포항공대 분자 모델링 시스템 개발 및 생화학 분자 내에서의 Zn(II)의 배위에 관한 이론적 연구

윤장우 포항공과대학 1992 국내석사

포항공대 분자그래픽스 시스템 개발 및 물분자 집합체에 관한 이론적 연구

김성욱 포항공과대학 1990 국내석사

Continuous Microfluidic Process for Electroporation and Extraction of Microalgal Cells

김보람 포항공과대학교 일반대학원 2020 국내박사

본 논문은 미세조류의 전기천공과 추출을 위한 연속 미세유체공정의 개발에 관한 연구이다. 본 연구에서는 미세조류의 바이오매스와 바이오리파이너리 경제성 확보를 위해 업스트림(upstream)과 다운스트림(downstream) 관점에서 공정 개선을 위한 미세유체시스템을 개발하였다. 제 1장에서는 미세유체 기반의 미세조류 연구에 대해서 전반적으로 다루었으며, 미세조류의 바이오매스 및 바이오리파이너리 공정을 업스트림과 다운스트림으로 분류하여 경제성확보를 위해 개발된 다양한 미세유체 시스템에 대해서 설명하였다. 또한 기존 보고된 연구 결과를 기반으로 본 학위 논문에서 다루고자 하는 연구의 전반적인 개요를 제시 하였다. 제 2 장에서는 업스트림관점에서 우량 균주 제작을 위한 새로운 도구의 필요성을 인지하고 미세유체기반의 형질전환 장치를 개발하였다. 동물세포와 달리 두꺼운 세포벽으로 형질전환 효율이 낮은 미세조류는 외래유전자의 저항성이 높아 유전 또는 대사 공학적 균주의 개량에 한계가 있다. 이와 같은 한계점을 극복하고자 전극 사이의 간격을 줄이고, 세포가 직접적으로 전극면과 닿지 않게 함으로써 세포 생존율이 높은 폴리이마이드 필름 기반의 전기 천공 장치를 개발하였다. 결과적으로 상용화 제품과 비교하여 약 50배 긴 지속시간에서 우수한 세포 생존율과 개선된 전달 효율을 검증하였다. 또한 ptCrCFP 외래 유전자를 주입하여 형질전환체의 제작을 형광단백질 발현과 PCR로 검증함으로써 실제 미세조류의 유전공학 장치로서의 응용 가능성을 확인하였다. 게다가 장시간 장치 사용에 따른 전극면 손상도와 전달 효율을 측정하여 3시간 동안 장치를 사용하였을 때 얻을 수 있는 콜로니의 수를 이론적으로 계산함으로써 형질전환의 한계를 확률적으로 개선 할 수 있는 연속미세유체 장치의 활용 가능성을 제시하였다. 제 3장에서는 미세조류 바이오매스와 바이오리파이너리 다운스트림 공정에서 갖는 에너지 대비 낮은 추출 효율 개선을 위한 연속흐름방식의 세포의 분리 미세유체장치를 개발하였다. 항산화 물질로 잘 알려진 아스타잔틴은 의학, 기능성식품, 사료, 화장품 등 응용 가치가 높아 전세계적으로 시장이 커지고 있는 추세이다. 특히, 자연유래 아스타잔틴은 합성 아스타잔틴 대비 안정성과 항산화활성도가 우수하기 때문에 관심도가 높아지고 있다. 그러나 합성 아스타잔틴 대비 낮은 경제성은 자연유래 아스타잔틴 시장의 한계이므로 생산성 증대 또는 추출 효율 증대와 같은 다양한 접근법을 통해 극복해야 할 문제이다. 본 연구에서는 세포 크기별 분리 공정 개발을 통해 아스타잔틴 함유량이 가장 높다고 알려진 Haematococcus.pluvialis의 추출률을 개선하고자 한다. 나선형 미세유체 채널 내부의 유동과 이차관성으로 발생하는 Dean 유동은 크기에 따라 세포가 받는 전체 힘이 달라 측면 이동을 유도한다. 결과적으로 출구의 위치에 따라 세포의 크기가 다르게 분리됨을 검증하였으며, 추가적으로 분리된 세포들은 유체역학 유동에 의해 발생하는 전단응력에 의해 아스타잔틴이 쉽게 추출됨을 검증하였다. 특히, 지용성 물질인 아스타잔틴은 구조적 특성으로 물상에 노출 되었을 때 액적을 형성하며 다른 물질과 혼합되지 않기 때문에 정제 공정에 소모되는 비용을 절감할 수 있는 장점이 있다. 미세유체 기반의 미세조류 추출 연속 공정 시스템은 큰 세포만을 선별하여 추출 효율을 증진시키고 작은 세포는 추가 배양 공정을 통해 추가 아스타잔틴 축적 후 추출하는 공정으로 응용이 가능하며 배양에 소모되는 시간을 감소함으로써 전반적인 다운스트림 공정을 개선 할 수 있는 혁신적인 제안이 될 것으로 기대한다. 제 4장에서는 본 학위 논문에서 수행한 연구에 대해서 요약하고 개발된 장치의 응용 가능성과 향후 전망에 대해서 제시하였다. In this thesis, a novel continuous microfluidic platform capable of precise control of hydrodynamic flow was developed and applied to upstream and downstream of the biomass and biorefinery in two distinct studies. Chapter 1 covers the study of microalgae based on microfluids, and describes various microfluidic systems developed to improve the process by classifying the biomass and biorefinery processes of microalgae into upstream and downstream. Also, based on the previously reported research results, an overall overview of the research in this dissertation was presented. In Chapter 2, we recognized the need for a new tool for producing a superior strain from an upstream perspective and developed a microfluidic transformation device. Unlike animal cells, microalgae that have low transformation efficiency due to thick cell walls and high resistance