1 AIT Asian Institute of Technology

Modified shewhart attribute control charts to detect defects for final inspection process at intel products Vietnam

AuthorHo Ngoc Vy
Call NumberAIT Thesis no.ISE-13-18
Subject(s)Inspection--Defect--Vietnam
Intel Corporation--Production
NoteThe requirements for the degree of Master of Engineering (Professional) in Industrial and Manufacturing Engineering with Area of Specialization in Industrial Engineering and Management, School of Engineering and Technology
PublisherAsian Institute of Technology
Series StatementThesis;no. ISE-13-18
AbstractFinal Inspection is the final gate that screens all nonconforming items by 100% inspection. Currently, there are no statistical limits that judge whether the process is in control. People are subjective to decide which number of defects is potentially caused by process excursion. All of this requires constructing Statistical Control Chart for Final Inspection process . Three - sigma limits Shewhart control chart for attribute data is potential candidate with superior properties. However, conventional Shewhart chart used normal approximation for binomial data is inadequate in the case of very low defect rate because of skewness of dist ribution. Many studies have been developed to overcome this problem. There are four majors of modified control charts: transforming data approaches; Optimizing the control limit value; Count Chart with geometric distribution (CCC or CPC chart); EWMA and CUSUM chart for detect small and moderate shift. Three selected control charts that are suitable for Final Inspection process had been evaluated by actual data. The numerical results show significant improvement in reducing the false alarm of Quesenberry Chart and Chen Chart compare to standardized Shewhart chart. CUSUM Arcsine control chart also demonstrate to detect small shift and moderate shift (one sigma typically). Nevertheless, all these chart still not capable to detect the process improvement since np value is too small to have positive Lower Control Limit. The Chen chart that approaches close to desired Average Run Length is selected for online monitoring. The CUSUM Arcsine chart that can detect small and moderate shift can be used as off - line cont rol chart that is reviewed by engineers.
Year2013
Corresponding Series Added EntryAsian Institute of Technology. Thesis;no. ISE-13-18
TypeThesis
SchoolSchool of Engineering and Technology (SET)
DepartmentDepartment of Industrial Systems Engineering (DISE)
Academic Program/FoSIndustrial Systems Engineering (ISE)
Chairperson(s)Voratas Kachitvichyanukul
Examination Committee(s)Huynh Trung Luong
Scholarship Donor(s)Intel Products Vietnam Company Limited
DegreeThesis (M.Eng.) - Asian Institute of Technology, 2013


Usage Metrics
View Detail0
Read PDF0
Download PDF0