제조셀 형성시 robustness를 고려한 신경망 기법

이준한 한국전문가시스템학회 1998 학술대회 Vol.1 No.1

사출성형공정에서 다수 품질 예측에 적용가능한 다중 작업 학습 구조 인공신경망의 정확성에 대한 연구

이준한,김종선 한국금형공학회 2022 한국금형공학회지 Vol.16 No.3

In this study, an artificial neural network(ANN) was constructed to establish the relationship between process condition prameters and the qualities of the injection-molded product in the injection molding process. Six process parmeters were set as input parameter for ANN: melt temperature, mold temperature, injection speed, packing pressure, packing time, and cooling time. As output parameters, the mass, nominal diameter, and height of the injection-molded product were set. Two learning structures were applied to the ANN. The single-task learning, in which all output parameters are learned in correlation with each other, and the multi-task learning structure in which each output parameters is individually learned according to the characteristics, were constructed. As a result of constructing an artificial neural network with two learning structures and evaluating the prediction performance, it was confirmed that the predicted value of the ANN to which the multi-task learning structure was applied had a low RMSE compared with the single-task learning structure. In addition, when comparing the quality specifications of injection molded products with the prediction values of the ANN, it was confirmed that the ANN of the multi-task learning structure satisfies the quality specifications for all of the mass, diameter, and height.

사출성형공정에서 CAE 기반 품질 데이터와 실험 데이터의 통합 학습을 통한 인공지능 품질 예측 모델 구축에 대한 연구

이준한,김종선 한국금형공학회 2021 한국금형공학회지 Vol.15 No.4

In this study, an artificial neural network model was constructed to convert CAE analysis data into similarexperimental data. In the analysis and experiment, the injection molding data for 50 conditions were acquired through thedesign of experiment and random selection method. The injection molding conditions and the weight, height, and diameterof the product derived from CAE results were used as the input parameters for learning of the convert model. Also theproduct qualities of experimental results were used as the output parameters for learning of the convert model. Theaccuracy of the convert model showed RMSE values of 0.06g, 0.03mm, and 0.03mm in weight, height, and diameter,respectively. As the next step, additional randomly selected conditions were created and CAE analysis was performed. Then, the additional CAE analysis data were converted to similar experimental data through the conversion model. Anartificial neural network model was constructed to predict the quality of injection molded product by using convertedsimilar experimental data and injection molding experiment data. The injection molding conditions were used as inputparameters for learning of the predicted model and weight, height, and diameter of the product were used as outputparameters for learning. As a result of evaluating the performance of the prediction model, the predicted weight, height,and diameter showed RMSE values of 0.11g, 0.03mm, and 0.05mm and in terms of quality criteria of the target product,all of them showed accurate results satisfying the criteria range.

인천인구사 2

이준한 인천대학교 인천학연구원 2011 인천학연구총서 Vol.14 No.-

사출성형 중 높은 전단속도에서 용융 수지의 점도 측정에 대한 연구

이준한,김종선 한국금형공학회 2019 한국금형공학회지 Vol.13 No.3

In this study, a mold for measuring melt viscosity was fabricated with a slit of 10mm in width, 0.5mm in thickness and 200mm in length. Two pressure sensors was installed in the slit to measure the viscosity of melt resin during injection molding. four different grades of PETG were used and the viscosity was measured in the injection molding machine with a screw diameter of 40 mm by setting the melt temperature to 220, 240, 260 ° C, and the injection speed of 10, 20, 30, 40 mm/s. The viscosity was calculated by the shear stress and shear rate, respectively, using the pressures measured at the pressure sensors during injection molding and the injection flow rate. The result showed that all four PETGs has a tendency of decreasing viscosity with increasing melt temperature and shear rate. This tendency was confirmed to be similar to that of PETG measured by general rheometer and the measured values were similar. Therefore, when the viscosity of the melt resin is measured with the injection rheometer, it was confirmed that reliable results of similar tendencies and levels as those of actual resins can be obtained at the range of high shear rate.

