## Are you sure?

This action might not be possible to undo. Are you sure you want to continue?

Accepting or rejecting a lot (of parts, components, etc.) based on the inspection of a sample drawn from it

Copyright Tapan P Bagchi

1

Acceptance Sampling

● ● ● ●

Accept/reject entire lot based on sample results Created by Dodge and Romig during WWII Not consistent with TQM of Zero Defects Does not estimate the quality of the lot

A “Lot” of goods

Copyright Tapan P Bagchi

2

**What is acceptance sampling?
**

Lot Acceptance Sampling A SQC technique, where a random sample is taken from a lot, and upon the results of appraising the sample, the lot will either be rejected or accepted A procedure for sentencing incoming batches or lots of items without doing 100% inspection The most widely used sampling plans are given by Military Standard (MIL-STD-105E)

Copyright Tapan P Bagchi 3

**What is acceptance sampling?
**

Purposes

Determine the quality level of an incoming shipment or at the end of production Judge whether quality level is within the level that has been predetermined

**But! Acceptance sampling gives you no idea about the process that is producing those items!
**

Copyright Tapan P Bagchi 4

Acceptance Sampling

Another area of quality control and improvement Closely connected with inspection and testing of product Inspection can occur at many points in a process

Acceptance Sampling: the inspection and classification of a Acceptance Sampling: the inspection and classification of a sample of nits selected at random from a larger batch or lot sample of nits selected at random from a larger batch or lot and ultimate decision about disposition of the lot – and ultimate decision about disposition of the lot – Lot Disposition or Lot Sentencing Lot Disposition or Lot Sentencing

When parts are received When parts are received After production After production

**Two common points of inspection
**

Copyright Tapan P Bagchi

5

Sampling Plans

Acceptance sampling: Form of inspection applied to lots or batches of items before or after a process, to judge conformance with predetermined standards Sampling plans: Plans that specify lot size, sample size, number of samples, and acceptance/rejection criteria Single-sampling Double-sampling Multiple-sampling and Sequential sampling

Copyright Tapan P Bagchi 6

**Types of sampling plans
**

Sampling by attributes vs. sampling by variables Incoming vs. outgoing inspection Rectifying vs. non-rectifying inspection

What is done with nonconforming items found during inspection Defectives may be replaced by good items

**Single, double, multiple and sequential plans
**

Copyright Tapan P Bagchi 7

**How acceptance sampling works
**

Attributes (“go no-go” inspection) Defectives-product acceptability across range Defects-number of defects per unit Variable (continuous measurement) Usually measured by mean and standard deviation

Copyright Tapan P Bagchi

8

**Why use acceptance sampling?
**

Can do either 100% inspection, or inspect a sample of a few items taken from the lot Complete inspection Inspecting each item produced to see if each item meets the level desired Used when defective items would be very detrimental in some way

Copyright Tapan P Bagchi

9

Sampling Terms

percentage of defects at which consumers are willing to accept lots as “good”

Acceptance quality level (AQL): the

**Lot tolerance percent defective (LTPD or RQL): the upper limit on the percentage of
**

defects that a consumer is willing to accept contained defectives exceeding the LTPD will be accepted

**Consumer’s risk: the probability that a lot Producer’s risk: the probability that a lot
**

Copyright Tapan P Bagchi

**containing the acceptable quality level will be rejected
**

10

**Why not 100% inspection?
**

Problems with 100% inspection Generally very expensive Can’t use when product must be destroyed to test it Handling by inspectors can induce defects Inspection may be very tedious so defective items may slip through even 100% inspection

Copyright Tapan P Bagchi

11

**A Lot-by-Lot Sampling Plan
**

N (Lot) n

Count Number Conforming Accept or Reject Lot

**Specify the plan (n, c) given N For a lot size N, determine
**

the sample size n, and the acceptance number c.

