Organic waste-derived carbonaceous electrodes for flow-electrode capacitive deionization of brackish water

Jaegwan Shin 강원대학교 대학원 2025 국내박사

본 학위논문에서는 유기성 폐기물 유래 탄소 전극을 활용한 흐름 전극 축전식 탈염 공정의 기수 담수화 기술로서 적용 가능성을 체계적으로 평가하였다. 본 학위논문의 주된 목적은 유기성 폐기물 유래 탄소 전극의 원료와 열분해 조건을 최적화함으로써 이의 유동 전극으로써 실현 가능성을 규명하고, 기수 담수화 성능과 연관된 유기성 폐기물 유래 탄소 전극의 주요한 속성(공극, 전기전도도, 젖음성 및 탭밀도)에 알칼리성 화학적 활성화 방법과 탄산칼슘 기반 템플레이팅 방법이 미치는 영향을 조사하며, 유기성 폐기물 유래 탄소 전극을 활용한 흐름 전극 축전식 탈염 공정의 경제성을 축전식 탈염공정과 비교하여 확인하는 것이다. 유기성 폐기물 유래 탄소 전극의 제조를 위한 최적의 원료 및 열분해 가스 분위기는 각각 쌀겨와 이산화탄소였다. 높은 열분해 온도는 쌀겨 유래 탄소 전극의 물리적 속성과 방향성을 향상시키는데 효과적이었으며, 이로 인해 흐름 축전식 탈염 공정의 전기화학적 속성 또한 향상하였다. 하지만, 800°C/이산화탄소 조건에서 제조된 쌀겨 유래 탄소 전극(RHCE800-C)을 활용한 흐름 전극 축전식 탈염 공정은 500°C, 650°C 와 800°C에서 제조된 쌀겨 유래 탄소 전극을 활용하였을 때보다 수율로 정규화된 기수 담수화 성능이 더 우수하였다. 따라서, 쌀겨 유래 탄소 전극의 열분해 온도는 800°C로 설정하였다. 탄산칼슘 기반 템플레이팅 방법은 알칼리성 화학적 활성화 방법보다 RHCE800-C의 공극, 전기전도도, 젖음성 및 탭밀도를 향상하는데 더 효과적이었다. 이로 인해, 탄산칼슘 기반 템플레이팅 방법(RHCE800-C와 탄산칼슘 템플레이트 혼합비 = 0.50)으로 개질된 RHCE800-C(RHCECC0.50-800-C)를 활용한 흐름 전극 축전식 탈염 공정이 RHCE800-C와 2.0 M 수산화나트륨 기반 화학적 활성화 방법으로 개질된 RHCE800-C(RHCESH2.0-800-C)보다 더 우수한 기수 담수화 성능을 나타냈다(기수 담수화 성능: RHCE800-C = 0.0488 mg/cm2‧min, RHCESH2.0-800-C = 0.0696 mg/cm2‧min 및 RHCECC0.50-800-C = 0.0804 mg/cm2‧min). RHCE800-C, RHCESH2.0-800-C 및 RHCECC0.50-800-C를 활용한 흐름 전극 축전식 탈염 공정의 담수화 기작은 전기 투석, 전기 흡착 및 유도전류 반응이었다. RHCE800-C, RHCESH2.0-800-C 및 RHCECC0.50-800-C의 제조 비용은 활성탄(MSP-20X)보다 매우 낮은 것을 확인하였다(제조비용: RHCE800-C = 33.93 $/kg, RHCESH2.0-800-C = 49.23 $/kg, RHCECC0.50-800-C = 37.50 $/kg 및 MSP-20X = 453.80 $/kg). RHCE800-C, RHCESH2.0-800-C 및 RHCECC0.50-800-C을 활용한 흐름 전극 축전식 탈염 공정의 비용으로 정규화된 기수 담수화 성능은 모두 활성탄 보다 높은 것으로 보아, RHCE800-C, RHCESH2.0-800-C 및 RHCECC0.50-800-C은 우수한 경제성을 갖는 것으로 사료된다. 특히, RHCECC0.50-800-C을 활용한 흐름 전극 축전식 탈염 공정의 비용으로 정규화된 기수 담수화 성능이 RHCE800-C와 RHCESH2.0-800-C 보다 상당히 우수하였다. 또한, RHCECC0.50-800-C을 활용한 흐름 전극 축전식 탈염 공정의 고유 소비 전력은 축전식 탈염 공정보다 낮기 때문에, RHCECC0.50-800-C을 활용한 흐름 전극 축전식 탈염 공정은 물 부족 현상을 해결하기 위한 담수를 생산하는 유망한 기술이라고 판단된다. Brackish water desalination via flow-electrode capacitive deionization (FCDI) utilizing organic waste-derived carbonaceous electrodes for producing fresh water was systematically evaluated. The main purposes of this dissertation were to (i) identify the effects of feedstock types and pyrolysis conditions (i.e., gas atmospheres and pyrolysis conditions) for fabricating organic waste-derived carbonaceous electrodes on brackish water desalination performance of the FCDI process, (ii) investigate the effects of the chemical activation using alkali agents and the calcium carbonate (CaCO3)-based templating method on the physicochemical properties of organic waste-derived carbonaceous electrodes related to brackish water desalination performance of the FCDI process, and (iii) demonstrate the feasibility of the FCDI process utilizing organic waste-derived carbonaceous electrodes for brackish water desalination. The optimal feedstocks and gas atmospheres for the fabrication of organic waste-derived carbonaceous electrodes were rice husks and carbon dioxide, respectively. Higher pyrolysis temperatures were effective in improving the physical properties and bulk conductivity of rice husk-derived carbonaceous electrodes, resulting in excellent electrochemical performance of the FCDI stack. Nonetheless, the FCDI process utilizing rice husks-derived carbonaceous electrodes produced at 800 °C/carbon dioxide conditions (RHCE800-C) exhibited a higher average salt removal rate (ASRR) normalized with yields compared to those of 500 °C, 650 °C, and 950 °C. Hence, the pyrolysis temperatures of rice husk-derived carbonaceous electrodes were set to 800 °C. To enhance the electrochemical performance of the FCDI stack, the CaCO3-based templating method was more beneficial in improving the physicochemical properties of RHCE800-C compared to the chemical activation using alkali agents (i.e., sodium hydroxide (NaOH)). As expected, the highest brackish water desalination performance was found for the FCDI process utilizing CaCO3-templated RHCE800-C at 0.50 of the mixing ratios of RHCE800-C and CaCO3-templates (RHCECC0.50-800-C; ASRR = 0.0804 mg/cm2‧min) than RHCE800-C (ASRR = 0.0488 mg/cm2‧min) and 2.0 M NaOH-activated RHCE800-C (RHCESH2.0-800-C; ASRR = 0.0696 mg/cm2‧min). Brackish water desalination via the FCDI process utilizing RHCE800-C, RHCESH2.0-800-C, and RHCECC0.50-800-C progressed through electrodialysis, electrosorption, and Faradaic reactions. The manufacturing costs of RHCE800-C (33.93 $/kg), RHCESH2.0-800-C (49.23 $/kg), and RHCECC0.50-800-C (37.50 $/kg) were considerably lower than activated carbons (MSP-20X; 453.80 $/kg). The higher cost-normalized ASRR values of the FCDI process utilizing RHCE800-C, RHCESH2.0-800-C, and RHCECC0.50-800-C (0.226–0.343 mg/cm2·min·$) compared to MSP-20X (0.028 mg/cm2·min·$) suggest that RHCE800-C, RHCESH2.0-800-C, and RHCECC0.50-800-C have excellent economic feasibility. Among them, the FCDI process utilizing RHCECC0.50-800-C exhibited the highest cost-normalized ASRR values (0.343 mg/cm2·min·$). Moreover, the FCDI process utilizing RHCECC0.50-800-C exhibited a lower specific energy consumption (0.63–1.32 kWh/m3) than membrane capacitive deionization (specific energy consumption = 0.70–1.80 kWh/m3). Hence, the FCDI process utilizing RHCECC0.50-800-C should be a promising option to secure alternative water resources for solving the water shortage through brackish water desalination.

