A Process Mining Based Approach to Complex Manufacturing Process Flow Analysis: A Case Study

SookYoung Son Graduate school of UNIST 2014 국내석사

With recent advances in IT infrastructure in manufacturing environments, a large amount of manufacturing data are collected and stored in a database at various stages of production. These data may include valuable information for manufacturing companies to improve their manufacturing processes. The method of manufacturing data analysis is crucial for understanding the manufacturing data. However, traditional manufacturing data analysis methods such as data mining, simulation, etc. have limitations for this purpose since those are difficult to provide overall process-level information. Therefore, in this thesis, a process mining based approach for analyzing complex manufacturing processes is proposed. Process mining is a useful tool for process-related knowledge acquisition since it enables users to derive not only manufacturing process models, but also several performance measures related to processes, resources, and tasks. This thesis suggests a framework for the manufacturing process analysis. To do this, it applies process mining techniques to perform four types of analysis, which are visualization of production flows, machine-to-machine inter-relationship analysis, machine utilization, and monitoring & diagnosis of task performance regarding yield rate and lead time. Furthermore, a case study is conducted to support the proposed framework with an event log of an electronic components manufacturing process.

Process Mining-driven Performance Analysis in Manufacturing Process: Cost and Quality Perspective

Tu Thi Bich Hong Graduate School of UNIST 2016 국내석사

The dynamics of globalization and high expectation of customers make manufacturing enterprise move towards three primarily competitive factors, namely, time, cost, and quality. The desire for continuous performance improvement of manufacturing processes is as old as manufacturing itself. However, the latest industrial revolution Industry 4.0 and the digital revolution have opened up new avenues, intertwined information systems and the operation processes. As a result, enterprises face a challenge in extracting value from a massive amount of events recorded by today‟s information systems. Process mining is well recognized as a valuable tool for observing and diagnosing inefficiencies in business processes based on event data. It turns out that process mining is a viable solution to this challenge. Nevertheless, significantly less attention has been paid on investigating cost and quality perspective in process mining. In these respects, this thesis suggests a framework for performance analysis in manufacturing processes based on process mining. The proposed approach focuse on cost and quality perspective. Specifically, the contributions of this thesis are in four-fold (i) to suggest a method to extend event log of manufacturing process with manufacturing information, i.e. cost, quality; (ii) to analyze manufacturing information, i.e. cost, quality with process model; (iii) to utilize various existing process mining techniques and develop new approaches to analyze and predict manufacturing cost; (iv) and to enable quality report in manufacturing process.

Improvement of manufacturing process for the new third generation recombinant factor VIII