**Reject lot if number of defects > c Specify course of action if lot is rejected
**

Copyright Tapan P Bagchi 12

Quality Definitions

Acceptance quality level (AQL) The smallest percentage of defectives that will make the lot definitely acceptable. A quality level that is the base line requirement of the customer RQL or Lot tolerance percent defective (LTPD) Quality level that is unacceptable to the customer

Copyright Tapan P Bagchi

13

**How acceptance sampling works
**

Remember

You are not measuring the quality of the lot, but, you are to sentence the lot to either reject or accept it

**Sampling involves risks:
**

Good product may be rejected Bad product may be accepted Because we inspect only a sample, not the whole lot!

Copyright Tapan P Bagchi 14

Acceptance Sampling

Copyright Tapan P Bagchi

15

**Three approaches for sentencing a lot
**

1. Single sampling 1. Single sampling

• For low cost or low impact material on the subsequent process • From trusted supplier

**3. Double/multiple sampling 3. Double/multiple sampling
**

• when 100% inspection is too costly • Trusted supplier with potential risk

**2. Rectified sampling 2. Rectified sampling
**

• Inspect every item in the lot, then remove the defective units by discarding, reworking, or returning to suppliers • For critical process. Defective input may result high failure cost • From new supplier

Copyright Tapan P Bagchi 16

**Types of Sampling Plan
**

Sampling Plan for Attributes Sampling Plan for Attributes Sampling Plan for Variables Sampling Plan for Variables

1. 2. 3. 4.

Single-sampling plan Double-sampling plan Multiple-sampling plan Sequential sampling plan

**Standard sample size
**

Smaller sample size on average

Copyright Tapan P Bagchi

17

**Advantages and Disadvantages of Sampling
**

SSP’s Advantages Less expensive because of less inspection Disadvantages Risk of accepting a lot of poor quality Risk of rejecting a lot of Works with single sample acceptable quality Protects both producer and consumer Requires planning and documentation Rejection on entire lot motivates quality Require extensive study improvement for on customer’s suppliers requirement

Copyright Tapan P Bagchi 18

Lot Formation

1.

**Lots should be such that …
**

produced on the same machines, by same operators, from common raw materials, at approximately the same time period

2.

**Larger lots are better than smaller lots
**

These are more representative of overall quality

3.

Lots should be conformable to the material handling systems and personnel

Copyright Tapan P Bagchi

19

Random Sampling

Inspected units should be selected at random Inspected units should represent all items in the lot Potential bad stories:

Sampled from the front, or top of pile Did not randomize Did not stratify the lot and did not sample from each stratum

Copyright Tapan P Bagchi

20

**The Single Sampling Plan
**

The most common and easiest plan to use but not most efficient in terms of average number of samples needed One sample drawn from the lot and 100% inspected Single sampling plan

**N = lot size n = sample size (randomized) c = acceptance number d = number of defective items in sample
**

Rule: If d ≤ c, accept lot; else reject the lot

Copyright Tapan P Bagchi

21

Take a randomized sample of size n from the lot N Inspect all items in the sample Defectives found = d

The Single Sampling procedure

d≤c? No Reject lot Return lot to supplier

Yes Accept lot

**Do 100% inspection
**

Copyright Tapan P Bagchi 22

**Single-Sampling Plans for Attributes
**

A lot of size N has been submitted for inspection Sample size n Acceptance number of defective c Lot sentencing is based on one sample of size n

Example: Example: N = 10,000 N = 10,000 n= 89 n= 89 c=2 c=2

N = 10,000

n = 89

Copyright Tapan P Bagchi

**Sample exactly n parts and inspect every part
**

23

**Producer’s & Consumer’s Risks due to mistaken sentencing of lot TYPE I ERROR = P(reject good lot)
**

α or Producer’s risk 5% is common

**TYPE II ERROR = P(accept bad lot)
**

β or Consumer’s risk 10% is typical value

Copyright Tapan P Bagchi

24

**Acceptance sampling contd.
**

Producer’s risk

Risk associated with a lot of acceptable quality rejected

Alpha α

= Prob (committing Type I error) = P(rejecting lot at AQL quality level) = producers risk

