• 한국수소및신에너지학회논문집
• /
• 제23권3호
• /
• pp.243-249
• /
• 2012

A lightweight 200W direct methanol fuel cell (DMFC) stack is designed and fabricated to power a small scale Unmanned Aerial Vehicle (UAV). The DMFC stack consists of 33-cells in which membrane-electrode assemblies (MEAs) having an active area of 88 $cm^2$ are sandwiched with lightweight composite bipolar plates. The total stack weight is around 3.485 kg and stack performance is tested under various methanol feed concentrations. The DMFC stack delivers a maximum power of 248 W at 13.2 V and $71.3^{\circ}C$ under methanol feed concentration of 1.2 M. In addition, the voltage of individual cell in the 33-cell stack is measured at various current levels to ensure the stability of DMFC stack operations. The cell voltage distribution data exhibit the maximum cell voltage deviation of 28 mV at 15 A and hence the uniformity of cell voltages is acceptable. These results clearly demonstrate that DMFC technology becomes a potential candidate for small-scale UAV applications.

• 한국수소및신에너지학회논문집
• /
• 제28권5호
• /
• pp.492-501
• /
• 2017

A 900 W scale direct methanol fuel cell (DMFC) stack is designed and fabricated for unmanned aerial vehicle (UAV) applications. To meet the volume and weight requirements, metallic bipolar plates are applied to the DMFC stack for the first time wherein POS470FC was chosen as bipolar plate material. To ensure good robustness of the metallic bipolar plate based DMFC stack, finite element method based simulations are conducted using a commercial ANSYS Fluent software. The stress buildup and deformation characteristics on bipolar plates and end plates are analyzed in details. The present DMFC stack exhibits the performance of 1,130 W at 32 V and 35.3 A, clearly demonstrating that it could successfully operate for UAVs requiring around 1,000 W of power.

• 한국수소및신에너지학회논문집
• /
• 제28권5호
• /
• pp.481-491
• /
• 2017

In this study, we developed 300 W lightweight DMFC system for charging secondary battery in small unit military operation. In order to reduce the volumetric shape and weight of the system considering the environment of the individual soldier's, the arranging of system components has been optimized. A metal bipolar plates made of STS-470FC have been implemented to the DMFC stack to meet the weight demand of the system. As a result of the performance test of the stack, the target value was satisfied by outputting 561 W exceeding 24% of the stack output 450 W required to output 300 W required for the entire system. Moreover, 2,655 hours exceeding 1,000 hours also has been satisfied. To ensure good robustness of the metallic bipolar plate based DMFC stack, finite element method based simulations are conducted using a commercial ANSYS Fluent software.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제17권9호
• /
• pp.2505-2528
• /
• 2023

The field of facial expression recognition (FER) has been actively researched to improve human-computer interaction. In recent years, deep learning techniques have gained popularity for addressing FER, with numerous studies proposing end-to-end frameworks that stack or widen significant convolutional neural network layers. While this has led to improved performance, it has also resulted in larger model sizes and longer inference times. To overcome this challenge, our work introduces a novel lightweight model architecture. The architecture incorporates three key factors: Depth-wise Separable Convolution, Residual Block, and Attention Modules. By doing so, we aim to strike a balance between model size, inference speed, and accuracy in FER tasks. Through extensive experimentation on popular benchmark FER datasets, our proposed method has demonstrated promising results. Notably, it stands out due to its substantial reduction in parameter count and faster inference time, while maintaining accuracy levels comparable to other lightweight models discussed in the existing literature.

• 신재생에너지
• /
• 제3권1호
• /
• pp.46-53
• /
• 2007

Direct Methanol Fuel Cell, DMFC is a potential power source for portable IT application. DMFC works at low temperature ($<100^{\circ}C$) without fuel processing. Methanol has high energy density, fuel economy, and easiness to handle. This paper focuses high efficient catalyst to increase utilization in the electrode, new membrane reducing methanol crossover, new material parts, and optimization of system integration. Lightweight and small-sized DMFC based on new materials, efficient stack, and improved system control will be applied to the 50W prototype system for the notebook computer.