백상훈 Graduate School, Korea University 2013 국내박사

그린진에프Ⓡ (beroctocog alfa)는 새로이 개발된 재조합 혈액응고 제8인자 (rFVIII)로서 A형 혈우병 치료에 사용되는 제품이다. 그린진에프는 B-domain이 제거된 형태의 rFVIII로 매우 균질성(homogeneity)있는 단백질로 생산되고, heavy chain 90 kDa 및 light chain 80 kDa 으로 구성되어 있다. 이러한 제품의 균질성은 quality control, characterization, 그리고 각 lot 별 안정적인 생산이 가능하여 GMP 규정에 맞는 제품 생산과 품질관리에 유리한 위치를 가진다. 이 제품은 세포배양공정 및 정제공정, 최종원액에 albumin 같은 동물 및 인간 유래 원료가 사용되지 않는 생산공정을 이용한 제 3 세대 rFVIII 제품이며, 이는 기존에 개발된 그린진 제품(2 세대 제품)에서 최종원액의 안전성 우려를 불식하기 위하여 albumin을 제거하고 아미노산을 안정제로 사용하는 형태로 개발되었다. 또한, 제품의 안전성을 극대화 시키기 위하여 그린진에프 공정에는 바이러스 불화화 및 바이러스 제거를 위한 S/D 공정 및 20 nm의 nanofiltration 공정 (기존 출시된 rFVIII 제품 중 가장 작은 pore size)이 추가되었다. 이러한 공정으로 새로 개발된 그린진에프는 원료 및 공정 중 오염 가능성을 경쟁 제품보다 기술적으로 가장 안전성이 높은 제품으로 개발하게 되었다. 본 연구에서는 그린진에서 그린진에프로 공정 개선을 위한 변경 및 upgrade 이후 국내 및 해외 GMP, 품목허가 가이드라인을 충족하는 연구 개발 과정을 나타내었다. 이를 위해 세포주에서부터 최종 GMP 제조공정까지의 개발과정 요약 및 공정 개선, 변경 사항을 연구하였으며, 공정 개선을 위해 추가된 바이러스 불화화/제거 공정 및 동결건조 후 rFVIII의 안정화를 위한 albumin-free 최종원액 조성 개발 연구결과를 나타내었다. 그린진 및 그린진에프의 동등성 확인을 위한 이화학적 특성, 생리화학적 특성 및 비임상 효력 시험 결과 비교를 통하여 제품 안전성 향상을 위한 공정 개선이 단백질 품질에 영향을 주는지의 여부를 확인하였다. 이러한 연구는 바이오의약품의 공정변경 시 필수적으로 요구되는 것이며, 따라서 본 연구에서는 동등성 확인 시험들이 ICH, FDA 및 EMA 가이드라인에 충족되도록 수행하였다. 이후 그린진에프 제조공정, 최종원액, 그리고 최종 분병 제품에 대한 process validation을 수행하여 공정의 robustness를 확인하였으며, 그린진에프 단백질 특성에 대한 시험을 수행하였다. 개발된 그린진에프의 공정은 높은 품질 및 안정성을 가지는 rFVIII을 생산하였으며, 모든 공정은 허가 가이드라인에 충족되도록 개발되었다. 바이러스 불화화/제거 공정의 검증 시험을 통하여 10~19 log reduction의 높은 바이러스 제거율을 확인하였고, albumin 안정제 대신 사용된 3개의 아미노산 조합의 안정제는 albumin 이나 기타 알려진 안정제들과 동등 이상의 결과를 보였고, 단백질 특성에 변화 없이 동결건조 전후 역가가 일정하게 유지되는 높은 안정성을 확인하였다. 동등성 시험에서 이화학적 특성, 생리화학적 특성, 비임상 효력/독성 시험 결과를 비교할 때 모두 동등함을 확인하였다. 이는 rFVIII의 단백질 특성 변화 없이 제조공정에 S/D 공정 및 nanofiltration의 이중 바이러스 불활화 및 제거공정을 추가하고, 최종원액에 동물유래 성분 배제된 공정을 도입하여 더욱 안전성 높은 제품 개발이 가능함을 나타내었다. 또한 변경 공정 전체에 대한 validation을 위해 3 batch에 대한 시험을 실시하여 robust한 결과를 확인하여, 본 공정이 rFVIII을 안정적으로 생산할 수 있는 것과 산업 규모에서 재현성 있는 생산 결과를 가짐을 확인하였다. 마지막으로 그린진에프 단백질 특성 분석을 통해 높은 homogeneity와 FVIII 단백질로서의 기능에 이상 없음을 확인하였다. 또 다양한 이화학적 특성 및 구조적 특성 분석을 수행하여 그린진에프 및 FVIII 단백질 특성 이해에 유용한 데이터를 확보하여 품목 허가를 위한 자료를 확립하였다. 결론적으로, 본 연구를 통해 추가된 바이러스 불화화 및 바이러스 제거를 위한 S/D 공정 및 nanofiltration 공정, 그리고 최종원액의 albumin 안정제를 제거하고 아미노산을 안정제로 사용하는 공정 개선으로 인한 FVIII 단백질 특성에 변화가 없음을 확인하였다. 또한 품목허가 가이드라인을 충족하는 연구 개발 자료를 확보하였다. 이를 통해 가장 높은 수준의 안전성과 우수한 homogeneity를 가지는 rFVIII 제품을 혈우병 A 치료에 사용할 수 있음을 증명 하였다. GreenGene FⓇ (GreenGene F) or beroctocog alfa (WHO INN name), is a new B-domain deleted recombinant factor VIII (rFVIII) developed to treat hemophilia A patients. It is a highly homogeneous factor VIII protein comprising of two peptides, the heavy and light chain, at 90 and 80 kDa. The homogeneity makes it in a better position to follow GMP regulations such as quality control, easier characterization, and consistent production. GreenGene F is a third generation rFVIII product developed from its predecessor product GreenGene (second generation) by replacing albumin stabilizers in the final formulation to an amino acid stabilizer to minimize the risk of contamination of viral pathogens from plasma derived proteins. Also, an S/D process was added for viral inactivation, and a 20 nm nanofiltration (smallest pore size among rFVIII products) was added for viral removal for further assurance of the safety. By applying these processes, the GreenGene F product has significantly reduced risk of viral contamination in the raw material and process and it can be technically regarded to be the safest product among rFVIII available. This study was made to show the development procedures fulfilling regulatory aspects for the approval of GreenGene F, which process is improved from its original product GreenGene. The process from cell line to final GMP manufacturing process was summarized, the process changes for improvement were described, and the study results of the added viral inactivation/removal process and albumin-free formulation development are explained. Next, comparability assessment between GreenGene and GreenGene F were pursed with the observation of physicochemical/physiological characteristics and in vivo efficacy to see if the modification of process for safer product manufacturing affected the rFVIII protein quality. This is a required study after change of biological product process; therefore studies were performed following the required regulations from ICH, FDA, and EMA guidelines. Process validation of the GreenGene F manufacturing process, drug substance, and drug product were made to confirm the robustness of the process. Finally, observations of the protein characteristics of GreenGene F were made. The GreenGene F process produced rFVIII with high protein quality and stability, with the whole process satisfying the regulatory guidelines. Validation studies of the added viral inactivation/removal process showed high viral reduction at 10-19 logs. The three amino acid mixture, developed to replace the albumin stabilizer in the formulation for lyophilization of recombinant FVIII, had comparative results with albumin and other commonly used stabilizers. The albumin-free formulation resulted in good preservation of recombinant FVIII during lyophilization without alternations in protein characteristics. The comparability studies showed equivalence in physicochemical characteristics, physiological properties, in vivo efficacy, and toxicity between GreenGene and GreenGene F. This proved that no changes in protein characteristics of rFVIII occurred from process changes in formulation, viral inactivation, and viral removal which enhance the safety of the product. Process validation showed robust results in the 3 batches of the manufacturing process performed, confirming that the process was stable for rFVIII production and has good reproducibility at industrial scale. Protein characterization studies of GreenGene F demonstrated the high homogeneity of the protein by analyzing the purified product. Studies of physicochemical properties and structural characterization defined the characteristics of the GreenGene F product and the FVIII protein, which would be useful data for approval. In conclusion, the studies confirmed that no changes in the manufactured FVIII protein characteristics appeared due to the manufacturing process changes, where S/D and nanofiltration steps were added to the original purification steps and albumin stabilizer replaced in the final formulation with albumin-free integrates. The studies provided the data satisfying the requirements in the regulatory guidelines for approval. Therefore, an rFVIII product with the highest level of safety while possessing superior homogeneity and quality could be provided for hemophilia A treatment.