Copyright Tapan P Bagchi

25

**Acceptance sampling contd.
**

Consumer’s risk

Receive shipment, assume good quality, actually bad quality

Beta β

= Prob (committing Type II error) = Prob (accepting a lot at RQL quality level) = consumers risk

**The Operating Characteristic (OC) curve for a sampling plan quantifies these risks
**

Copyright Tapan P Bagchi 26

Take a randomized sample of size n from the lot of unknown quality p Inspect all items in the sample Defectives found = d

The Single Sampling procedure

d≤c? No Reject lot Return lot to supplier

Yes Accept lot

**Do 100% inspection
**

Copyright Tapan P Bagchi 27

**The Operating Characteristic or OC Curve
**

The OC curve indicates a sampling plan’s ability to discriminate between good and bad lots

Copyright Tapan P Bagchi

28

**What is the Operating Characteristic (OC) Curve?
**

It is a graph of the % defective (p) in a lot or batch vs. the probability that the sampling plan will accept the lot Shows probability of lot acceptance Pa as function of lot quality level (p) It is based on the sampling plan Curve indicates discriminating power of the plan Aids in selection of plans that are effective in reducing risk Helps to keep the high cost of inspection down

Copyright Tapan P Bagchi

29

**OC Definitions on the Curve
**

100% 90% Probability of Accepting Lot 75%

α = 0.10

50%

β = 0.10

Go

AQL

od

Indifferent .06 .09

.03

Copyright Tapan P Bagchi

RQL Bad

30

25%

Lot Quality (Fraction Defective)

**OC Curve Terms
**

Acceptable Quality Level (AQL) Percentage of defective items a customer is willing to accept from you (a property of mfg. process) Lot Tolerance Percent Defective (LTPD) Upper limit on the percentage of defects a customer is willing to accept ( a property of the consumer) Average Outgoing Quality (AOQ) Average of rejected lots and accepted lots Average Outgoing Quality Limit (AOQL) Maximum AOQ for a range of fractions defective

Copyright Tapan P Bagchi

31

**The Perfect OC Curve
**

100% Probability of Accepting Lot

75%

This curve distinguishes perfectly between good and bad lots.

50% What would allow you to achieve a curve like this?

25%

.03

.06

.09

Copyright Tapan P Bagchi 32

Lot Quality (Fraction Defective)

OC Curves

100% Probability of Accepting Lot OC Curves come in various shapes depending on the sample size and risk of α and β errors

This curve is more discriminating This curve is less discriminating

75%

50%

25%

**.03 .06 .09 Lot Quality (Fraction Defective)
**

Copyright Tapan P Bagchi 33

OC Curve and its interpretation

The oc-curve enables us to evaluate the probability of acceptance (Pa) for any true lot quality level-on a what-if basis. This way, one can design sampling plans that perform the way one wants. The OC curve is Interpreted as follows: If the lot quality is 0.093 fraction defective, then Pa, is 0.05. If the lot quality is 0.018 fraction defective, then Pa, is 0.95.

Copyright Tapan P Bagchi 34

**Operating Characteristic Curve
**

1.00

α = 0.05 Probability of acceptance, Pa

{

0.80

**OC curve for n and c
**

0.60

0.40

0.20

β = 0.10

{

0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 0.20

AQL

Proportion defective p

Copyright Tapan P Bagchi

LTPD

35

**OC Curve helps visualize producer’s and consumer’s points
**

Type I and Type II decision errors correspond logically to the two decision points on the oc-curve. Type I error -- Wrongful Rejection A type I error is associated with the producer's point -- to reject when the true value of the quality characteristic is AQL. The risk of rejecting an AQL lot is the producer's risk (α = alpha risk) Type II error -- Wrongful acceptance A Type II error is to accept when the true value of the quality characteristic is RQL -- at the consumer's point. The risk of accepting a lot, if it is an RQL lot, is the consumer's risk (ß = beta risk).

