Earticle

최근 인공지능을 활용한 서비스 고급화는 선택이 아닌 필수가 되었다. 그에 따라 인공지능의 연구도 가속화되어 효 율적인 인공지능 연구를 위한 방법 또한 주목 받고 있다. 대표적인 방법으로는 연구 과정에서 효과적으로 실험을 관 리할 수 있는 도구를 사용하는 것이다. 기존의 딥러닝 연구는 파편적인 작업 방식을 기반으로 한 협업과 학습 결과 최적화를 위한 반복 작업으로 인해 비효율적이었다. 본 연구는 이러한 문제점을 개선하기 위해 인공지능 분야 중 딥 러닝에 초점을 맞추어 편리하고 생산적인 연구 환경을 제공할 수 있는 협업 기반의 딥러닝 실험 관리 도구인 digo(dig와 go의 합성어로 반복적인 딥러닝 연구를 표현하는 단어를 조합)를 설계 및 구현하였다. 기계학습 연구자 들을 대상으로 실험 및 설문조사를 실시하여 딥러닝 실험 관리 도구의 성능을 검증하였고, 하이퍼 파라미터 자동 최 적화 및 학습 결과 시각화 기능의 편의성을 확인하였다.

Recently, advanced service using artificial intelligence has become a necessity, not an option. As a result, research on artificial intelligence has been accelerated, drawing attention to methods for efficient artificial intelligence research. A typical method is to use tools to effectively manage experiments in the course of the study. Existing deep learning studies have been inefficient due to collaboration based on fragmentary work methods and repetitive tasks for optimizing learning results. To improve these problems, this work designs and implements Digo(a combination of words that represent repetitive deep learning research as a compound word of dig and go), a collaborative-based deep learning experiment management tool that can provide a convenient and productive research environment, focusing on deep learning among artificial intelligence. Experiments and surveys were conducted on machine learning researchers to validate the performance of deep learning experimental management tools, and to confirm the convenience of hyperparameter automatic optimization and learning result visualization features.

2차 침전지 유출수 총인 농도의 실시간 제어를 위하여 Auto-Jar가 장착된 응집제 주입량 최적제어 시스템(Advanced Molit Coagulant-dosing Optimization System, A-MCOS)을 개발하여, 선행 연구의 MCOS가 운전되었던 J-STP에서 2014년 5월 7일부터 23일까지 현장적용 가능성을 검토하였다. Pilot plant 운전 결과, A-MCOS에 의한 응집제 절감율은 정율주입의 대조군과 비교시 평균 20.5%를 나타내었다. 응집제 사용 비용과 응집제 사용에 따른 슬러지발생 처리비용을 합산하여 경제성을 분석하였다. 그 결과 A-MCOS 실험군의 처리비용 증가액은 5.0 won/m3으로 선행연구의 MCOS 운전체계로 발생된 처리비용 증가액의 1/3 수준이었으며, 대조군 대비 18.0%의 비용절감 효과를 나타내었다. 이는 AMCOS가 선행연구의 MCOS보다 향상된 비용효율성을 지니며, TP처리를 위한 현장적용 가능한 대안이 될 수 있음을 시사해주었다.

A-MCOS (Advanced Molit Coagulant-dosing Optimization System) equipped with an Auto-Jar was devised for a realtime control of total phosphorus (TP) in secondary effluent, then tested for field applicability starting from May 7 till May 23 in 2014 at J-STP where MCOS was operated in our previous study. The mean coagulant reduction rate by the A-MCOS was 20.5%, compared with a control where coagulant was applied at a constant rate. Total cost for A-MCOS or control operation was estimated by summing up coagulant cost and sludge treatment cost. Thereby, the cost increase by A-MCOS would be 5.0 won/m3, corresponding to one-third of that by MCOS, which would result in 18% of cost reduction compared with a control operation. These results indicate that A-MCOS can be an alternative for TP treatment, with an improved cost efficiency over MCOS.

Spirulina platensis is a unique photosynthetic cyanobacterium that is used as a commercial feed for animals and a dietary supplement for humans. S. platensis has significant auto-flotation activity due to its gas vesicles. The natural flotation activity of these cells provides an effective method for harvesting of Spirulina biomass. The present research found that hydrodynamic pressure, light level and bicarbonate concentration affected the natural flotation of S. platensis. In particular, auto-flotation increased as light level decreased (5.5-times greater at no supplemental light than at 300 μ㏖ photons/㎡/s), as bicarbonate level increased (5-times higher at 0.6 than 0.2 M) and as hydrodynamic pressure decreased (indicated by depth of the culture). These results should be considered when designing photobioreactors for direct harvesting by auto-flotation.

Objective: The aim of this study is to design an auto showering system regarding bed type for old people or bedridden patients. Background: The rapid growth concerning the aging population leads to an increase in elderly bedridden patients living in senior care centers and hospitals where care givers have difficulty bathing older people with limited mobility. Method: In this study, a showering equipment for experiment was based on anthropometric and researching existing products. The nozzle position was optimized by showering tests and simulations using showering equipment. Results: The problems regarding the existing products were analyzed and the nozzle position was optimized through showering tests. The number, position, and spraying angle of the nozzle were optimized through showering simulations. The automatic showering system concerning bed type was designed by considering their results. Conclusion: When designing an automatic showering system, the most important design element involves determining the position of nozzle and angle of water injection and the number of nozzles. The system was developed by applying the results through user centered-research. Application: The user centered-research for developing products was applied directly in order to develop automatic baths, showering products etc. Furthermore, it was also available to be applied towards senior friendly products.