Profile monitoring and anomaly detection based on data mining algorithms

박승환 Korea University 2017 국내박사

This dissertation represents the profile monitoring method for the anomaly detection based on data mining techniques. The profile monitoring is mainly the use of control charts for cases in which the quality of a process or product can be characterized by a functional relationship between a response variable and one or more explanatory variables. In the advanced manufacturing environment, the profile monitoring is a critical task according to an increase of a manufacturing data. A conventional manufacturing process has generally used a scheduling technique and a conventional process control in order to increase yield of products and improve efficiency of facility. However, with automation of most systems and development of technology, there is a growing need for a process management based on data and statistics. Therefore, even many manufacturing companies are introducing management techniques applying hi-tech equipment and technology, more elaborate process management is required as process variables and complexity are increasing. Usually process anomaly detection is a widely known technique in a process management field. The technique is mostly conducted through a control chart, but a machine learning algorithm is also often used these days. For successful process anomaly detection, online monitoring is required so that detection results can be noticed by workers in real-time. Sensors are usually attached to equipment of process to detect process anomaly, and they generate various types of data. The data collected from sensors represent nonlinear, time-variant and multivariate properties, which are important for general data mining field. Considering these characteristics, we propose new process management techniques. First, this study proposes a monitoring chart for anomaly detection of nonlinear process signals generated by semiconductor manufacturing processes. In these manufacturing processes, fault detection and classification (FDC) and statistical process control (SPC) have been established as fundamental techniques to improve production efficiency and yield. Nonlinear process signals are collected through automatic sensing during each operation cycle of a manufacturing process. As these cyclic signals nonlinearly vary on the process state, the usage of the prevalent SPC chart is limited. Therefore, we propose a more efficient monitoring chart considering nonlinear and time-variant characteristics. Using the principal curve, a nonlinear smoothing algorithm, we construct a time-variant centerline that represents the standard pattern of the process. Then, control limits are calculated with time-variant variances over the course of the process. To evaluate performance, the proposed method was applied to industrial data for chemical vapor deposition (CVD), a semiconductor manufacturing process. We employed the misdetection ratio of signals to evaluate the performance. The proposed method demonstrated superior performance compared to other existing methods. Second, the analysis of profile data collected from the process equipment is a critical issue in improving the facility capacity and process efficiency. In particular, multi-profile monitoring is essential for process control because an advanced manufacturing process consists of numerous pieces of equipment and their related sensors. The main goal of this study is to build a monitoring chart using a Profile Integrated Measure (PIM) from multi-profile data in order to observe an overall condition of various points in the process. To deploy the proposed algorithm, multi-profile data needed to be preprocessed and applied to the prediction model. The PIM is calculated from the prediction model and reflects the relationships between the multi-profile data property, which has normal/abnormal states. The proposed algorithm constructs a model using the PIM of a normal state and identifies the performance of the model. Experiments with the simulation datasets modified from the manufacturing process validate the effectiveness and applicability of the proposed algorithm.