Flow-3D를 이용한 교각 주변 흐름의 수치해석

김창성 명지대학교 대학원 2005 국내석사

According to researchers, the influential factors of scouring are generally divided into three factors: the flow conditions, the type and position of structures, and the characteristics of bed materials. In addition, scouring is affected by the 3-dimensional turbulent boundaries, the unsteady flow, the movement of sediment in the scour-hole areas, the approach flow velocity and depth, the width of the bridge foundation/pier, and the particle size of bed materials. However, it is difficult to estimate the scour depth near bridge piers when all conditions are factored in at once. Therefore, for reasonably accurate estimates of scour depth, it is essential to consider sufficiently the flow force and resisting force for scour. That is, to determine the shear stress concerning the bed material distribution is needed. In this study, the experiments were performed under the condition of a steady state of flow. As a result, scouring occurred at velocity ratios of V/Vc=0.476, and the scour depth was increased linearly as the velocity ratio increased. In addition, the average values of shear sress ratio at zero scouring depth in both rectangular and circular piers were approximately 7(tauc/tauapproach), and in the case for same size bed particle materials, the shear stress ratio of rectangular piers were higher than circular piers. The results of this study can be used for the fundamental material for estimating the scour depth of bed materials. 최근 들어 다양한 산업에서 3차원 수치해석을 이용하는 영역이 넓혀지고 있다. 국내에서도 3차원 수치해석을 이용하여 댐이나 정수처리시설 설계 등 수리분야에 적용하였고, 점점 그 이용이 확대되고 있다. 특히 교량 및 수제 등의 수공구조물로 인한 주변 흐름형상은 2차원 해석이 불가능하여 현재 수리모형실험을 통한 해석이 주를 이루고 있다. 이에 본 연구에서는 현재 3차원 수치해석에 많이 이용되는 상용모델인 FLOW-3D를 이용하여 3차원 흐름에 대한 수치해석을 수행하였으며 수리모형 결과와 비교하여 적용성을 검토하였다. 그 결과 모의 및 실험결과 수리모형과 수치모의 결과는 서로 비슷한 흐름양상을 보이고 있었다. 또한 Fr수와 조도계수(Roughness)를 매개변수로 하여 다양한 상태에서의 FLOW-3D의 민감도를 검토하였으나 실제 흐름과는 상이함을 보였다. 또한 다양한 난류모형을 적용하여 그 결과를 실제 흐름과 비교해보았다. 그 결과 3차원모형에서는 교각 후면부와 좌, 우측면부에서의 흐름이 실제 수리 모형과는 상이한 결과를 보이는데 이는 FLOW-3D의 벽 경계 및 와류에 대한 난류모델에 따른 수치해석적인 차이로 사료된다.