Copyright Tapan P Bagchi 36

Decision Criteria

Probability of accepting lot

1.00

Ideal

Not very discriminating

“Good”

“Bad”

0

**Lot quality (fraction defective)
**

Copyright Tapan P Bagchi 37

**Consumer’s and Producer’s Risk
**

1 α = .10 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 “Good” Indifferent 0.1 β = .10 0 0 .05 .10

AQL

Probability of accepting lot

LTPD “Bad”

.15

.20

.25

**Lot quality (fraction defective)
**

Copyright Tapan P Bagchi 38

**Example: QC Curve for n = 10, c = 1
**

1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0

.9139 .7361 .5443 .3758 .2440 .1493 .0860

Probability of acceptance

.10

.20

Copyright Tapan P Bagchi

.30

.40

.50

39

Fraction defective in lot

Types of OC Curves

Type A Gives the probability of acceptance for an individual lot coming from finite production Type B Give the probability of acceptance for lots coming from a continuous process or infinite size lot

Copyright Tapan P Bagchi

40

**OC Curve Calculation
**

The Ways of Calculating OC Curves

Binomial distribution (Type B) Hypergeometric distribution (Type A)

**Pa = P(r defectives found in a sample of n)
**

Poisson formula

P(r) = ( (np)r e-np)/ r!

Larson nomogram

Copyright Tapan P Bagchi

41

**OC Curve Calculation by Poisson distribution
**

A Poisson formula can be used

P(r) = ((np)r e-np) /r! = Prob(exactly r defectives in n) Poisson is a limit Limitations of using Poisson

**n ≤ N/10 total batch Little faith in Poisson probability calculation when n is quite small and p quite large.
**

For Poisson, Pa = P(r ≤ c)

Copyright Tapan P Bagchi

42

For us, Pa = P(r ≤ c)

Copyright Tapan P Bagchi

p

43

**OC Curve Calculation by Binomial Distribution
**

Note that we cannot always use the binomial distribution because

Binomials are based on constant probabilities

N may not be infinite p changes as items are drawn from the lot

Copyright Tapan P Bagchi

44

**OC Curve by Binomial Formula
**

Pa .998 .980 p (%) .01 .02 .03 .04 .05 .06 .07 .08 .09 .10 .11 .12

45

Using this formula with n = 52 and c=3 and p = .01, .02, ...,.12 we find data values as shown on the right. This givens the plot shown below.

.930 .845 .739 .620 .502 .394 .300 .223 .162 .115

Copyright Tapan P Bagchi

**The Ideal OC Curve
**

●

Ideal curve would be perfectly perpendicular from 0 to 100% for a fraction defective = AQL It will accept every lot with p ≤ AQL and reject every lot with p > AQL

Pa 1.0

●

0.0 AQL

Copyright Tapan P Bagchi

p

46

**Guidelines for choosing Producer’s and Consumer’s decision points (AQL and RQL)
**

To choose the Producer's Point (AQL) in practice Lots at the producer's point quality level (AQL) should be accepted most of the time. Define AQL accordingly. Take into account historical quality levels and the consequence of returning a lot to the producer at AQL quality level. Choose the producers risk of rejecting a lot that is of AQL quality. Typical: α = 0.05. To choose the Consumer's Point (RQL) in practice Lots at the consumer's point quality level (RQL) should be rejected most of the time. Take into account the economic consequence of the consumer’s accepting a lot at RQL quality level and define RQL accordingly. Choose the consumers risk of accepting a lot that is of RQL quality. Typical: β = 0.05.

Copyright Tapan P Bagchi 47

**The OC Curve by binomial formula
**

Operating characteristic (OC) curve

Curve plots the probability of accepting the lot (Pa) versus the lot fraction defective (p)

Copyright Tapan P Bagchi

48

The OC Curve for Single SP

Copyright Tapan P Bagchi

49

Ideal OC curve

Theoretically , it can be achieved by 100% inspection of the lot Due to sampling, ideal OC curve cannot be perfectly achieved. If sample size (n) is large, the OC curve shape will approach the ideal OC curve

Copyright Tapan P Bagchi 50

Properties of OC Curves

The acceptance number c and sample size n are most important factors in defining the OC curve Decreasing the acceptance number (c) is preferred over increasing sample size (n) The larger the sample size n the steeper is the OC curve (i.e., it becomes more discriminating between good and bad lots)