Study on the conversion from batch to continuous manufacturing processes for amiodarone hydrochloride tablets

윤주현 국민대학교 일반대학원 2025 국내석사

This study aimed to effectively design and optimize the manufacturing process for amiodarone hydrochloride tablets by applying continuous wet granulation. The properties of the resulting granules and tablets were evaluated to compare the differences between a conventional batch process using a high-shear mixer (HSM) and a continuous process using a twin-screw granulator (TSG). Furthermore, twin-screw rpm was identified as a Critical Process Parameter (CPP) affecting granule and tablet properties. A Design of Experiments (DoE), specifically a central composite design, was conducted to establish optimal process conditions. The independent variables (X) investigated were twin-screwrpm and milling size, while the dependent variables (Y) included dissolution, flowability, assay, disintegration, and friability. The results from the comparative study indicated no significant differences in the in-process controls (IPCs) for tablets manufactured by either process. Moreover, the dissolution profiles demonstrated equivalence to the reference drug, with a similarity factor (f2) greater than 50. Furthermore, measurement of the actual duration for each process revealed that the continuous process achieved a time reduction of over 80% relative to the batch process. In conclusion, this study confirms the feasibility of converting the manufacturing of a drug product from a traditional batch-based process to a continuous one. 본 연구는 연속 습식 과립화 공정을 적용하여 아미오다론 염산염 정제 제조를 효과적으로설계하고 최적화하는 것이다. 배치공정에서 High shear mixer(HSM)와 연속공정에서 Twin screw granulator(TSG) 사용에 따른 차이점을 비교하기 위해 과립물 및 정제의 특성을평가하였다. 또한, 과립화 공정 중 과립 및 정제 특성에 영향을 미치는 Critical Process Parameters(CPP)로 Twin screw rpm를 선정하여 Design of Experiment(DoE) 중 중심 합성설계법으로 실험을 수행하였다. X값으로 Twin screw rpm, milling size를 선정하였고, Y값으로dissolution, flowability, assay, disintegration, friability를 선정하여 최적의 연속공정 조건을도출하고, 이를 배치공정과 비교하였다. 배치공정과의 비교 연구 결과 제조된 정제들의 공정중시험(IPC)에서 유의한 차이가 없었고, 용출 프로파일 또한 유사성 인자(F2)가 50 이상으로대조약과의 동등성을 확보하였다. 또한, 각 공정의 실제 소요시간을 측정한 결과 배치공정 대비연속공정의 소요시간이 85% 이상 단축되는 것을 확인하였다. 결론적으로 본 연구를 통해배치공정 기반 약물을 연속공정 기반으로의 전환 가능성을 확인하였다. This study evaluates the formulation possibilities of amiodarone hydrochloride, an antiarrhythmic agent used for the treatment and prevention of arrhythmia, for development into a continuous manufacturing process dosage form. Amiodarone hydrochloride is a calciumchannel blocker and an antiarrhythmic agent listed on the national essential medicines list. We predicted and tested the physicochemical properties and essential preformulation results of amiodarone hydrochloride using our proprietary AI formulation design software. In this study, apparent and equilibrium solubility tests were conducted in various solvents, and degradation was examined to confirm the compatibility between amiodarone hydrochloride and various excipients. Additionally, flowability was assessed to predict the physical properties of the active pharmaceutical ingredient (API). The solubility test results showed that amiodarone hydrochloride was practically insoluble in aqueous solutions but was 'soluble' in methanol. Furthermore, it was confirmed to be more soluble in Polysorbate 80 (Tween 80) than in Sodium Lauryl Sulfate among the tested surfactants. In the compatibility study, the interaction between amiodarone hydrochloride and 21 different excipients was observed for up to 4 weeks under long-term (25°C/60%RH) and accelerated (40°C/75%RH) storage conditions. The flowability study of the API resulted in a Carr's index of 30. Through this research, we confirmed the potential for developing amiodarone hydrochloride into an oral tablet dosage form with improved flow properties by using a glidant. The purpose of this study was to develop an oral formulation of amiodarone hydrochloride optimized for manufacturing via a continuous wet granulation process. A significant challenge arose when a formulation, previously established as pharmaceutically equivalent in a batch process, was converted to continuous production. Specifically, the initial dissolution rates were diminished by 20% under simulated gastric pH conditions and by 30% under simulated intestinal pH conditions. To address this issue and enhance the initial dissolution rates amiodarone hydrochloride tablets were manufactured by optimizing the proportions of the binder and the binding solution. The formulation's critical quality attributes (CQA) were assessed using in-process controls (IPC) and dissolution profiling. In a supplementary investigation the effects of varying the type and ratio of lactose on IPC results and dissolution profiles were also compared. The in vitro studies revealed no statistically significant impact of lactose type or ratio on dissolution rate. However the newly manufactured amiodarone hydrochloride tablets with an adjusted binder and binding solution ratio showed a 33% increase in the initial dissolution rate in the gastric pH environment and a 24% increase in the intestinal pH environment compared to the tablets that had established pharmaceutical equivalence in the batch process. In conclusion the novel tablet formulation developed in this study is anticipated to be highly suitable for efficient application in a continuous wet granulation process.