An automatic pipe routing system is proposed and implemented. Generally, the pipe routing design as a part of the shipbuilding process requires a considerable number of man hours due to the complexity which comes from physical and operational constraints and the crucial influence on outfitting construction productivity. Therefore, the automation of pipe routing design operations and processes has always been one of the most important goals for improvements in shipbuilding design. The proposed system is applied to a pipe routing design in the engine room space of a commercial ship. The effectiveness of this system is verified as a reasonable form of support for pipe routing design jobs. The automatic routing result of this system can serve as a good basis model in the initial stages of pipe routing design, allowing the designer to reduce their design lead time significantly. As a result, the design productivity overall can be improved with this automatic pipe routing system.

LCD(Liquid Crystal Display)는 자체 발광능력이 없으므로 그 후면에서 LCD 화면을 밝혀주는 BLU(Backlight Unit)를 필요로 한다. BLU의 핵심부품으로써 광원으로부터 나온 빛을 LCD의 패널 방향으로 이끌어내는 역할을 담당하고 있는 부분이 바로 도광판이고, 빛을 패널 방향으로 고르게 분포시키기 위해 도광판 하부에 점 모양의 확산물질을 인쇄한다. 이 인쇄패턴은 광원으로부터의 거리에 따라서 그 크기를 다르게 함으로써 패널 부분으로 균일한 평면광을 생성하도록 한다. 따라서 보다 균일한 평면광을 생성하는 BLU를 개발하기 위해서는 주어진 조건에서 패널 부분으로 균일한 평면광을 생성하도록 설계된 인쇄패턴을 찾는 것이 관건이다. 이러한 인쇄패턴을 Simplex Search 알고리즘을 사용하여 자동으로 찾게 함으로써 수작업으로 찾는 것보다 빠른 시간 내에 더욱 효율적인 인쇄패턴을 찾도록 한다.

본 논문에서는 선행연구에서 신뢰성을 확보한 지능형 브레이커의 해석모델을 바탕으로 지능형 브레이커의 타격성능 개선을 위한 설계변수 민감도 분석 및 다목적함수 최적화를 다룬다. 본 연구에서는 타격출력만을 향상시키는 기존의 연구를 보완하여 타격출력과 안정성 모두를 개선시키는 것을 목표로 한다. 연구를 진행하는 순서는 다음과 같다. 먼저 정확한 민감도 분석 및 최적화를 진행하기 위하여 해석 시나리오를 설정한다. 그 후 지능형 브레이커의 설계변수 민감도를 분석하고, 상위 민감도를 가지는 변수를 추출한다. 마지막으로 추출한 변수를 사용하여 다목적함수 최적화를 진행하고, 초기 성능과 최적화된 성능의 비교를 통해 기존 지능형브레이커에서 타격성능이 얼마나 향상되었는지 파악한다. 이러한 연구를 통해 국내 기술력으로 기존 해외 선진사의 제품보다 타격성능이 향상된 제품을 개발할 수 있는 가능성을 확보할 수 있다.

This paper discusses the design parameter sensitivity analysis and multi-objective function optimization for improving the impact performance of an auto-sensing breaker based on the analytical model of the same, which secured reliability in a previous research. The study aims to improve both impact power and stability by complementing the existing research that only improved the impact power. The study sequence is as follows: first, the analysis scenarios for the accurate sensitivity analysis and optimization are set up. Second, the sensitivity of the design parameter of the auto-sensing breaker is analyzed, and the variables with high sensitivity are extracted. Third, the extracted variables are used to optimize the multi-objective functions, and the optimized performance is compared with the initial performance to see how the impact performance on the existing auto-sensing breaker has improved. This study is based on domestic technology, and will allow the development of products with a better blowing performance than their existing overseas counterparts.

The impact energy in low speed crash must be absorbed locally in the front structure with no damage in the rear part of an auto-body. The repair cost in a low speed crash could be reduced by establishing energy absorbing element such as a crash box. In this paper, the front structure of an auto-body is designed to reduce the repair cost in a low speed crash accident. The strength and energy absorbing ability of a crash box is also related with 5mph bumper performance. A compromise solution is acquired from an optimization process wish DOE and RSM to satisfy the 5mph bumper and 15kph crash performances.

In this paper, optimization for frontal crashworthiness is carried out for the weight reduction design of an auto-body member with the advanced high strength steels(AHSS) such as 780TRIP and 780DP. The frontal crashworthiness is evaluated in order to optimize thicknesses for the front rail member of the ULSAB-AVC, Thicknesses of the front rail member with AHSS are optimized by comparison of crushing distance, absorbed energy and the deceleration for the auto-body with the response surface methodology. The results demonstrate that the crashworhiness of the front rail member with the optimum thicknesses of the AHSS is similar to analysis results obtained from the ULSAB-AVC project. The results also show that the weight reduction design is performed by substituting the AHSS for conventional structural steels such as 440E in the auto-body members.