The Examination of Flow Theory in Leisure and Work by Applying ESM : 경험표집법(ESM)을 이용한 여가와 일의 플로우 이론 연구

김홍일 경기대학교 대학원 2010 국내박사

사람이 특정 활동에 참여해서 그들이 가진 기술과 인식한 도전수준이 균형을 이룰 때 일어나는 최적의 경험을 Csikszentmihalyi (1975)는 플로우 (flow)라 정의하였고 그 개념을 정립한 이래로 여러 학문분야에서 사람들의 심리상태나 주관적 경험을 연구하기 위해 플로우 개념을 사용하였다. Csikszentmihalyi의 4 채널 플로우 모델에 의하면, 높은 수준에서 도전과 기술이 균형을 이룬 상태는 플로우 경험을 위한 필요충분조건이라는 것을 대전제로 하고 있고 연구자들은 이러한 대전제를 당연시하며 플로우를 측정하는 그들의 연구에 널리 사용하였다. 그러나 플로우 대전제에 대한 연구 대부분은 경험적이 아닌 이론적 연구에 지나지 않는다. 따라서, 본 연구는 기존의 플로우 대전제가 참인지 거짓인지를 경험적으로 검증하고자 하였다. 또한, 플로우 차원들과 여가 및 일 관련 활동들을 4 채널 모델에 범주화하고 여가와 일 관련 활동들 간에 플로우의 빈도, 강도, 그리고 차원에 있어서 차이가 있는지를 알아보고자 하였다. 마지막으로 핸드폰의 문자메세지 (TM-ESM)를 통해 기존 ESM이 응답자들의 생생한 경험을 잘 반영하고 있는 지를 조사하고자 하였다. 이러한 연구목적을 달성하기 위해 TM-ESM과 심층 인터뷰를 병행하여 양적 연구방법과 질적 연구방법을 모두 사용하여 자료를 수집하고 빈도분석과 t-test 등을 통해 수집한 자료를 분석하였다. 분석 결과, 플로우는 도전과 기술수준이 높은 수준에서 균형을 이루는 상태 뿐만 아니라 그이외의 경우 (불균형상태 포함)도 경험할 수 있는 것으로 나타났다. 따라서, 기존의 플로우 대전제는 참이 아니고 정확하지 않은 것으로 판명되었다. 즉, 높은 수준의 도전과 기술수준의 균형은 플로우 경험을 위한 필요충분조건이 아닌 충분조건이어야 한다는 것이다. 한편, 플로우의 차원들(명확한 목표, 즉각적인 피드백, 의식과 행동의 조화, 집중, 컨트롤 및 통제, 즐거움, 시간왜곡, 텔레프레전스, 자기목적적 경험)은 Csikszentmihalyi (1993, 1996)의 9가지 차원들과 대동소이한 결과를 보여주었다. 플로우은 높은 수준에서 도전과 기술이 균형을 이룰 때 여가 활동보다 일 관련 활동에서 더욱 자주 일어나는 것으로 나타난 반면 그 이외의 상태 (불균형 혹은 낮은 수준에서 균형)에서는 여가 활동에서 더 자주 플로우 경험을 하는 것으로 나타나 다른 연구들의 결과와 상반되었다. 즉, 사람들의 도전과 기술 수준에 따라 여가와 일 관련 활동의 플로우 빈도는 달라질 수 있다는 것을 의미한다. 플로우 강도는 일반적으로 사람들이 여가활동을 할 때 더욱 높을 것으로 나타났으나 낮은 수준에서 기술과 도전의 균형과 낮은 도전수준과 높은 기술수준의 불균형일 경우 반드시 통계적으로 여가활동에서의 플로우 경험이 강도가 높다고 할 수는 없는 것으로 나타났다. 또한, 여가활동의 플로우 차원 (피드백, 즐거움, 텔레프레전스)은 일 관련 활동의 차원 (명확한 목표, 자기목적적 경험)과 다르게 나타났다. TM-ESM을 통해 수집한 자료를 응답시간에 따라 15분 내외로 분류하여 두 집단 간의 플로우 경험에 차이를 분석한 결과, 통계적으로 유의한 차이를 보였다. 이는 응답시간을 통제할 수 없는 기존 ESM 자료수집의 한계를 보여주는 것이다. 본 연구는 플로우 대전제를 경험적 연구를 통해 검증함으로써 플로우 이론에 긍정적인 기여를 할 것으로 사료된다. 그리고 이러한 검증을 통해서 플로우 대전제에 대해 더욱 정확하게 이해하고 연구에 적용하여 향후 플로우 연구에 올바른 방향성을 제시하였다. 플로우 차원의 조사를 통해 사람들의 플로우 내재적 동기에 대해 이해하고 개인의 삶과 심리적 안녕을 향상시키는 데 기여할 것이다. 그러나, 본 연구는 편의표본추출법을 사용하여 결과를 일반화하는데 한계가 있다. 또한, 인터뷰를 통한 경험치의 표본 수가 많지 않아 플로우 차원을 범주화에 어려움이 있다. 조사대상자를 대학생으로 제한하였기 때문에 여가와 일 관련 활동에 있어 제한된 형태를 보여 일반적인 여가 및 일 활동으로 간주하기에는 어려움이 따른다. 이러한 한계점을 극복하기 위해서는 다른 대상과 환경을 토대로 한 후속 연구가 지속되어야 할 것이다. Since Csikszentmihalyi (1975) introduced and conceptualized flow as an optimal experience that occurs when an individual perceives an equivalence of challenge and skill during an activity, the flow concept has been studied in several disciplines to identify individual’s states of mind and subjective experiences. According to Csikszentmihalyi’s research on four channel model of flow, it is assumed that flow occurs if the levels of an individual’s skills and perceived challenges are equally high and vice versa. Namely, the equivalence of skills and challenges above average levels is the necessary and sufficient condition to experience flow. Therefore, researchers naturally accepted and prevalently used the premise of flow as proposed by Csikszentmihalyi to measure the flow. Despite many works on flow theory and models, the premise of flow has been studied conceptually or using by an inappropriate approach and analysis, rather than empirically. The purpose of the study is to verify the premise of flow empirically whether it is true or not; to identify and classify flow dimensions into the contexts of four channel model of flow respectively if flow occurs in the condition of the equivalence between challenges and skills; to categorize activities in each context of four channel model of flow; to investigate the difference between leisure and work activities in terms of frequency, level (intensity), and dimensions of flow respectively; to examine whether the existing ESM capture individual’s vivid experiences or not by using a new device like a cellular phone; to identify the advantage of the new technology application; ultimately to enhance the theory of flow. In order to achieve the purpose of the study, both a quantitative and a qualitative approaches were used for data collection and analysis including an Experience Sampling Method (ESM) using a text message survey instead Experience Sampling Form (ESF) among the self-reporting methods, and in-depth interviews. A quantitative and qualitative ESM was imperative not only to obtain vivid experience data for the empirical verification of the premise of flow by frequency analysis, but to compare the difference between leisure and work in terms of frequency, level, and dimensions of flow by frequency analysis and t-test. In addition, a qualitative in-depth interview method was used to find out flow dimensions by the levels of skill and challenge in each context of the four channel model. The results of the present study showed that the prevalently used premise of flow was not accurate and true. According to the frequency analysis that tested the premise of flow, flow could occur even when an individual’s perceived challenges and skills at high levels are not equivalent. Therefore, flow is more likely to occur in the condition of congruence between challenge and skill rather than in the rest conditions. That is, the equivalence of challenges and skills may be assumed to be the sufficient condition for the flow to occur, not the necessary and sufficient condition. In other words, the challenges and skills at high levels may not be necessarily equivalent to produce flow. From the validation perspective, it seems that the four channel model of flow may not be powerful enough to apply to the daily life setting or different types of activities. More precise conditions, subjects, and the scope for the application of the four channel model should be considered in order to make the four channel model of flow more valid. The nine flow