Integrated Manufacturing Process Design of Al6061 Alloy Bolts for Fastening Offshore Platforms

Park, Sung Cheol 한국해양대학교 대학원 2019 국내석사

최근 수송 분야에서 경량화에 대한 관심이 증가하면서, 각종 수송 기계 부품에 비철금속을 적용하는 사례가 늘고 있다. 특히, 6xxx 계열의 알루미늄 합금은 낮은 비중량, 우수한 기계적 특성 및 높은 내식성으로 인해 소형 선박에서부터 해양 구조물에 이르기까지 다양한 해양 분야에 적용되고 있다. 패스너를 사용한 기계식 체결은 사용이 쉽고 적용이 용이하여 알루미늄 합금과 강재의 접합에 널리 사용되어왔다. 현재까지도 산업 현장에서는 알루미늄 합금과 강재의 체결을 위해 스테인리스강 볼트가 사용되고 있다. 그러나 알루미늄 합금의 모재에 기존의 스테인리스강 볼트를 사용하게 되면 중량의 증가뿐만 아니라 알루미늄 모재에서 갈바닉 부식이 발생하게 되고 이것은 특히 해양 환경에서 더욱 가속화된다. 따라서 스테인리스강 볼트를 알루미늄 합금 볼트로 대체하는 것은 추가적인 경량화를 도모할 수 있을 뿐만 아니라 해양 환경에서 이러한 갈바닉 부식의 생성을 미연에 방지할 수 있다. 현재까지 알루미늄 합금 볼트 성형에 대해 일부 연구들이 수행되어 왔다. 그러나 대부분의 연구들은 재료의 성형성 평가에 국한되어 있으므로 성형 공정에 대한 명확하고 체계적인 설계 기준이 없다. 또한, 볼트를 제조하기 위해 필수적인 트리밍 및 전조 공정의 설계가 누락되어있다. 따라서 본 연구에서는 해양구조물 체결을 위한 Al6061 합금 볼트의 통합 제조공정 설계를 수행하였다. 이를 위해 이론 설계 및 수치 해석을 진행하였다. 이론 설계에서는, 스크류 규격과 기하학적 관계를 고려하여 전조 공정의 초기 소재경과 압입 깊이를 계산하였다. 헤딩 공정을 위한 초기 소재 치수는 체적일정법칙으로 구하였다. 결함 예측에 사용되는 공정 한계를 고려하여, 공정 단수를 설정하고, 헤딩 공정의 설계 규칙에 의해 예비 형상을 결정하였다. 또한 설계된 헤딩 공정에 대해 슬래브법을 사용하여 성형 하중을 예측하였다. 이론 설계를 바탕으로 헤딩, 트리밍 및 전조 공정에 대한 유한요소해석을 수행하였다. 헤딩 및 트리밍 공정 중에 발생하는 결함을 예측하기 위해 연성파괴기준을 적용하였다. 또한 전조 공정에서 금형의 압입량, 이송 속도 및 분당 회전수, 트리밍 공정에서는 펀치의 곡률 반경, 하부 금형의 랜드 폭 및 펀치와 하부 금형간의 정지 거리를 각각 설계 변수로 선정하고 다구찌법을 통해 공정을 최적화하였다. 제안 된 설계 방법을 검증하기 위해 알루미늄 합금 볼트 성형 실험을 수행하였고 높은 치수 정확도를 갖는 건전한 형상의 Al6061 합금 볼트를 얻을 수 있었다. In recent years, with the increasing interest in weight reduction in the transportation sector, the application of non-ferrous metals to various transportation machinery parts has been increased. In particular, the 6xxx series of aluminum alloys are commonly used in marine applications from small vessels to offshore structures due to its low specific weight, high quality mechanical properties, and high resistance to corrosion. In order to join aluminum alloy and steel, mechanical joint by fasteners has been widely used because it is easy-and-practical for the application. Until now, stainless steels bolts have been used for fastening aluminum alloys and steels to industrial sites. However, when the stainless steel bolt is used for the base material of aluminum alloy, weight increases and galvanic corrosion occurs in aluminum, which is accelerated in marine environment particularly. Therefore, substituting aluminum alloy bolts for stainless steel ones can prevent such corrosion promotion in marine environment as well as extra weight increase. To date, several studies have been conducted on aluminum alloy bolt forming. However, most studies are limited to the evaluation of formability of materials, so there is no clear and systematic design criterion for the forming process. Furthermore, the design of the trimming and thread-rolling process, essential for manufacturing bolts, was not considered. Therefore, the integrated manufacturing process design of Al6061 alloy bolts for fastening offshore platforms was performed in this study. To do this, we carried out theoretical design and numerical analysis. At theoretical design, the initial rod diameter and the penetration depth in thread-rolling process were calculated according to the screw standards and geometric relation. Thereafter, the dimensions of the initial workpiece for the heading process were obtained by volume constancy law. Considering process limitations to predict the defects, the number of the stage was set, and preform was determined by the design rule in the heading process. In addition, the forming load for the designed heading process was predicted by using the slab method. Based on theoretical design, FE-analysis for the heading, trimming and thread-rolling process was conducted. The ductile fracture criterion was applied to predict the defects during the heading and trimming process. In addition, the Taguchi method is used to optimize the trimming and thread-rolling process with the set of design parameters, such as the penetration depth, transfer velocity, and revolutions per minute of rolling dies in the thread-rolling process, and the blade radius of the punch, land width of the bottom die, and stop distance between the punch and bottom die in the trimming process, respectively. To verify the proposed design, the aluminum alloy bolt forming experiment was carried out and sound Al6061 alloy bolt with high dimensional accuracy was obtained.