to foreign genes. Thus have limitations in improving genetic or metabolic engineering strains. In order to overcome such limitations, an electroporation device with high cell viability based on a polyimide film was developed by reducing the gap between electrodes and preventing the cells from directly contacting the electrode surface. As a result, excellent cell viability and improved delivery efficiency were verified at a duration approximately 50 times longer than that of commercial products. In addition, the production of transformants by ptCrCFP foreign genes was verified by fluorescence protein expression and PCR to confirm the applicability of actual microalgae as a genetic engineering device. The result of damage on surface of electrodes demonstrated that possibility of reuseable device. In Chapter 3, we developed a microfluidic device for separating cells in a continuous flow method to improve the extraction efficiency with low consume energy in the microalgae biomass and biorefinery downstream processes. Astaxanthin, well known as an antioxidant, has high application value in medicine, functional food, feed and cosmetics, and the market is growing worldwide. In particular, natural astaxanthin is of increasing interest due to its superior stability and antioxidant activity compared to synthetic astaxanthin. However, lower economic efficiency than synthetic astaxanthin is a limitation of the naturally derived astaxanthin market, which is a problem that must be addressed through various approaches such as increased productivity or increased extraction efficiency. In this study, we aim to improve the extraction rate of Haematococcus.pluvialis, which is known to have the highest astaxanthin content, through the development of a separation process by cell size. The flow inside the micro channel and the secondary inertial Dean flow generated by spiral channel induces lateral movement due to the different total force received by the cells depending on the size. As a result, cells was separated differently according to the location of the outlet, and additionally, the separated cells easily extracted astaxanthin by shear stress generated by hydrodynamic flow. The developed continuous process-based microfluidic system includes a sequential process strategy to improve overall downstream processes. The separated large cells flow directly to the extraction process, and the small cells can flow to the additional culture process as a supplementary step and then to the final extraction step. As such, the passive-based separation system is expected to be applicable to a large-scale system because of the high throughput compared to an active separation system based on optical analysis of a single cell. The innovative platform and results presented in this thesis provide new tool to overcome limitation of transformation in the challenging area of upstream, and provide insight that can improve downstream.

