Professional Documents
Culture Documents
IEM 4103 Quality Control & Reliability Analysis IEM 5103 Breakthrough Quality & Reliability
IEM 4103 Quality Control & Reliability Analysis IEM 5103 Breakthrough Quality & Reliability
Introduction
Control Chart for Fraction Nonconforming
𝑝-chart
𝑛𝑝-chart
Control charts for Nonconformities (Defects)
𝑐-chart
𝑢-chart
Textbook: Chapter 7
2/24/2021
2
6.1 Introduction
Nonconforming: Defective
Conforming: Non-defective
Attributes: Quality characteristics of conforming or nonconforming
Think: What is the difference between attributes and variables?
Fraction nonconforming: The ratio of the number of nonconforming units in a
sample to the total number of units in that sample, to estimate it,
𝐷 # of nonconforming
𝑝Ƹ = =
𝑛 sample size
2/24/2021
3
Attributes vs. Variables
2/24/2021
4
Review: binomial distribution
5
Review: basic model of control charts
2/24/2021
6
6.2 Control Chart for Fraction Nonconforming
2/24/2021
7
Sample fraction nonconforming
𝐷
𝑝Ƹ =
𝑛
The distribution of the random variable 𝑝Ƹ can be obtained from the binomial.
Furthermore, the mean and variance of 𝑝Ƹ are
𝜇𝑝ො = 𝑝
𝑝(1 − 𝑝)
𝜎𝑝2ො =
𝑛
2/24/2021
8
𝑝-chart with known 𝑝
Depending on the values of 𝑝 and 𝑛, sometimes the lower control limit LCL < 0.
In these cases, we customarily set LCL = 0 and assume that the control chart
only has an upper control limit.
2/24/2021
9
𝑝-chart with unknown 𝑝
σ𝑚
𝑖=1 𝐷𝑖 σ𝑚
𝑖=1 𝑝Ƹ 𝑖
𝑝ҧ = =
𝑚𝑛 𝑚
2/24/2021
10
𝑝-chart with unknown 𝑝 (cont’d)
Note: The control limits defined here should be regarded as trial control limits.
2/24/2021
11
Discussion for 𝑝
If the control chart is based on a known or standard value for the fraction
nonconforming 𝑝, then the calculation of trial control limits is generally
unnecessary.
However, one should be cautious when working with a standard value for 𝑝.
Since in practice the true value of 𝑝 would rarely be known with certainty, we
would usually be given a standard value of 𝑝 that represents a desired or target
value for the process fraction nonconforming.
If this is the case, and future samples indicate an out-of-control condition, we
must determine whether the process is out of control at the target 𝑝 but in
control at some other value of 𝑝.
2/24/2021
12
Example
2/24/2021
13
Example (cont’d)
2/24/2021
14
Example (cont’d)
2/24/2021
15
Example (cont’d)
2/24/2021
16
Example (cont’d)
2/24/2021
17
Example (cont’d)
2/24/2021
18
Example (cont’d)
2/24/2021
19
Example (cont’d)
2/24/2021
20
Example (cont’d)
2/24/2021
21
Example (cont’d)
Completed
2/24/2021
22
Performance evaluation
Type I error
𝛼 = 𝑃 𝑝Ƹ ≥ UCL 𝑝0 + 𝑃 𝑝Ƹ ≤ LCL 𝑝0 = 𝑃 𝐷 ≥ 𝑛UCL 𝑝0 + 𝑃 𝐷 ≤ 𝑛LCL 𝑝0
Type II error
𝛽 = 𝑃 𝑝Ƹ < UCL 𝑝1 − 𝑃 𝑝Ƹ ≤ LCL 𝑝1 = 𝑃 𝐷 < 𝑛UCL 𝑝1 − 𝑃 𝐷 ≤ 𝑛LCL 𝑝1
ARL1 = 1/(1 − β)
2/24/2021
23
𝑛𝑝 Control Chart
UCL = 𝑛𝑝 + 3 𝑛𝑝(1 − 𝑝)
Center line = 𝑛𝑝
LCL = 𝑛𝑝 − 3 𝑛𝑝(1 − 𝑝)
24
Example
Set up an 𝑛𝑝 control chart for the orange juice concentrate can process
Use sample 1-30 (table in slides 14)
2/24/2021
25
Example (cont’d)
2/24/2021
26
Properties of 𝑛𝑝 chart
Advantage
𝑛𝑝 chart is a scaling of the vertical axis by the constant 𝑛, provide the same
information as 𝑝 chart
𝑛𝑝 chart needs less calculation
often used when 𝑛 is constant and 𝑝 is small
Limitation
Not easy for interpretation when 𝑛 is varied (UCL, LCL and CL all vary)
Only plot # of nonconforming without considering sample size, hard to take action
2/24/2021
27
6.3 Control Charts for Nonconformities
𝑐 chart
𝑢 chart
2/24/2021
28
Motivation
2/24/2021
29
Review: Poisson distribution
𝑒 −𝑐 𝑐 𝑥
𝑝 𝑥 = 𝑥 = 0,1,2, ⋯
𝑥!
The mean and variance of the distribution are
𝐸 𝑥 = 𝑐 and 𝑉𝑎𝑟(𝑥) = 𝑐
Assumptions
The average occurrence rate 𝑐 (per unit) is a known constant
Occurrences are equally likely to occur within any unit of time/area
Occurrences are statistically independent
2/24/2021
30
𝑐 chart
Center line = 𝑐
LCL = 𝑐 − 3 𝑐
𝑐ҧ =?
2/24/2021
32
Example (cont’d)
2/24/2021
33
Example (cont’d)
Additional data
2/24/2021
34
Example (cont’d)
2/24/2021
35
Discussion for inspection unit
2/24/2021
36
𝑢 chart
There is no reason why the sample size must be restricted to one inspection
unit. In fact, we would often prefer to use several inspection units in the sample,
thereby increasing the area of opportunity for the occurrence of nonconformities.
Let 𝑥 = # of nonconformities in a sample of 𝑛 inspection unit
𝑥~𝑃𝑜𝑖𝑠𝑠𝑜𝑛(𝑐)
37
𝑢 chart
UCL = 𝑢 + 3 𝑢/𝑛
Center line = 𝑢
LCL = 𝑢 − 3 𝑢/𝑛
𝑐𝑖 σ𝑚
𝑖=1 𝑢𝑖
𝑢𝑖 = , 𝑢ത =
𝑛 𝑚
2/24/2021
38
Example
2/24/2021
39
Example (cont’d)
Data
2/24/2021
40
Example
Why LCL = 0?
2/24/2021
41
Thank you!
Any Questions?
2/24/2021
42