dimensions detected in the present study are quite similar to the nine dimensions (clear goals, immediate feedback, matched skills and challenges, concentration, altered sense of time, merging of action and awareness, a sense of control, autotelic experience, loss of self-consciousness) presented by Csiksentmihalyi (1993, 1996) except two elements, telepresence and fun (enjoyment), which were used in other studies as flow constructs or dimensions. According to the findings of the present study, flow experiences are more likely to occur during work-related activities only in the condition of high challenges and high skills. However, flow experiences were more often reported during leisure activities compared to work activities in the other three conditions of high challenge and low skills; low challenge and low skills; and low challenges and high skills, which oppose results from previous studies. In terms of frequency of flow experience between leisure and work, the current results could be differentiated by a different condition, such as levels of challenges and skills in this study. Results of the t-test revealed statistically significant differences between work activities and leisure activities in terms of the level of overall flow experience at the .05 level. Therefore, the intensity of overall flow experience in leisure was higher than that in work-related activities. However, after closer examination of these results, it cannot be said that the levels of flow experience in Context Ⅲ and Ⅳ were higher in leisure compared to work activities. Findings of the interviews revealed that leisure has different flow dimensions from work. As the results, feedback, fun, and telepresence were detected as flow dimensions during leisure activities. Conversely, other flow dimensions, such as clear goals and autotelic experiences were found in work-related activities. However, concentration and time distortion appeared in both leisure and work activities as flow dimensions. According to the results of the t-test (see Table 15), there was a statistically significant difference between responding times within 15 minutes and over 15 minutes in terms of flow experiences at the .05 level (t=-3.16). The results revealed that experiences responded within 15 minutes were different from experiences responded over 15 minutes. The difference of experiences seemed to be caused by information loss over time. Therefore, the results proved that more vivid experience data could obtain through the TM-ESM by controlling the responding times of the participants. The present study can considerably contribute to the flow theory by verifying the premise of flow inductively made by a series of Csikszentmihalyi’s studies since the premise of flow is a fundamental issue for the theory of flow. Also, results may lead researchers to the right direction for future research without misunderstandings, thus providing a cornerstone to evaluate the suitability and ultimately reestablish the premise of flow more precisely. Classifying dimensions of flow in four quadrants of the four channel model of flow contributes to the understanding of the participant’s intrinsic motivations of flow experiences, depending on the levels of challenges and skills during the daily activities. Therefore, dimensions classified by the levels of challenges and skills could provide an insight into improving the quality of life and psychological well-being. In terms of the generalization issue, the results of the present study may not be applicable to other populations and areas. Random sampling method was not used due to limited resources and practical difficulties. A convenience sampling method was the most feasible method for the purpose of this study. Therefore, the results should be applied to other populations with caution. In addition, a sample size of experiences was not large enough to make inferences about flow dimensions divided by the levels of challenge and skill in each quadrant, especially the second and fourth quadrants, which were lack of dimensions caused by the insufficient experience cases. Finally, in order to overcome the issues on generalization and the sample size, continuous attempts of further research should be carried out to replicate this study with other populations and settings in various areas. Control variables (e.g. income, sex, age and education etc.) are used in order to assess or clarify the relationship between independent and dependent variables. Age and education were limitedly used as control variables to make the relationship between flow and equivalence of skill and challenge in the high levels more valid. Participants were recruited from a pool of undergraduate students in twenties from a “The modern society and leisure” class during the 2010 spring semester. However, leisure and work activities in the current study were limited to undergraduates’ ones and cannot be a representative sample of general population. For example, work activities are mostly limited to study-related activities because the sample of the current study was selected among undergraduate students. Participants may misunderstand and misinterpret the questions although the instrument was carefully designed and constructed to minimize possible errors caused by using Experience Sampling Method (ESM) with a new tool, cellular phone, by text messages. In addition, a long period and high frequency of the text message survey may have caused negative effects, for example, ignoring or skipping some required answering procedures. In order to reduce unreliable and missing data, participation was limited to volunteer undergraduate students who were interested in this study.

