Chapter 17

Process Improvement Using Control Charts

Copyright © 2014 by The McGraw-Hill Companies, Inc. All rights reserved.McGraw-Hill/Irwin

Process Improvement Using Control Charts

17.1 Quality: Meaning and Historical Perspective

17.2 Statistical Process Control and Causes of Variation

17.3 Sampling a Process, Rational Subgrouping and Control Charts

17.4 and R Charts

17-2

Process Improvement Using Control Charts Continued

17.5 Comparison of a Process with Specifications: Capability Studies

17.6 Charts for Fraction Nonconforming17.7 Cause and Effect, Defect Concentration

Diagrams (Optional)

17-3

17.1Quality: Meaning and Historical Perspective

Quality◦Fitness for use◦Extent to which customer expectations are met

Types of quality◦Quality of design◦Quality of conformance◦Quality of performance

LO17-1: Discuss the principles and importance of quality improvement.

17-4

History of the Quality Movement1924 Statistical Quality Control/Control Charts,

Shewart/Bell Telephone1920’s Statistical Acceptance Sampling, Bell

Telephone1946 American Society for Quality Control created1950 W. Edwards Deming introduces statistical

quality control in Japan1951 Deming Prize established in Japan1980’s Total Quality Management (TQM)1988 Malcolm Baldrige National Quality Awards

established1990’s ISO 9000, international quality standards

adopted

LO17-1

17-5

17.2 Statistical Process Control and Causes of Process Variation

Historical inspection approach◦ Inspection of output◦Action on output

Scrap, rework, downgrade (expensive!)Statistical process control◦Monitor and study process variation◦Goal: Continuous process improvement◦Preventing by quality through process

improvement

LO17-2: Distinguish between common causes and assignable causes of process variation.

17-6

Causes of Process Variation

Common causes◦Typical (random) variation inherent in process

design◦Process in statistical control

Assignable causes◦Unusual process variation◦ Intermittent or permanent process changes◦Not common to all process observations◦Process not in statistical control

LO17-2

17-7

17.3 Sampling a Process and Rational Subgrouping and Control Charts

Must decide which process variables to study◦Best to study a quantitative variable

This means we are employing measurement dataWe will take a series of samples over time◦Usually called subgroups◦Usually of size two to six◦Usually observed over a short period of time

Want to observe often enough to detect important process changes

LO17-3: Sample a process by using rational subgrouping.

17-8

Control Charts

A control chart employs a center line, upper control limit and lower control limit

The center line represents average performance

The upper and lower control limits are established so that when in control almost all plot points will be between the limits

LO17-3

17-9

17.4 and R Charts

and R charts are the most commonly used control charts for measurement data◦ chart plots subgroup means versus time◦R chart plots subgroup range versus time

chart monitors the process meanR chart monitors the amount of variabilityThese two charts must be used together

LO17-4: Use and R charts to establish process control.

17-10

Pattern Analysis

An observation beyond the control limits indicates the presence of an assignable cause

Other types of patterns also indicate the presence of an assignable cause

These patterns are more easily described in terms of control chart zones◦A, B, C

LO17-5: Detect the presence of assignable causes through pattern analysis.

17-11

17.5 Comparison of a Process with Specifications: Capability Studies

Natural tolerance limits for a normally distributed process in statistical control will contain about 99.73 percent of the process observations and is given by

If the natural tolerance limits are inside the process specification limits, we say that the process is capable of meeting specifications

222

3,33dRx

dRx

dRx

LO17-6: Decide whether aprocess is capable of meeting specifications.

17-12

17.6 Charts for Fraction Nonconforming

Sometimes we inspect items and simply decide if they conform to standards or not◦Nonconforming: does not meet standards

Defective◦Conforming: meets standards

Use a p chartObserve subgroups of n units over time◦Determine the number nonconforming

LO17-7: Use p charts to monitor process quality.

17-13

17.7 Cause-and-Effect Concentration Diagrams (Optional)

A cause-and-effect diagram for “why tables are not cleared quickly in a restaurant”

Also known as Ishikawa diagrams or fishbone charts

LO17-8: Use diagramsTo discern the causes of quality problems (Optional).

Figure 17.26 17-14