• 한국정보처리학회:학술대회논문집
• /
• 한국정보처리학회 2019년도 추계학술발표대회
• /
• pp.760-761
• /
• 2019

Smart manufacturing is a process of integrating computer-related technologies in production and by doing so, achieving more efficient production management. The recent development of supercomputers has led to the broad utilization of artificial intelligence (AI) and machine learning techniques useful in predicting specific patterns. Despite the usefulness of AI and machine learning techniques in smart manufacturing processes, there are many fundamental issues with the direct deployment of these technologies related to data management. In this paper, we focus on solving the outlier detection issue in smart manufacturing applications. More specifically, we apply a state-of-the-art outlier detection technique, called Elliptic Envelope, to detect anomalies in simulation-based collected data.

• 한국시뮬레이션학회논문지
• /
• 제10권3호
• /
• pp.71-82
• /
• 2001

Considering material handling delay which occurs between storage and processing stations, we propose an algorithm to estimate the required storage capacity, i.e., number of aisles and number of openings in vertical and horizontal directions in each aisle, of an automated storage/retrieval(AS/R) system. Due to the random nature of storage and retrieval requests, proportion of single and dual commands is not known in advance. Two design criteria, maximum permissible overflow probability and maximum allowable storage/retrieval(S/R) machine utilization, are used to compute the storage capacity. Most of studies assume that storage capacity of AS/R systems is given, although it is a very important decision variable in the design phase. Therefore, the proposed model can be effectively used in the design phase of new AS/R systems.

• 산업융합연구
• /
• 제3권1호
• /
• pp.19-31
• /
• 2005

This paper deals with the sequencing problem in the operation of the manufacturing systems with the constraint of buffer capacity. Some of studies for this theme have been progressed for several years. And then most of them considered only one objective, such as maximum lateness, machine utilization, makespan, mean flowtime and so on. This study deal with two objectives of the delivery for customers and the idle time of machines for producers. For the decision of sequence, the utility function is used. The developed heuristic algorithm presents a good solution. Through a numerical example, the procedures of the job sequencing is explained.

• 신재생에너지
• /
• 제18권2호
• /
• pp.18-25
• /
• 2022

In this study, we propose a wind power generation prediction system that applies machine learning and data mining to predict wind power generation. This system increases the utilization rate of new and renewable energy sources. For time-series data, the data set was established by measuring wind speed, wind generation, and environmental factors influencing the wind speed. The data set was pre-processed so that it could be applied appropriately to the model. The prediction system applied the CNN (Convolutional Neural Network) to the data mining process and then used the LSTM (Long Short-Term Memory) to learn and make predictions. The preciseness of the proposed system is verified by comparing the prediction data with the actual data, according to the presence or absence of data mining in the model of the prediction system.

• KEPCO Journal on Electric Power and Energy
• /
• 제7권2호
• /
• pp.301-307
• /
• 2021

After the spread of the data collection devices such as smart meters, energy data is increasingly collected in a variety of ways, and its importance continues to grow. However, due to technical or practical limitations, errors such as missing or outliers in the data occur during data collection process. Especially in the case of customer-related data, billing problems may occur, so energy companies are conducting various research to process such data. In addition, efforts are being made to create added value from data, which makes it difficult to provide such services unless reliability of data is guaranteed. In order to solve these challenges, this research analyzes prior research related to bad data processing specifically in the energy field, and propose new missing value processing methods to improve the reliability and field utilization of energy data.

• International Journal of Internet, Broadcasting and Communication
• /
• 제16권1호
• /
• pp.163-168
• /
• 2024

Technologies that are considered to be a part of the fourth industrial revolution include holograms, augmented reality, and virtual reality. As technology advances, the industry's scale is growing quickly as well. While the development of technology for direct use is moving slowly, awareness of floating holograms-which are considered realistic content-is growing as the industry's scale and rate of technological advancement continue to accelerate. Specifically, holograms that have been incorporated into museums and exhibition spaces are static forms of content that viewers gaze at inertly. Additionally, their use in educational fields is very passive and has a low rate of utilization. Therefore, in order to improve usability from the viewpoint of viewers of realistic content, such as exhibition halls or museums, we introduce realistic content control technology in this study using a machine learning framework to recognize hands. It is anticipated that using the study's findings, manipulating realistic content independently will enhance comprehension of objects presented as realistic content and boost its applicability in the industrial and educational domains.