Flow regime and suction rate of gas-liquid swirl ejector depending on swirl intensity and back pressure

김상훈 Graduate School, Yonsei University 2021 국내박사

버블 컬럼이 장착된 아랫방향 스월 노즐 기체-액체 이젝터에 대해 후방 압력과 스월 강도에 따른 유동 양식과 흡입 유량 변화를 실험적으로 계측하고 노즐을 통한 제트의 미립화 및 이젝터 내 유동 특성 등을 바탕으로 변화 원인을 분석하였다. 제트의 스월 강도가 약할 경우 노즐 출구로부터 먼 거리까지 제트는 액주 형태로 유지되며 높은 축방향 운동량을 갖고 이 경우 외부 기체의 흡입유량은 Impinging effect에 의존한다. 반면 제트의 스월 강도가 강할 경우 접선방향 운동량과 난류의 증가로 액주가 빠르게 액적으로 미립화되고 축방향 운동량이 거리에 따라 급격히 감소하며 Mixing shock 상단에서의 가스와 액적(또는 액주) 사이의 운동량 전달이 외부 기체의 흡입유량에 큰 영향을 준다. 따라서 스월 강도에 따라 약한 스월과 강한 스월로 구분하였고 약한 스월의 경우 후방압력 증가에 따라 유동양식이 Coaxial, Mixed, Bubble flow로 천이하는 반면, 스월 강도가 증가할수록 후방압력 증가 시 유동양식이 Coaxial flow에서 Bubble flow로 바로 천이하였으며 디퓨저와 버블 컬럼에서 회전유동이 관찰되었다. 후방압력 증가에도 불구하고 유동양식 천이에 의해 외부 가스 흡입량이 증가하였으며 스월 강도가 강할수록 증가폭이 상승하였다. Bubble flow 영역에서, 강한 스월의 경우 후방압력 증가에 따른 가스 흡입유량의 감소율이 약한 스월의 경우보다 작게 나타났으며 동일한 후방압력에 대해 높은 가스 흡입유량이 나타났다. 그러나 스월 강도가 증가할수록 최소 기체 흡입유량의 감소로 유동양식 천이 압력이 증가하고 Snb가 0.14 이상인 경우 대기압 및 상대적으로 낮은 후방압력 조건에서 Bubble flow 유동양식이 나타나지 않는다. 윗방향 스월 노즐 이젝터의 실험 결과와 마찬가지로 Bubble flow 유동양식일 때 바디 스월 넘버 0.14에서 가장 높은 흡입 유량과 에너지 효율을 보였다. Bubble flow 유동 양식에서 축방향 노즐 이젝터의 실험 결과는 선행 연구의 실험적 관계식과 일치하였으나, 강한 스월을 사용한 이젝터의 실험결과는 선행연구의 관계식으로 나타낼 수 없다. 본 실험 결과를 바탕으로 강한 스월 이젝터에 대한 실험적 관계식을 결정하였으며 오일러 수의 지수 값은 -0.301이다. 이는 축방향 노즐 이젝터에서 사용 된 오일러 수의 지수 값보다 큰 값이며 효율의 증가를 의미한다. 스월 노즐을 사용한 기체-액체 이젝터의 경우 일차 유체의 종류 및 유량, 이젝터의 형상 등이 동일한 조건에서도 스월 강도와 후방 압력에 따라 유동 양식과 가스 흡입유량이 다르게 나타난다. In this study, variation of the flow regime and the gas suction rate relative to the back pressure and the swirl intensity were investigated using downward swirl nozzle gas-liquid ejector with bubble column. And, academic understanding was improved through theoretical explanation using the minimum flow rate, the rotation of the fluid in the ejector, and the spray characteristics of the jet. When the swirl intensity of the jet was weak, jet was maintained in the form of liquid column and had high axial momentum for a long distance from the nozzle outlet. In this case, the gas suction flow rate depends on the Impinging effect. On the other hand, when the swirl intensity of the jet was strong, because the tangential momentum and disturbance caused by the turbulent was increased, the liquid column was atomized rapidly and axial momentum was decreased dramatically according to axial direction. In this case, the gas suction rate was heavily influenced by the momentum transfer between the liquid jet (or droplets) and the entrained gas before the Mixing shock. Therefore, Weak swirl and Strong swirl were defined according to the swirl intensity. In the case of the Weak swirl, as the back pressure was increased, the flow regime was changed in the order of Coaxial, Mixed, and Bubble flow. But, as the swirl intensity was increased, the flow regime was changed directly from the Coaxial flow to the Bubble flow, and rotational flow was observed in the diffuser and bubble column. Despite the increase in the back pressure, the gas suction rate was increased by the transition of the flow regime, and it was increased more significantly by going to higher swirl intensity. In the Bubble flow, the Strong swirl ejector had lower decreasing rate with rising the back pressure and higher gas suction rate at the same back pressure than the Weak swirl ejector. However, the transition pressure in flow regime increased with rising swirl intensity due to the reduction of the minimum flow rate. As a result, when Snb is 0.14 or more, the Bubble flow regime did not appear under atmospheric or relatively low back pressure conditions. Similar to the previous study on the upward swirl gas-liquid ejector, the highest gas suction rate and energy efficiency were observed at an Snb of 0.14. In the Bubble flow regime, experimental results using axial nozzle (Snb= 0) in this study could be fitted by empirical correlation in the previous study. However, it was impossible for the Strong swirl ejector. Using the experiment results of this study, the new exponent of empirical correlation was determined. In the case of the Strong swirl, the exponent value of the Euler number was -0.301. This value is greater than that of axial nozzle ejector, and this means an increasing in the energy efficiency. Even under the same conditions such as kinds of primary fluid, flow rate of primary fluid, and geometry of ejector, in the case of gas-liquid ejectors using swirl nozzle, the flow regime and the gas suction rate were changed depending on the swirl intensity and the back pressure.