A Study on the Changing Combustion system of Coke Oven for the Reducing NOx

박성덕 포항공과대학교 철강대학원 2020 국내석사

포항제철소의 코크스오븐은 총 5기가 있으며 ‘73년 1기 건설을 시작으로 2~3년 간격으로 ‘83년 5B 코크스오븐이 건설되었으며, ‘07년 고로 증산 등 생산량 변동에 따른 수급 충족을 위해 5A 설비가 추가로 건설하게 되었다. 최근 건설된 5A 코크스오븐 외 나머지 설비는 건설당시 환경법 기준을 근거로 설계되어 연소 구조가 연소실 하부 1곳에서만 불꽃이 발생하는 1단연소 구조로 구성이 되어있다. 최근 환경법 강화에 따라 과거에는 적용되지 않았던 코크스오븐 설비의 NOx 농도제약이 생기게 되면서 이전에 건설된 코크스오븐이 환경법에 저촉될 Risk가 높아졌다. 코크스오븐 설비는 Silica 연와로 구성되어 있으며, Silica 연와 특성상 온도변화에 따라물리적 강도변화가 크게 발생하므로 조업 중 온도강하를 통한 설비구조 변경이 불가능하다. 통상적으로 코크스오븐은 최초 화입이후 40~50년간 지속적으로 열간상태를 유지하면서 조업 하고, 회사 경영여건에 따라 신설 또는 Shut Down을 결정하게 된다. 따라서 이미 40년 이상 사용하여 수명이 다해가는 노후 코크스오븐에 환경법 기준 충족을 위한 가스처리 설비의 신설 투자는 경제적이지 못하다 이에, 본 연구는 코크스오븐 연소시스템 및 NOx 발생 Mechanism 검토를 통해 설비와 운전조건의 최소한의 변동으로 강화되는 환경법 기준을 충족하고 안정적 조업환경을 구성하도록 하는 것에 있다. There are a total of five coke ovens at Pohang Steel Mill, and 5B coke ovens were built every two to three years, starting with the construction of the first one in 1973, and additional 5A facilities were built to meet supply and demand due to changes in production, including increase of furnace production in 2007. Except for the recently constructed 5A coke oven, the remaining facilities are designed based on environmental law standards at the time of construction, and the combustion structure consists of a single combustion structure in which flames occur only in one lower The recent tightening of environmental laws has created NOx concentration constraints on coke oven facilities that have not been applied in the past, raising the risk that previously constructed coke ovens will violate environmental laws. Coke oven facilities consist of silica refractories, and due to the nature of silica kites, physical strength changes occur greatly due to temperature changes, so it is impossible to change the structure of the facility through temperature drop during operation. Usually, coke ovens operate continuously in hot condition for 40 to 50 years after initial loading, and decide on new or shut down according to the company's management conditions. Therefore, it is not economical to invest in new gas treatment facilities to meet environmental law standards in old coke ovens that have been used for more than 40 years. Thus, this study is designed to ensure that the Coke-Oven combustion system and the NOx-generated Mechanism review meet the environmental law standards strengthened by minimal variation in the facility and operating conditions and form a stable working environment.

Nowcasting Korean GDP growth using Machine Learning with Economic Policy Uncertainty feature

정승민 포항공과대학교 융합대학원 2022 국내석사

GDP growth is an indicator of a country's economic situation and is a crucial factor in financial decisions. Nevertheless, since it has a problem of being announced lately, 'Nowcasting', the prediction of GDP growth at present, is being treated as an essential issue. Due to the recent increase in uncertainty, studies to increase the accuracy of Nowcasting are primarily divided into two directions. One is to reflect uncertainty as a variable, and the other direction is to use ML models as predictive models. However, there has yet to be an attempt to incorporate both approaches. Therefore, this study aims to integrate both approaches to generate a prediction model for the GDP of Korea. The proposed method first extracts common factors through the Dynamic Factor Model to reduce the dimensions of 83 economic indicators affecting GDP growth. Then, the Economic Policy Uncertainty value, an indicator of uncertainty, is combined with the reduced factors, and they are used as input features of prediction models. Finally, several machine learning models are used to predict GDPs. To validate the proposed approach, we conduct experiments with Korean GDP-related data. In the experiment, we construct two data sets with and without the Economic Policy Uncertainty value to explore the impact of the uncertainty. Random Forest, Gradient Boost, and XGBoost are used as ML-based prediction models, while OLS regression is used as a conventional prediction model. The experimental result shows that including the EPU feature provides higher prediction accuracies for all four models. In addition, the performances of the ML models are more elevated than that of OSL regression.

