Prerequisite(s):
No prerequisites are required.
Text
Book:
Montgomery,
Douglas C., Introduction to Statistical Quality Control, John Wiley
and Sons, 4th ed. (2001)
Course Description
Statistical process monitoring (SPM) is a method of monitoring and
controlling the process variation through statistical analysis. The
purpose is to detect abnormality in the process and identify the causes
behind this variation. This course is about statistical techniques to
monitor and control process and quality variables. The objective is to
give the students the idea about SPM and how it can improve quality and
productivity of a process. Therefore, statistical tools introduced in this
course can be used with various kinds of data from different disciplines.
The course will focus on the statistical process monitoring and quality
control techniques used in science and engineering. The content covers
statistical process monitoring charts for variables and attributes.
Descriptive statistics including mean, standard deviation, variance,
probability distributions will be given. The concept of univariate charts
such as Shewhart, cumulative sum and exponentially weighted moving average
charts will be followed by autocorrelation and crosscorrelation in process
data. The techniques for multivariable processes with correlated data will
be introduced. The definition and guidelines of experimental design and
factorial experiments will be covered.
Course Outline
I.
Introduction
I1. What is Statistical Process Monitoring (SPM)
I2. General Statistics: mean, standard deviation, variance, median,
continuous and
discrete probability
distributions, significance tests, confidence intervals, normality
plot
I3. Tools of Statistical Process Control: Histogram, Pareto chart, cause
and effect diagram,
scatter
diagram, control chart
II.
Univariate Control Charts
II.1. Control Charts for Attributes: Charts for defects and
nonconformities
II.2. Control Charts for Variables: X, Range, and S charts
II.3. CUSUM Charts
II.4. EWMA Charts
II.5. Process Capability Analysis
III.
Autocorrelation in Data
III.1. Definitions
III.2. Diagnosis of Autocorrelation
III.3. Effects of Autocorrelation
III.4. Control Charts for Autocorrelated data
IV.
SPM of Multivariate Process
IV.1. Why do we need Multivariate SPM Techniques?
IV 2. Multivariate SPM Charts
IV.3. Multivariate SPM based on Residuals
IV.4 Principal Component Analysis
V.
Design of Experiments (If time allows)
V.1. What is Experimental Design?
V.2. Experiments with one Factor
V.3. Factorial Experiments
Teaching
Methods
Students
will be assigned homework every two weeks. A group project (two or three
people) will be assigned in the first half of the semester. Groups need to
find a system to which they are going to apply univariate and multivariate
SPM techniques. At the end of the semester, groups will present their work
to the class. Software will be used extensively. Excel, any other
statistical software (Statistica, SPSS, Minitab,...) or Matlab can be used.
Grading
Homework assignments: 20
%
Midterm Exam: 30 %
Term Project: 40%
Presentation: 10%
Matrix
Calculations in Microsoft Excel
(PDF
file)