Flow correction to weak flow in particle tracking microrheology

이혜경 서울대학교 대학원 2015 국내석사

Particle tracking microrheology(PTM) is one of passive measurements of thermally excited probes, in which no external force is applied. PTM has been developed to directly image the position of a colloidal particle as a function of time, and allowed analysis of individual particle trajectories. Because of its advantages, PTM is a valuable tool capable of measuring the local dynamics in heterogeneous systems. However, viscoelastic properties of solutions under flow are also of great interest in various processes. Especially in drying process, because of the effect of internal flows in a drying solution, measurement of local dynamics was often limited to a qualitative description of the particle flow, or to the visualization of the velocity field. To overcome this difficulty and to measure the rheological properties, the flow must be corrected. Hence, flow correction is important in the application of PTM to drying process of polymer solutions. In this study, we tracked the motion of probing particles in one dimensional flow. In case of weak two dimensional flow, to eliminate the effect of motion due to the external flow by using noise reduction, ensemble averaged mean square displacement was investigated. We proved that the modulus after flow correction is the same with the one without flow.

Approaches in Generative Speech Models: Restoration and Enhancement in Degraded Speech Signals

양다희 한양대학교 공학대학원 2025 국내박사

In real-world communication scenarios, speech signals are often degraded by adverse factors such as background noise, reverberation, bandwidth limitations, and packet loss. These distortions hinder both human perception and the performance of speech-driven applications, including voice assistants and automatic speech recognition systems. Recent advances in deep learning have led to significant progress in speech processing, broadly categorized into discriminative and generative approaches. While discriminative models learn direct mappings from input to target signals, they often suffer from limited generalization and over-smoothed outputs. In contrast, generative models—while offering greater flexibility and perceptual quality—are often constrained by high computational cost and slow inference due to their iterative generation process. This dissertation proposes practical generative approaches that aim to improve efficiency, adaptability, or reconstruction quality across three key restoration tasks. First, for packet loss concealment (PLC), we propose Flow-PLC, a generative model based on the flow-matching framework. Flow-PLC learns a vector field that deterministically transforms a source distribution into clean waveforms, enabling fast and high-fidelity reconstruction. By avoiding iterative sampling, this approach significantly improves inference speed while maintaining high reconstruction quality, making it well-suited for real-time applications. Second, for speech enhancement under noisy conditions, we introduce a token-based generative framework that combines autoregressive modeling with flow matching to improve robustness against diverse noise conditions. We adopt a non-neural tokenization method called dMel, which discretizes Mel spectrograms while retaining both semantic and acoustic information. An autoregressive language model is trained to predict clean dMel sequences from noisy inputs, and a flow-matching-based dequantizer is employed to refine and reconstruct Mel spectrograms from the predicted tokens. This framework effectively leverages the sequence modeling capability of language models and the fine-resolution reconstruction ability of flow matching for improved enhancement performance, resulting in more robust speech enhancement under severe noise distortions. Finally, for general speech restoration involving simultaneous distortions—such as noise, reverberation, and bandwidth reduction—we propose FLOWER, a conditioning framework that provides normalizing flow-based Gaussian guidance to the generative model. This guidance is obtained from a normalizing flow model during training, but at inference time, it is sampled directly from a Gaussian distribution, enabling lightweight deployment. Consequently, FLOWER improves restoration performance without increasing model complexity. keywords: Flow matching (FM), generative model, general speech restoration (GSR), language model (LM), normalizing flow (NF), packet loss concealment (PLC), speech enhancement (SE)