Impaired Mammalian Epimorphic Regeneration in the Absence of Adaptive Immunity

이지은 포항공과대학교 융합대학원 2025 국내석사

Regeneration refers to the restoration of damaged tissues or organs to their original structure and function. While mammals exhibit limited regenerative capabilities, digit tip regeneration represents a rare example of epimorphic regeneration, characterized by the formation of a blastema, a mass of undifferentiated cells. This study investigates the role of adaptive immunity in mammalian digit tip regeneration using Rag1-KO mice, which lack functional adaptive immunity. Immunostaining revealed the presence of T cells during digit tip regeneration, suggesting their involvement in the process. In Rag1-KO mice, the absence of adaptive immunity led to reduced nail area and shortened digit tip length. These external changes were accompanied by internal structural deficits, including a reduced mesenchymal area and impaired bone regeneration, characterized by diminished volume and weakness. Single-cell RNA sequencing further demonstrated decreased mesenchymal cell proliferation and late-stage osteogenic differentiation in Rag1-KO mice, confirming that the absence of adaptive immunity influences proliferation and differentiation of blastema. Furthermore, the absence of adaptive immunity resulted in increased infiltration of innate immunity, including macrophages and neutrophils, disrupting the blastema microenvironment. These findings suggest that adaptive immunity indirectly regulates proliferation and differentiation of blastema by modulating innate immunity. This study provides important insights into the role of adaptive immunity in mammalian epimorphic regeneration and offers potential directions for enhancing regenerative outcomes in mammals.

감성분석 기반 7 가지 감정을 반영한 재구매 예측 및 고객 구매행동 분석

박해균 포항공과대학교 융합대학원 2024 국내석사

최근 들어 온라인 플랫폼의 발달로 인해 디지털 커머스(digital commerce)가 크게 성장하고 있다. 이에 따라, 소비자들이 주로 온라인 플랫폼을 통한 구매의사결정을 하면서 온라인 구매리뷰의 중요성이 기업과 소비자 측면에서 대두되고 있는 추세에 있다. 이와 관련하여, 기존의 CSA(Customer Sentimental Analysis)와 재구매예측 선행연구들이 많이 진행되었으나, 각각 긍부정의 단순한 2가지 감정에 대한 분석과 텍스트 데이터나 그 외의 정형 데이터 중 일부만을 활용하는 한계점을 보여왔다. 이를 보완하고자 본 연구에서는 Phase 1의 7가지 감정분류 모델에서 7가지 감정인 분노(Anger), 혐오(Disgust), 두려움(Fear), 기쁨(Happiness), 슬픔(Sadness), 놀람(Surprise), 중립(Neutrality)을 분류하도록 학습하였고 93%의 정확도를 보였다. 이어진 Phase 2의 재구매예측 모델에서는 리뷰 데이터에서 Phase 1을 통해 추출한 7가지 감정을 반영하였고, 감정을 반영하지 않은 모델의 정확도인 89.04%에 비해, 90.25%로 상대적으로 높은 성능의 모델을 도출하였다. 이후 사후분석을 통해, 국내의 가장 대표적인 O2O(Online to Offline) 플랫폼인 네이버 쇼핑의 2024년 실제 데이터에 RFM Framework 기반의 K-Means 클러스터링을 통해 분석하였다. 결과적으로 본 연구는 고객군을 각 감정 점수를 반영하여 세분화함으로써 CRM(Customer Relationship Management) 측면에서 비재구매 고객을 관리하기 위해서 극단적인 부정적 감정에 해당하는 분노, 혐오가 발생하지 않도록 VoC(Voice of Customer)에 대한 처리나, 제품 구매후 서비스(A/S)를 통해 관리하는 것이 중요하다는 점을 제시하였다. The development of online platforms in recent years has catalyzed significant growth in digital commerce. Consequently, the importance of online purchase reviews has become prominent for businesses and consumers alike. However, previous research in Customer Sentiment Analysis (CSA) and Repurchase Prediction has shown limitations, primarily focusing on binary sentiment analysis or utilizing only some textual or structured data. To address these limitations, this study undertakes two phases. In Phase 1, a 7 sentiment classification model is trained to classify seven emotions—Anger, Disgust, Fear, Happiness, Sadness, Surprise and Neutrality—with an accuracy of 93%. Subsequently, in Phase 2, a repurchase prediction model is developed reflecting the seven emotions derived from Phase 1. This model achieves a notable accuracy of 90.25%, outperforming the model that does not incorporate emotions with an accuracy of 89.04%. Following this, in the subsequent post-analysis, RFM Framework-based K-Means clustering is employed on real data from Naver Shopping, a prominent O2O platform in South Korea, from the year 2024. In conclusion, this study highlights the critical role of managing extreme negative emotions such as Anger and Disgust in CRM (Customer Relationship Management) to effectively address non-repurchasing customers. Strategies can include proactive handling of customer dissatisfaction through Voice of Customer (VoC) initiatives and enhancing post-purchase services to optimize product quality.

