Multivariate quality control using an integrated artificial neural network scheme: a case study in plastic injection molding industry
Manufacturers using traditional process control charts to monitor their processes often encounter out-of-control signals indicating that the process mean has changed. Most manufacturers are unaware how much these changes in the mean inflate the variance in the process output. In actual, many manu...
Main Author: | |
---|---|
Format: | Thesis |
Language: | English English English |
Published: |
2013
|
Subjects: | |
Online Access: | http://eprints.uthm.edu.my/6622/1/24p%20AFIZ%20AZRY%20MOHD%20KHAIRY.pdf http://eprints.uthm.edu.my/6622/2/AFIZ%20AZRY%20MOHD%20KHAIRY%20COPYRIGHT.pdf http://eprints.uthm.edu.my/6622/3/AFIZ%20AZRY%20MOHD%20KHAIRY%20WATERMARK.pdf |
Summary: | Manufacturers using traditional process control charts to monitor their processes
often encounter out-of-control signals indicating that the process mean has changed.
Most manufacturers are unaware how much these changes in the mean inflate the
variance in the process output. In actual, many manufacturing processes involve two
or more dependent variables and attempting to monitor such variables separately
using univariate SPC charting scheme would increase false alarms and leading to
wrong decision making The problem becomes more complicated when dealing with
small mean shift particularly in identifying the causable variables. In this project
study, advances SPC scheme which applied Artificial Neural Networks that were
designed to enable balanced monitoring and accurate diagnosis were used. Its
performance and effectiveness in actual manufacturing practice were compared with
common use traditional process control chart. Thermoplastic Injection Molding was
selected in this study since most of the industries applied this method to produce low
cost parts. It is important to ensure that good quality parts can be produced according
to requirement. The potential benefit from advance SPC schemes was it able to
performed rapid detection of process disturbance. However the accuracy of mean
shifted diagnosis performance need to improve. |
---|