Analysis of relationships between defects and variables in wire mesh production using association rules and simple regression analysis

• Main Authors: Aree Wiriyaphongsanon, Chompu Inwai Rungchat
• Format: Conference Proceeding
• Published: 2018
• Subjects: Computer Science
• Online Access: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85054567235&origin=inward; http://cmuir.cmu.ac.th/jspui/handle/6653943832/62669
• Institution: Chiang Mai University

© Proceedings of 2018 the 8th International Workshop on Computer Science and Engineering, WCSE 2018. All rights reserved. Many categories of defects frequently occur in the production process of wire mesh used for construction. This research has the main objective of finding variables or factors in production which are related with the amounts of defects occurring, to find ways of reducing them in future. The procedure of this research started from studying the wire mesh production process in the case study company and selecting a product for the study. For the next step, as there were many categories of defect, these categories were prioritized to select those which required prior improvement using the Pareto principle and the principle of Association Rules. The latter were used to analyze relationships between defects such as whether any kinds of defects occurred together, to help reduce the categories of defects requiring study. The research found that there were two important categories of defect which should be addressed as priority, type S1 defect and type S2 defect. From there analysis was conducted to identify the causes of each kind of defect with a Cause and Effect Diagram. A brainstorming session with the relevant personnel found that machinery settings were the factors causing opportunities for the greatest amounts of defects to occur. The researchers then collected data of detailed machine settings and defects which had occurred in the past. Simple Regression Analysis was then used to identify the machine setting factors related with the amounts of defects. The research found that the type S2 defect was related with only the X8 factor, while the type S1 defect was related with the factors X3, X4 and X10