Bank Customer Behavior and Time Prediction using Process Mining

이은채 포항공과대학교 융합대학원 2023 국내석사

Utilizing data to predict customers' future behavior is an effective marketing strategy, particularly in the financial sector. The digital transformation in this industry has led to a shift towards digital operations and the availability of various interaction channels. With a wealth of customer data from these channels, data mining efforts are actively underway to identify potential customers and predict outcomes such as product subscriptions, customer churn, and loan approvals. However, the current predictive models face limitations in accurately characterizing customer behavior patterns as a single variable. Understanding customers' actions is crucial for predicting their future behavior. Therefore, this study aims to demonstrate that process mining can effectively incorporate customers' past behavior, leading to improved overall performance (accuracy, precision, recall, and F1-score) of the predictive model. The primary objective of this study is to create more informative and critical features for the machine learning model using process mining techniques. The outcome of this machine learning classification model is binary, which may not intuitively align with setting up marketing strategies that result in more satisfied and loyal customers. Hence, another key goal is to predict not only the customer's future actions but also the expected time until their next action. To achieve this, a method of predictive business process monitoring was employed to determine the time of the next event based on the customer's event log of actions. Case studies were conducted using bank customer data to evaluate the suggested prediction methods. The results indicate that predictions based on individuals' action logs can assist banks in providing better service and support to customers while enabling the implementation of a targeted customer management strategy.