• 한국지능시스템학회논문지
• /
• 제11권6호
• /
• pp.505-509
• /
• 2001

본 연구에서는 학습기근을 갖는 결정론적 볼츠만 머신의 은닉충 뉴런에 비단조 활성화 함수를 적요한 경위의 학습성능을 XOR 문제와 ADD 학습에 대하여 수지 시뮬레이션을 통하여분석한다. 단조 활성화함수를 사용한 경우와 비교하여 학습 수렴률, 학습안정도, 및 학습 속도에 있어서 성능이 크게 향상됨을 확인하였다. 또한 네트워크의 막전위 분포를 조사함으로서 end-cut-off 타입의 비단조 함수를 이용한 경우에 나타는 다음 층의 뉴런에 영향을 주지 않는 뉴런의 출현, 즉, 신경회로망에 있어서 은닉층 뉴런늬 수을 자율적으로 조정하는것을 확인하였따. 이것은 학습문제에 대하여 네트워크 은닉층 뉴런의 수를 명확하게 결정할수 없는 현재의 상황에 있어서는 새로운 돌파구가 될것으로 기대된다.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 2007년도 학술대회 논문집 전문대학교육위원
• /
• pp.112-116
• /
• 2007

If any of the set parameters such as the environment temperature, mold temperature are not maintained at a consistent level, the fail rate of injection molding products is increased. The price of the injection molding machine is very high, so in order to maximize the utilization of the machine that is required the production of a number of different products with minimum fail rate using a single machine. To prevent the defect products by an inspection process with perfect quality is very important to minimizing production of defect products in the molding process. Vision inspection systems are widely utilized in various manufacturing industries for quality assurance purposes. The vision inspection system consists of CCD camera and lighting system to capture the image of the subject of inspection, an image comparison algorithm using to determine the pass/fail of the products, and mechanical devices for the operation of the whole system. This research focuses on the development of the vision inspection system to process the inspection of an automobile parts. We developed a mechanical devices for the inspection of the injection molding products and an image comparison algorithm to determine the pass/fail result of the inspection based on the molding image and the accepted product image.

• 대한산업공학회지
• /
• 제19권2호
• /
• pp.23-34
• /
• 1993

Although there are numerous studies that address the problem of optimal machine grouping and part family classification for cellular manufacturing, little research has been reported that studies the conditions where cellular manufacturing is appropriate. This paper, in order to evaluate and compare the job shop with the GT cellular shop, the performance of those shops were simulated by using SIMAN. We tested the effect of independent variables including changes of product demands, intercell flow level, group setup time, processing time variability, variety of material handling systems, and job properties (ratio of processing time and material handling time). And also performance measures (dependent variables), such as machine utilization, mean flow time, average waiting time, and throughput rate, are discussed. Job shop model and GT cellular shop written in SIMAN simulation language were used in this study. These systems have sixteen machines which are aggregated as five machine stations using the macro feature of SIMAN. The results of this research help to better understand the effect of production factors on the performance of cellular manufacturing systems and to identify some of the necessary conditions required to make these systems perform better than traditional job shops. Therefore, this research represents one more step towards the characterization of shops which may benefit from cellular manufacturing.

• ETRI Journal
• /
• 제42권5호
• /
• pp.686-699
• /
• 2020

In this paper, we propose a supervised-learning-based spatial performance prediction (SLPP) framework for next-generation heterogeneous communication networks (HCNs). Adaptive asset placement, dynamic resource allocation, and load balancing are critical network functions in an HCN to ensure seamless network management and enhance service quality. Although many existing systems use measurement data to react to network performance changes, it is highly beneficial to perform accurate performance prediction for different systems to support various network functions. Recent advancements in complex statistical algorithms and computational efficiency have made machine-learning ubiquitous for accurate data-based prediction. A robust network performance prediction framework for optimizing performance and resource utilization through a linear discriminant analysis-based prediction approach has been proposed in this paper. Comparison results with different machine-learning techniques on real-world data demonstrate that SLPP provides superior accuracy and computational efficiency for both stationary and mobile user conditions.