Flow Management for Voice/Data Transport over UDP/TCP Based Networks

정성호 Georgia Institute of Technology 2000 해외박사

음성 및 데이터 통합망은 망과 그 운영측면에서의 비용절감, 망 자원 이용률의 증가, 대역폭이용의 융통성 등 여러 가지 장점을 제공한다. 따라서 서비스 제공업체에서는 ATM(Asynchronous Transfer Mode) 기술 또는 IP(Internet Protocol) 기술에 기초한 단일 통합 교환 및 수송망에서의 음성 및 데이터 통합 서비스 제공에 관심을 기울이고 있다. ATM과 IP 기술은 음성 및 데이터가 전송되는 방법에 있어서 서로 다른 점을 보인다. 본 논문의 목적은 ATM 및 IP 통신망에서의 음성 및 데이터 전송과 관련된 주요 이슈를 제시하고 해결방안을 제안한다. ATM은 IP 기술에 비해 서비스 품질 (Quality of Service) 제공에 있어 우수하며 음성전송을 위해 필요한 서비스 품질지원에 적합한 것으로 인식되어 왔다. ATM을 이용한 음성 전송과 관련된 주요 이슈 중의 하나는 ATM 망에서의 음성전송 능력에 대한 성능평가 및 향후 존속가능성의 평가이다. IP 기술과 관련된 주요 이슈 중의 하나로는 고품질 음성전송을 위한 서비스 품질보장을 들 수 있다. IP 통신망에서는, 음성 트래픽이 UDP(User Datagram Protocol)를 이용하여 수송된다. 이러한 UDP 기반 음성 응용의 성능에 관한 요구사항을 실현하기 위하여 망 폭주현상 발생시 음성 응용의 성능이 심하게 저하되지 않도록 적절한 대역폭을 지속적으로 제공하는 것이 필요하다. UDP flow는 대개 망 폭주현상 발생시 전송율을 줄이지 않는 특성이 있어 비응답형 flow라고 불리우기도 한다. 따라서 TCP(Transmission Control Protocol) flow와 같이 망 폭주 현상시 전송율을 줄이는 응답형 flow보다 더 많은 대역폭을 사용하게 되며, 궁극적으로는 'Internet Meltdown' 현상을 초래할 수 있다. 따라서 UDP flow에게 적절한 대역폭을 할당하는 것 뿐만 아니라 응답형 TCP flow를 보호함으로써 모든 flow에게 적절한 서비스 품질을 제공하는 것이 필요하다. 본 연구에서는, 먼저 음성전송을 위한 ATM 통신망의 성능 및 존속가능성을 평가한다. 평가항목에는 Trunking 효율, 서비스품질 변수, 망 존속가능성 변수평가가 포함된다. 아울러 IP 통신망에서의 flow관리를 위한 프레임워크를 제안하고, 제안된 프레임워크에 기초한 일련의 라우터 기반 서비스 품질 제공 메커니즘을 제안한다. 제안된 메커니즘에는 큐관리, 자원예약, 미터링 등이 포함되며, 이들은 UDP flow에 적절한 대역폭의 할당, TCP flow 보호, TCP flow 간의 대역폭 공정성 등을 제공한다.

Flows of Complex Fluids in Porous Media with Applications to Composites Manufacturing