Human Action Recognition and Classification in Extreme Conditions

김정윤 포항공과대학교 융합대학원 2025 국내석사

최근 이미지 분류 모델은 필요한 각자의 상황에 따라 우수한 성능과 정확도를 자랑한다. 주목할 부분은 대부분의 분류 모델에서 사용하는 데이터셋은 대부분 높은 화질의 이미지 및 영상 자료이다. 하지만 군, 재난, 구조, 경계와 같이 특수한 상황에서의 Vision 데이터는 상대적으로 낮은 화질을 가지고 있다. 실제로 해상도, 밝기, 채도가 훌륭한 일반적인 데이터셋에서 80% 이상의 정확도를 보여주는 모델을 대상으로 인위적인 조정을 통해 데이터셋의 수준을 낮출 경우 20% 이하로 정확도가 크게 떨어졌음을 확인했다. 본 연구는 이러한 극단적인, 특수한 상황에서의 Human action recognition 성능을 보장하기 위한 방법을 제안한다. 먼저 기존 데이터셋의 해상도, 밝기, 채도를 인위적으로 조정하여 주어진 특수한 상황에서 획득하는 비전 데이터와 유사한 특성을 갖도록 한다. 이미지 분류 모델은 오리지널 데이터셋과 함께 해당 특성을 흡수한 데이터셋을 함께 학습함으로써 최적의 가중치를 얻고 주어진 상황에서도 성능을 보장하는 방법을 고안했다. 실제로 군 상황을 예로 들어 해상도, 밝기, 채도가 조정된 데이터셋을 함께 학습한 결과, 오리지널 데이터셋으로 학습 및 평가했을 때와 유사한 정확도로 회복하는 것을 확인했다. 이를 통해 특수한 상황에서 제한된 데이터의 부족을 보완하고, 전혀 새로운 도메인이나 사전 학습된 모델이 적합하지 않은 데이터셋에 대해서도 성능을 보장할 수 있다. 이는 군 뿐만 아니라 악천후 속의 재난 상황이나 국가 중요시설 감시, 경계 등의 여러 특수한 상황에서 유용하다. Recent advancements in image classification models have demonstrated remarkable performance and accuracy across diverse applications. These models are predominantly trained on high-quality image and video datasets, which serve as benchmarks for their capabilities. However, visual data captured in specialized scenarios, such as military operations, disaster response, or border surveillance, often suffers from lower quality due to challenging environmental conditions. Notably, our analysis revealed a significant drop in performance when the dataset quality was intentionally degraded—models that achieved over 80% accuracy on standard datasets with excellent resolution, brightness, and saturation saw their accuracy plummet to below 20% under these adjustments. This study proposes a novel approach to ensure reliable human action recognition performance in extreme and specialized scenarios. To achieve this, we artificially adjust the resolution, brightness, and saturation of existing datasets to emulate the characteristics of visual data captured in such challenging environments. By training image classification models on both the original dataset and the modified dataset, we derive optimal weights that maintain robust performance even under these adverse conditions. Using a military scenario as a case study, we demonstrate that incorporating the adjusted datasets during training restores accuracy to levels comparable to those achieved with the original dataset under standard conditions. This approach addresses the limitations posed by insufficient data in specialized scenarios, ensuring reliable performance even when datasets deviate significantly from pre-trained model domains or when entirely new domains are introduced. The proposed method proves applicable not only to military operations but also to other extreme situations such as disaster response under adverse weather conditions, surveillance of critical national infrastructure, and border security.

전산모사 및 I-V 모델링을 이용한 20-nm 노드 Ge2Sb2Te5 상변화 메모리 특성 분석

장현동 포항공과대학교 일반대학원 2020 국내석사

Phase-change memory (PCM) using chalcogenide materials has been spotlighted as non-volatile memory, which can fill the gap of memory hierarchy between DRAM and NAND Flash. However, as research on scaling down PCM devices has been actively conducted, problems such as resistance drift, cell-to-cell disturbance, undesired programming, and high RESET current have emerged. In particular, the chalcogenide material in the amorphous state has caused side effects, thereby raising the need for research for in nm-class PCM devices. Therefore, in this thesis, electrical and thermal performances of the PCM are characterized and discussed using technology computer-aided design (TCAD) and integrated circuit characterization and analysis program (IC-CAP) tools. Electrical and thermal performances of 20-nm node PCM described the bandgap model are extensively analysed according to physical parameters and geometry using fully-calibrated TCAD simulation. Increasing the maximal crystallization rate (r0) and decreasing the activation energy (Eact) reduces SET resistance and SET programming current, thus resistance ratio can be increased and consumes less power. This result suggests that Eact is more sensitive to electrical performance than r0. SET and RESET current decreases as impact ionization factor (II) increases. Also, programming current decreases and heat efficiency increases as thermal boundary resistance (TBR) thermal conductivity decreases and TBR metal resistivity increases. Decreasing the cell height reduces SET resistance, so increases read latency. However, current which produces the same Joule heating effect increases, thus increasing power consumption. Therefore, read latency and programming current are in a trade-off relationship when the cell height varies. Finally, these parameters can vary the threshold voltage, thus when designing the PCM cell, these parameters must be considered to meet the desired specifications. Threshold switching and snap-back mechanism is explained by the non-equilibrium carrier distribution and non-uniformity of electric field along the amorphous layer. Then, the appropriate barrier lowering change model that describe carrier transport in the subthreshold region is selected by comparing two models. Therefore, the subthreshold region, threshold switching, negative differential resistance, and ON region are implemented using analytical model of PCM. As a result, this model can be applied to circuit simulations of PCM devices with current equations based on physical computation.