QA- lec 5 Flashcards
1
Q
Control charts
A
- Shewhart plots
- Cusum plots
- Cumulative sum chart

2
Q
Random and systematic error
Dot plot

A

3
Q
Standard deviation of the entire population
A

4
Q
Standard deviation of the population based on a sample
A

5
Q
Real mean
A
- Tank of product has real mean (u) and an actual SD (sigma)
- Samples from tank have a mean and SD which is an estimate of the true mean
- In order to use the sample results to estimate the tank (population) deviation (s) use n-1
6
Q
Properties of the normal distribution
A

7
Q
Confidence intervals
A
- n= sample size- replicate in sample
- Confidence limits rounded to
- 95% CI= u= x + 2s / Square root of n
- (x= true mean) 2s= pop Sd n=replicatres in sample
- 99.7= same but with 3s

8
Q
Shewhart charts
A
- Over a long period of time we can assess the population deviation (sigma) (process capacity)
- n= sample size
- u= target mean
- CI rounded to
- warning lines
- Action lines

9
Q
Shewhart chart for mean values
A

10
Q
Areas under the normal curve
A
- 1 in 45 obeservations normally fall above (below) 2sigma
- 1 in 750 normally fall above and below 3sigma

11
Q
Control charts Shewhart plots

A

12
Q
Controls using shewhart plots
A
- Out of control examples
- Two successive points outside warning line
- One point outside action line
- Eight successive points lying one side of the target value line
- Establish other warning levels to enable correction before production stops
- E.G. six successive points lying one side of the target value line
13
Q
Cusum plot
A
- ARL- average run length
- This is the average number of measurements required to deyect a change in the process has changed
- This period of time can be length in shewhart plots
- This can be reduced by using cusum charts

14
Q
Shewhart v cusum plot of data

A

15
Q
Use of V-mask Initial data
A
- If points fall above or below the mask then the results are said to be out of control

16
Q
Data integrity
A
- Current regulatory hot topic
- The extent to which all data are complete, consistent and accurate throughoutout the data lifecycle
- From initial data collection and recording, through processing (including)