• 한국소성가공학회:학술대회논문집
• /
• 한국소성가공학회 1999년도 춘계학술대회논문집
• /
• pp.188-194
• /
• 1999

Fiber reinforced composite materials are widely used in automotive industry to produce parts that are large, thin. lightweight. strong and stiff. It is very important to know a charge shape in order to have good products in the compression molding. In particular, the product such as a bumper beam is composed of the random and unidirectional fiber mats. This study analyzes numerically the characteristics of flow fronts such as a bulging phenomenon made by changing viscosity of random mat and unidirectional fiber mat and slip parameters. And it is discussed that the effect of ratio of viscosity A and stack type on mold filling parameters

• Composites Research
• /
• 제12권5호
• /
• pp.23-30
• /
• 1999

Fiber reinforced composite materials are widely used in automotive industry to produce parts that are large, thin, lightweight, strong and stiff. It is very important to know a charge shape in order to have good products in the compression molding. In particular, the product such as a bumper beam is composed of the random and unidirectional fiber mats. The characteristics of flow fronts such as a bulging phenomenon for random mat and unidirectional fiber mat and slip parameters are studied numerically. And the effects of viscosity ratio and stack type on mold filling parameters are also discussed.

• 한국정보통신학회논문지
• /
• 제15권6호
• /
• pp.1301-1306
• /
• 2011

RTOS는 임베디드 시스템 환경에서 멀티태스킹 동작을 설계하는데 강력한 도구이다. 그러나 협소한 메모리를 가진 MCU에서는 응용프로그램에 비해 기존의 RTOS가 차지하는 비율이 크기 때문에 적용하기 어려웠다. 본 논문에서는 기존의 RTOS에서 사용빈도가 적은 기능을 제거하고, 스케줄링과 자원 관리의 기능만 가지고 최소한의 코드로 동작하는 경량화된 RTOS를 설계하였다. 공유 스택을 사용하여 사용자 메모리를 확보하며, 태스크의 문맥 전환시에 발생하는 오버헤드를 감소시키고, TCB등의 사이즈를 축소하는 기법을 사용하였다. 설계 및 검증 결과, 커널의 사이즈를 1KB이하로 축소할 수 있었고, 커널과 응용 프로그램의 비율을 고려해 볼 때, 본 논문에서 설계한 RTOS는 4KB이상의 프로그램 메모리를 가진 MCU에서 사용할 수 있다.

• 한국정보전자통신기술학회논문지
• /
• 제11권1호
• /
• pp.26-33
• /
• 2018

A web application running on the Internet is causing many difficulties for a program developer, and it requires to process multiple sessions at the same time due to the occurrence of excessive traffic. Web applications should be able to process concurrent requests efficiently and in real time. Node.js is a single-threaded server-side JavaScript environment implemented in C and C ++ as one of the latest frameworks to implement event models across the entire stack. Nodes implement JavaScript quickly and robust to achieve the best performance using a JavaScript V8 engine developed by Google. In this paper, it will be explained the operation principle of Node.js, which is a lightweight real-time web server that can be implemented in JavaScript for real-time responsive web applications. In addition, this application was practically implemented through automatic alignment system for group scheduling to demonstrate event-based real-time response web processing.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제14권3호
• /
• pp.1188-1203
• /
• 2020

The human expression contains a lot of information that can be used to detect complex conditions such as pain and fatigue. After deep learning became the mainstream method, the traditional feature extraction method no longer has advantages. However, in order to achieve higher accuracy, researchers continue to stack the number of layers of the neural network, which makes the real-time performance of the model weak. Therefore, this paper proposed an expression recognition framework based on densely concatenated convolutional neural networks to balance accuracy and latency and apply it to humanoid robots. The techniques of feature reuse and parameter compression in the framework improved the learning ability of the model and greatly reduced the parameters. Experiments showed that the proposed model can reduce tens of times the parameters at the expense of little accuracy.