A process integrated engineering output management model for project-based manufacturing

한진택 연세대학교 대학원 2015 국내박사

In a project-based ETO industry, design, purchasing, production, and quality processes are closely integrated to meet a delivery schedule, cost limitation, and quality specification. When an engineering change occurs in a project execution stage, miscommunication of design change may generate a significant cost and delay.The Research suggests a model for collecting, managing and utilizing engineering outputs integrated with product life-cycle process of project-based industry. Project output acquisition is designed to collect outputs without additional efforts at the time outputs are registered accordingly with the project schedule. Once revision is occurred in the project period, the revision is designed to be transferred to the corresponding process by using ECO process.For this, the Research suggests a model to capture a design change and share activities for processing change done by each stakeholder. The framework also captures ECR, ECO, and a relationship among them with a result of processes. The data could be used for analyzing defects of each process, improving a design quality, and a plan for cost reduction.For capturing, sharing, and monitoring every change of drawing element and BOM, the model extracts a drawing and BOM from a product structure of a 3D CAD system. Extracted data is sent to PLM/ERP system for planning, purchasing, manufacturing, and quality control processes.To increase usability of the outputs as enterprise knowledge, accumulated outputs are managed and stored by applying multiple classifications from perspectives of process activity, contents, and enterprise knowledge management during accumulation process is carrying out. As a result of applying the suggest system in electric transformer EPC project performing concurrent-multiple project, usability and a degree of accumulating outputs that includes revision history and process accuracy are improved compared to previously used PDM system.The research is focused on collecting, acquiring project outputs and change management. Using engineering data which is collected based on the research and actual data from project processes, a framework for Evaluating and estimating project status and profitability will be developed.