Voter Turnout “Surge and Decline”: Asian New Democracies

이준한 한국학술연구원 2004 Korea Observer Vol.35 No.2

This paper addresses electoral turnout in Asia’s new democracies, that is, Bangladesh, East Timor, In-donesia, Korea, Mongolia, Nepal, Pakistan, the Phil-ippines, Taiwan, and Thailand. Vote/Voting Age Population data collected by the International Insti-tute for Democracy and Electoral Assistance suggest the “surge and decline” model aptly portrays the dy-namics and patterns of post-transition voter turnout, as initial democratic elections swelled participation, only to be followed by a drop in most of these nations. Turnout in the new Asian democracies is expected gradually to decline before reaching a largely stabi-lized and normalized level.

서해5도의 정치ㆍ행정

이준한 인천대학교 인천학연구원 2015 인천학연구총서 Vol.33 No.-

결선투표제의 비판적 고찰

李埈漢 한국의정연구회 2010 의정논총 Vol.5 No.2

이 논문은 결선투표제의 정치적 효과에 대하여 비판적으로 고찰하는 목적을 갖는다. 결선투표제는 현재 한국에서 개헌의 주요한 대안 가운데 하나이다. 따라서 이 논문은 결선투표제에 대한 이해를 확장시키고 결선투표제 도입에 대한 논의의 폭을 넓히는 데 기여하고자 한다. 이러한 목적에 도달하기 위하여 이 논문은 전 세계적인 차원에서 진행된 다양한 결선투표제에 대한 연구결과를 종합하고 비교해본다. 그 가운데 이 논문이 주목하는 것은 결선투표제의 다섯 가지 정치적 효과이다: 첫째, 결선투표제의 정통성 약화 효과, 둘째, 결선투표제의 투표율 저하 경향, 셋째, 결선투표제의 선거결과 역전 경향, 넷째, 결선투표제의 정당정치의 불안정화, 다섯째, 결선투표제의 네가티브 선거연합 구축 경향이다. This essay surveys on the trends and contours of the extensive literature on the runoff systems and their political effects on the globe. In doing so, this essay attempts to contribute to expand the understandings of the runoff systems both in political science community and political arena in Korea. The major findings of this essay are as follows: First, the runoff systems tend to manipulate election outcomes, despite the noble idea of majority rule. Second, voter turnout is more likely to decline in the second round. Third, come-from-behind candidates are likely to win in the second round so that the reverse of outcome is likely to occur in the runoff systems. Forth, the runoff systems tend to foster electoral competitions and party fragmentations. Finally, the runoff systems are likely to induce a formation of unstable coalition against extremist parties before the second round.

제20대 국회와 언론의 역할

이준한 관훈클럽 2016 관훈저널 Vol.- No.139

인공신경망을 이용한 사출성형품의 무게 안정성 제어에 대한 연구

이준한,김종선 한국산업융합학회 2020 한국산업융합학회 논문집 Vol.23 No.5

In the injection molding process, the controlling stability of products quality is a very important factor in terms of productivity. Even when the optimum process conditions for the desired product quality are applied, uncontrollable external factors such as ambient temperature and humidity cause inevitable changes in the state of the melt resin, mold temperature. etc. Therefore, it is very difficult to maintain prodcut quality. In this study, a system that learns the correlation between process variables and product weight through artificial neural networks and predicts process conditions for the target weight was established. Then, when a disturbance occurs in the injection molding process and fluctuations in the weight of the product occur, the stability control of the product quality was performed by ANN predicting a new process condition for the change of weight. In order to artificially generate disturbance in the injection molding process, controllable factors were selected and changed among factors not learned in the ANN model. Initially, injection molding was performed with a polypropylene having a melt flow index of 10 g/10min, and then the resin was replaced with a polypropylene having a melt floiw index of 33 g/10min to apply disturbance. As a result, when the disturbance occurred, the deviation of the weight was –0.57 g, resulting in an error of –1.37%. Using the control method proposed in the study, through a total of 11 control processes, 41.57 g with an error of 0.00% in the range of 0.5% deviation of the target weight was measured, and the weight was stably maintained with 0.15±0.07% error afterwards.