuracy. Optimal interpolation is one of the data assimilation techniques. It is very effective and widely used in air quality forecasting fields. The technique, however, requires too much memory space and long execution time. It makes the PM10 air quality forecasting difficult in real time. We propose a fast optimal interpolation data assimilation method for PM10 air quality forecasting using a new kernel tridiagonal sparse matrix and CUDA massively parallel processing architecture. Experimental results show the proposed method is 5~56 times faster than conventional ones.

• 대한설비공학회:학술대회논문집
• /
• 대한설비공학회 2006년도 하계학술발표대회 논문집
• /
• pp.938-943
• /
• 2006

The primary purpose of a computer room of data center and associated infrastructure is to support the operation of critical IT equipment. Traditionally, most owners of large critical data centers have been more than willing to accept a reasonable amount of computer room worker discomfort if necessary to support critical IT systems. All electrical equipment produces heat, which must be removed to prevent the equipment temperature from rising to an unacceptable level. Most information technology equipment and other equipment found in a data center or network room is air-cooled. Sizing a cooling system requires an understanding of the amount of heat produced by the equipment contained in the enclosed space, along with the heat produced by the other heat sources typically encountered.

• 환경영향평가
• /
• 제9권4호
• /
• pp.323-338
• /
• 2000

A Korean air dispersion modeling software, AirMaster, was developed on a basis of dispersion theories adopted in U.S. EPA's ISC3 (Industrial Source Complex - version 3) model to assess the air quality impact from the stacks. Key characteristics of AirMaster are as follows: 1) The building downwash effect can be easily simulated; 2) The screen, long term, and short term models can be run independently; 3) The input data to run the model such as meteorological and terrain data are supplied automatically from the databases in AirMaster; and 4) The modeling procedure is easy and simple under the GUI window environment. In order to validate AirMaster, comparisons with ISC3 model and Indianapolis tracer experiment were carried out. It was shown that AirMaster was identical to ISCST3 and ISCLT3 models in predicting the 1 hr to annual concentrations from the stack under various stack emission and meteorological conditions. The 1 hr concentrations predicted by AirMaster also showed a good agreement with the Indianapolis tracer measurements.

• 융합정보논문지
• /
• 제12권4호
• /
• pp.260-269
• /
• 2022

본 논문에서는 기존에 전문가에 의해서 이루어지던 국가 대기오염 측정망 데이터들의 이상 탐지 작업을 인공지능을 통해 자동화하고자 심층 신경망을 이용한 이상 탐지 모델을 제안하였다. 환경과학원에서 제공받은 기상자료 데이터의 결측치 및 이상치를 분석하여 학습데이터를 생성하였으며 비지도 학습 방식의 BeatGAN 모델에 기반하여 커널 구조 변경과 합성곱 필터층 및 전치 합성곱 필터층의 추가를 통해 새로운 모델을 제안하여 이상 탐지 성능을 높이고자 하였다. 또한 제안하는 모델의 생성적 특징을 활용하여 새로운 데이터를 생성하고 이를 학습에 사용하는 재학습 알고리즘을 구현 및 적용하여 기존 BeatGAN 모델뿐 아니라 다른 비지도 학습 모델인 Iforest, One Class SVM과 비교하였을 때 제안모델의 성능이 가장 높았음을 확인할 수 있었다. 본 연구를 통해 실제 산업현장에서 센서의 이상, 점검 등의 여러 요인으로 인해 학습 데이터가 부족한 상황에서 추가적인 비용없이 과적합을 피하며 제안하는 모델의 이상탐지 성능을 올릴 수 있는 방법을 제시할 수 있었다.

• 한국분무공학회지
• /
• 제19권2호
• /
• pp.64-68
• /
• 2014

A driving vehicle performance which is driven by FTP-75 mode was simulated by computer. Throttle valve position, engine speed, air mass flow rate, fuel consumption et al. were computer simulated. A set of engine part load performance data, automatic transmission shift map and vehicle specifications were used for the computer simulation. Throttle valve position, engine speed, air mass flow rate et al. measured for evaluating the computer simulation results by driving the vehicle with FTP-75 mode on a chassis dynamometer. GT-Power$^{(R)}$ software was used for the computer simulation of the driving vehicle performance. Experimental fuel consumption rate was measured by using an ECU HILS fuel injection system. The experimental data and simulation results were compared. The computer simulation of the driving vehicle performance predicts the measured data well comparatively.

• 설비공학논문집
• /
• 제30권2호
• /
• pp.83-91
• /
• 2018

A datacenter is a high energy consuming facility whose cooling energy consumption rate is 10~20 times larger than general office buildings. The higher the temperature of supply air from a CRAC (computer room air-conditioner) is supplied, the more energy efficient cooling is possible because of improving the COP of a chiller and advanced range of outdoor air temperature available for the economizer cycles. However, because the temperature of cold air flowing into server computers varies depending on air mixing configurations in a computer room, the proper supply air temperature must be considered based on the investigation of air mixing and heat dissipation. By these, this study aims to understand the effects of variation of the supply air temperature on the air flow distributions, temperature distributions and rack cooling efficiencies. Computational fluid dynamics (CFD) aided in conducting the investigation. As a result, the variation of the supply air temperature does not affect the air flow distributions. However, it mainly affects the temperature distribution. From the results of CFD simulations, Rack cooling indices (RCIHI and RCILO) were evaluated and showed the ideal state set at $19^{\circ}C$ of the supply air temperature.