플라즈마 전자빔을 이용한 신개념 금속 적층 제조 공정의 개발을 위한 초합금 분말의 전자빔 예열 및 적층 특성에 관한 연구

이호진 조선대학교 일반대학원 2018 국내박사

A powder bed fusion (PBF) process is one of representative metal additive manufacturing (AM) processes. The PBF process can be classified into laser and electron beam types according to the applied heat source. The power of the electron beam is greater than that of the laser. In addition, the electron beam is hardly influenced by the reflectivity of the applied material unlike the laser. The PBF process using electron beam is one of challengeable metal AM processes. However, only one PBF process using an electron beam, so called electron beam melting (EBM) process, has been developed. EBM process adopts a thermal electron beam using a hot cathode. The EBM process has several disadvantageous characteristics including a high level of vacuum environment, a short service life of the cathode, a high temperature of the cathode, a small diameter of beam, a high temperature of the building plate, etc.. The development of a novel PBF type AM process using an electron beam is needed to overcome demerits of the EBM process. The aim of this thesis is to investigate preheating and deposition characteristics of super-alloy powders for the development of a novel metal additive manufacturing process using a plasma electron beam. A plasma electron beam has several benefits, including a relatively low level of vacuum environment, a long service life of the cathode, a relatively large diameter of the beam, a low temperature of the cathode, as compared to the thermal electron beam. Due to these merits, a novel PBF type AM process using a plasma electron beam was proposed in this thesis. The proposed process consisted of preheating step to prevent spreading of powders, deposition step to create the deposited bead and re-melting step to improve surface roughness and density. Various experiments and numerical analyses were carried out to investigate the influence of process parameters on preheating and deposition characteristics of powders. The experiments were performed for two conditions, including no heating condition using the plasma electron beam (without heating of building plate) and heating condition using the plasma electron beam (with heating of building plate), according to heating of the building plate. The building plate was heated to 150 oC before loading of the building plate in the vacuum chamber for the case of the no heating condition. However, the building plate was heated using a plasma electron beam in the vacuum chamber before experiments for the case of the heating condition. The working distance, the probe current of the electron beam, the acceleration voltage of the electron beam, the travel speed of the table, and the length of the bead were chosen as process parameters. Stellite21 powders with 45-150 mm of diameter were selected as the used material. The experiments were carried out using a plasma electron beam finishing system. A commercial software ABAQUS was used to perform numerical analyses. Temperature dependent material properties including state and porosity of powders were used in the numerical analyses. Sintering and melting temperatures of the Stellite21 powders were estimated via sintering and melting experiments using a fiber laser and an infrared camera. Preheating experiments were performed for the case of no heating condition of the building plate. From the results of the experiments, the effects of the working distance, the probe current and the travel speed on spreading characteristics of powders and the formation of preheated beads were examined. In addition, appropriate working distance, probe current and travel speed were estimated to create the desired preheating bead. In order to investigate heat transfer characteristics during preheating of powders using a plasma electron beam, three-dimensional finite element analyses (FEAs) were performed. The heat source model of the plasma electron beam for preheating step were proposed through the measurement of intensity of the probe current and the regression method using normalized radius and intensity. Using the results of the FEAs, the effects of the probe current and the travel speed on the temperature distribution in the vicinity of the preheated region were investigated. The influence of the temperature distribution in the vicinity of the preheated region on the spreading phenomenon of powders was discussed. Deposition experiments were carried out for the case of no heating condition of the building plate. The influence of the probe current, the travel speed and the bead length on defects, morphologies, dimensions of deposited beads was examined. From the results of the examination, a proper deposition condition for the case of no heating condition of the building plate was estimated. In addition, a planner part with repeated deposition beads was fabricated using the proposed condition. FEAs for melting of powders during the deposition step were carried out to investigate the influence of the probe current and the travel speed on the temperature distribution in the vicinity of the melted region. The influence of the probe current and the travel speed on the formation of the melted region was examined using the results of FEAs. Through the comparison of results of experiments and those of FEAs, creation mechanism of the deposited bead for the case of no heating condition of the building plate was discussed. Using results of experiments and numerical analyses of preheating and deposition steps without heating of the building plate, multi-layer deposition experiments were carried out. A three-dimensional part with several defects was fabricated by the experiments. From the results of the experiments, it was noted that automatic successive building system is needed to reduce temperature changes during the deposition and delamination between successive layers. In addition, it was considered that the building plate should be heated in the vacuum chamber before feeding of powders to improve joining characteristics between the building plate and the deposited part. A PBF type experimental set-up with automatic feeding, deposition and specimen loading devices was developed to perform preheating and deposition experiments in a vacuum chamber. In order to investigate the probe current and focusing of the electron beam, a Faraday cage and a beam focusing tester were assigned in the experimental set-up. Using the experimental set-up, preheating and deposition experiments were carried out for the case of heating of the building plate. In those experiments, the building plate was heated during 10 minutes using the defocused plasma electron beam in a vacuum chamber. After heating of the building plate, powders were fed on the building plate using the automatic feeding device. Preheating conditions for the case of heating of the building plate were estimated through the examination of the influence of the probe current and the travel speed on the formation of the preheated bead. The effects of the acceleration voltage, the probe current and the travel speed on the formation, the ingredient variation and the surface roughness of the deposited bead were investigated through deposition experiments for the case of heating of the building plate. From the results of the investigation, the deposition window and the proper deposition condition for the case of heating of the building plate were estimated to fabricate the desired deposited bead. Multi-layers deposition experiments were carried out to investigate feasibility of the fabrication of three-dimensional parts using the proposed AM process. Estimated proposed preheating and deposition conditions for the case of heating of building plate were applied to the multi-layers deposition experiments. In multi-layers deposition experiments, a single layer was created from heating of building plate, feeding of powders, preheating of fed powders, deposition of bead, and surface re-melting of the deposited bead. Through the repetition of the creation method of the single layer, three-dimensional parts were fabricated. Planner part with a single layer, rectangular parts with 3 and 10 layers, cylindrical part with 5 layers, stepped part with 10 layers, and square net part with 11 layers were fabricated. The re-melting step was not applied to the fabrication of the square net part. From the results of multi-layers deposition experiments, it was noted that the proposed AM process can fabricate three-dimensional parts with a high density.