Liu, Hai Long 경상대학교 대학원 2013 국내박사

A wide variety of environmental and technological processes involve problems of a fluid flow through porous media, such as spreading of contaminants in soil, the enhanced oil recovery in petroleum industry, the trickle bed reactor for catalyzed reaction, various filtration processes in chemical engineering and composites manufacturing. The thesis covers the most of all numerical, theoretical and experimental aspects on both Newtonian and non-Newtonian fluid flows through porous media. In the first topic, we present a new three-dimensional finite element technique to solve flows in a representative porous volume with fibrous microstructures. Through the extensive numerical simulations for various fiber and fabric architectures, we investigate the relationship between the permeability and fiber architectures in order to establish a reasonable approximation method in estimating the permeability of such complex 3D architectures. Specifically we discuss the applicability and the limitation of the macroscopic permeability averaging rule for those purposes, using the permeability of simple structural building blocks. In second topic, we studied flow behaviors of a non-Newtonian fluid in spherical microstructures by a direct numerical simulation. A shear-thinning (power-law) fluid through both regular and randomly packed spheres has been numerically investigated in a representative unit cell with the tri-periodic boundary condition. The flow mobility of regular packing structures, including simple cubic (SC), body-centered cubic (BCC), face-centered cubic (FCC), as well as randomly packed spheres, has been investigated quantitatively by considering the amount of shear-thinning, the pressure gradient and the porosity as parameters. Furthermore, the mechanism leading to the main flow path in a highly shear-thinning fluid through randomly packed spheres has been discussed. In the third topic, we derive the analytical solution of a power-law fluid through periodic arrays of square and hexagonal packing cylinders by the lubrication theory to characterize the mobility dependence on the pressure drop, the amount of shear-thinning and the fiber volume fraction. Based on the theoretical modeling and direct numerical simulations, we propose a simple geometrical scaling rule for the mobility by separating the fluid rheological contribution from the rheological contribution. This simple scaling rule will be employed to approximate the dependence of mobility on the pressure gradient for various packing structures, which facilitates simple and reliable prediction of the flow mobility for power-law fluids in porous media. In the fourth topic, we present numerical simulations for flow behaviors of viscoelastic fluid passing through the porous media. We employ the DEVSS/DG (Discrete Elastic-Viscous Splitting / Discontinuous Galerkin) finite element scheme combined with the mortar-element method for the bi-periodic boundary condition and the fictitious domain method for a proper representation of fibers in a fluid. The matrix logarithm has been incorporated to achieve stable solutions at high Weissenberg number. By employing an Oldroyd-B and Leonov models as constitutive equations, we discuss effects of elasticity on the flow resistance in typical porous microstructures of viscoelastic flows: square-, hexagonal-, as well as random-packing unidirectional fibers. The increase of flow resistance has been observed for the first time, as far as authors know, as the Weissenberg number exceeds a critical value by numerical simulation. The significant growth of the flow resistance has turned out to be related with the extensional stretching of polymers in viscoelastic flows in porous media and the energy dissipation of the viscoelastic flow in porous media has also been analyzed. The last topic is about the experimental work on the flow visualization for the the VARTM process. In this experimental study, an ultrasound visualization system has been set up for in-situ monitoring of the resin flow impregnating through opaque carbon fiber reinforcements. The flow front advancement through the carbon fabrics covered by a bagging film can be identified and visualized by the high frequency B (brightness) mode ultrasound imaging technique. The resin advancement in the opaque carbon fabrics has turned out to form a non-uniform plug flow along the pressure gradient direction and the potential void formation can be observed from the mesoscopic resin flow behavior: the inter-tow regions have been preferentially filled by the resin fluid and the fiber tow region behaves as a sink that probably remain as a void defect. The local unsaturated transient velocity of the resin flow has been also evaluated, which is particularly important in understanding saturation behaviors in dual scale fabrics and is hardly measurable by other means. In spite of the hardware limitations on the resolution, the proposed ultrasound visualization system can provide a less expensive and portable visualization tool to understand the microstructure of opaque reinforcements and monitor the resin flow behavior during the VARTM process in the industrial composite manufacturing environment.

Flow 유형이 휴대폰 서비스 이용에 미치는 영향에 관한 연구

이효선 전북대학교 경영대학원 2008 국내석사

Mobile phones are the necessity of the people in present. And the mobile phones are still evolving and evolving. Now the new tech competition with various functions and capacities to push out the old ones to the outside of the market and create new demands is continuously going. In almost every field, the role of mobile phone is increasing in accordance of modern people's demands, who lives their lives very fast and busily, and the current market circumstances. Based on the model proposed by Novak & Hoffman(2000) theoretically, this study was segmented the mobile phone users by flow types, found the users' features by types and conducted focused on the differences for using the mobile phone by segmented markets. In order to understand the differences of usage of the mobile phone users by flow types, I tried to find whether the differences of the features of population statics, the motivation and the degree of using the mobile phone, the recognition and satisfaction of the quality, the considering factor of selecting telecom companies, etc. were according to types. According to the result of this study, the Flow type of the mobile phone users was segmented into 5 types such as Flow, controlled, bored, indifference and fear, and as the result of the population statics, in Flow or controlled type, the proportion of the female students were relatively high. This means that the female student had more tendency of feeling Flow in using the mobile phone than the male students. Then, in order to find which differences were in using the mobile phone by flow types, I analyzed various variants. First, as the result of the cross analysis in order to find whether there were differences in using the mobile phone by flow types except the telephone call with joining telecom companies, there were no notable differences of these variants by flow types. Moreover, as the result of dispersion analysis in order to find whether there were the differences of the motivation of using the mobile phone by flow types, it was analyzed that there were notable differences between groups on most motivation of considered 15 motivations of utilization. It represented that the motivations of using the mobile phone were different by flow types. Then, as the result of the dispersion analysis in order to find where there were the differences of the how much concern about the considering factors when people selected telecom companies by flow types, there were no differences in 'the initial joining cost' and 'the company image' and it was analyzed that there were notable differences between groups with 7 considering factors such as 'the monthly utilization fee' or 'the company name recognition', etc. This means that the degree of consider the various factors was different for each flow type when they select the telecom companies. In order to research which differences were in the behavior of using the mobile phone, as the dispersion analysis, it was found that there were differences of 'the monthly average utilization fee of the mobile phone', 'the monthly average utilization time of the mobile phone' and 'the monthly average utilization numbers of the mobile phone'. Especially the degree of utilization of the controlled group was the highest and the one of the Flow group was quite high. However, in 'the monthly average additional service utilization fee', there was no notable differences statistically. In addition, as the result of researching the level of the evaluation of service quality on the mobile service by flow types and the level of customer satisfaction, there were no differences between types on the evaluation of service quality, but there were notable differences on customer satisfaction. In the level of customer satisfaction, the flow type was the highest and the second highest one was the controlled group. With referring this study, marketers need to understand the features by flow types and plan the marketing strategies for each type with segmenting the market.

Sustainable Electrical Energy Harvesting via Electrified Flow

이승환 울산과학기술원 Ulsan National Institute of Science and Te 2026 국내박사