Primary and secondary metabolite profiling of Korean fermented red pepper paste (gochujang) according to industrial manufacturing processes

신기루 건국대학교 대학원 2015 국내석사

본 연구에서는, 4가지 산업 공정을 거쳐 제조되는 고추장을 가스 및 액체 크로마토그래피-질량분석기를 이용하여 분석하고, 다변량 통계 분석을 사용하여 종합적인 대사체 프로파일링을 실시하였다. 각각의 제조 공정별 고추장 시료는 부분최소제곱법-판별분석(PLS-DA)을 통하여 뚜렷하게 나뉘는 것으로 나타났다. 아미노산, 유기산, 당, 당알콜, 아이소플라본, 플라보노이드, 알칼로이드 및 리소인지질은 각각의 공정을 구분 짓는 주요한 대사체로 확인되었다. 단당류는 증가하고 이당류는 감소하는 현상이 공정 1에서 나타났으며, 증자찹쌀, 증자 밀이 첨가되면서 이러한 첨가물이 함유하고 있는 리소인지질이 증가하는 것이 공정 2에서 나타났다. 고추장의 맛에 기여하는 아미노산은 공정 3에서 증가하는 것을 알 수 있었고, 공정 4에서는 고추양념에 함유 되어 있는 루테올린-글르코사이드, 아피제닌-글루코사이드, 루테올린, 캡사이신, 그리고 다이하이드로캡사이신과 같은 2차 대사산물이 고추양념이 첨가됨에 따라 증가하는 양상이 나타났다. 또한 메주가루를 함유 하고 있는 공정을 통하여 아이소플라본 배당체는 공정 3 이후에 점차적으로 감소하였으며, 동시에 아이소플라본 아글리콘은 증가하는 것을 알 수 있었다. 따라서 본 연구 결과, 대사체 프로파일링이 고추장 제조 공정을 평가 가능한 방법이며, 대사체학 기법이 품질관리의 도구로 사용될 수 있다는 것을 입증하였다. In this study, the comprehensive metabolite profiling of gochujang, a Korean fermented red pepper paste, at industrial 4-process was conducted by gas and liquid chromatography-mass spectrometry with multivariate analysis. From the partial least squares-discriminant analysis of primary and secondary metabolites, the patterns were clearly distinguished according to the processes (process 1: 1st fermentation, process 2: 1st addition, process 3: 2nd fermentation, process 4: 2nd addition). Amino acids, organic acids, sugars, sugar alcohols, isoflavones, flavonoids, alkaloids, and lysoPCs were identified by each process. Monosaccharide increased while disaccharide decreased in process 1 due to use carbon source and lysoPCs increased due to add additives such as steamed glutinous rice and steamed wheat in process 2. Most of the amino acids related to taste of gochujang increased in process 3. In process 4, some secondary metabolites such as luteolin-glucoside, apigenin-glucoside, luteolin, capsaicin, and dihydrocapsaicin derived from red pepper were increased. Isoflavone glycosides gradually decreased after process 3 while isoflavone aglycones increased in process 3-4 including meju powder. The result of this study proved that metabolite profiling is possible to evaluate gochujang manufacturing processes and metabolomics techniques can be used as quality control tool.

Some important issues after fault identification in multivariate manufacturing processes : application to TFT-LCD manufacturing processes

이규종 Graduate School, Korea University 2013 국내석사

Multivariate control charts have been widely recognized as efficient tools for detection of abnormal behavior in multivariate processes. However, these charts provide only limited information about the contribution of any specific variable to an out-of-control signal. To address this limitation, some fault identification methods have been developed to identify contributors to an abnormality. In real situations, however, a couple of tasks should be further considered with these contributors to improve their applicability and to facilitate interpretation of faults. This study presents a rank sum-based summarization technique and a decision tree algorithm to facilitate the interpretation of fault identification results. Experimental results with real data from the manufacturing process for a thin film transistor-liquid crystal display demonstrate the applicability and effectiveness of the proposed methods.