Copyright Tapan P Bagchi

51

Effects of n on OC curves

As sample size (n)

movement of OC

increases, the OC curve shape will approach the ideal OC curve ~ more discriminating Note that c is kept in proportional to n

Copyright Tapan P Bagchi

52

Effects of c on OC curves

Increasing acceptance number

movement of OC

(c) does not significantly change the shape of OC But Producer’s risk (β) increases

Copyright Tapan P Bagchi

53

OC curve

Acceptable quality level (AQL) The poorest quality level for the supplier’s process that a consumer would consider to be acceptable A property of the supplier’s manufacturing process, not a property of the sampling plan Lot tolerance percent defective (LTPD) The protection obtained for individual lots of poor quality Also called rejectable quality level (RQL) LTPD is a level of lot quality specified by consumer, not a characteristic of the sampling plan Sampling plans can be designed to have specified performance at the AQL and the LTPD points

Copyright Tapan P Bagchi 54

Type-A and Type-B OC Curves

Copyright Tapan P Bagchi

55

**Poisson distribution for Defects
**

Poisson parameter: λ = np P(r) = (np)r e-np/r! = Prob(exactly r defectives in n) This formula may be used to formulate equations involving AQL,RQL, α and β to given (n, c). We can use Poisson tables to approximately solve these equations. Poisson can approximate binomial probabilities if n is large and p small. Q. If we sample 50 items from a large lot, what is the probability that 2 are defective if the defect rate (p) = .02? What is the probability that no more than 3 defects are found out of the 50?

Copyright Tapan P Bagchi

56

Hypergeometric Distribution

Hypergeometric formula:

⎛ n − r ⎞⎛ r ⎞ ⎜ ⎜ N − M ⎟⎜ M ⎟ ⎟⎜ ⎟ ⎝ ⎠⎝ ⎠ P(r ) = ⎛n⎞ ⎜ ⎟ ⎜N⎟ ⎝ ⎠

r defectives in sample size n when M defectives are in N. This distribution is used when sampling from a small population. It is used when the lot size is not significantly greater than the sample size. (Can’t assume here each new part picked is unaffected by the earlier samples drawn). Q. A lot of 20 tires contains 5 defective ones (i.e., p = 0.25). If an inspector randomly samples 4 items, what is the probability of 3 defective ones?

Copyright Tapan P Bagchi 57

Properties of OC Curves for finite size lots

Copyright Tapan P Bagchi

58

Properties of OC Curves

If the acceptance level c is changed, the shape of the curve will change. All curves permit the same fraction of sample to be nonconforming.

Copyright Tapan P Bagchi

59

“Rectifying” Inspection

When the lot is rejected, it requires corrective action: 100% inspection and replacement of defectives by good parts

Rectified

Passed

Copyright Tapan P Bagchi

60

Average Outgoing Quality (AOQ)

Expected proportion of defective items passed to customer

**Pa p ( N − n) AOQ with rectifying inspection = N
**

Average outgoing quality limit (AOQL) is

The “maximum” point on AOQ curve

Copyright Tapan P Bagchi

61

**Average outgoing quality AOQ
**

Given the lot size of N, sample size of n, and fraction defective of p n items in the sample that, after inspection, contain no defectives, because all discovered defectives are replaced N – n items that, if the lot is rejected, also contain no defectives N – n items that, if the lot is Copyright accepted p(N-n) defectives Tapan P Bagchi

If N >> n, then

AOQ = Pa p

62

Average Quality

Average outgoing quality (AOQ):

Average of inspected lots (100%) and uninspected lots

**AOQ = Pa× p⎜⎜⎜ N − n ⎟⎟⎟ ⎠ ⎝ N
**

⎛ ⎞

Pa = Probability of accepting lot p = Fraction defective N = Lot size n = Sample size

Copyright Tapan P Bagchi

63

**AOQLimit (AOQL) Estimation
**

0 0.05 0.1 0.1 0.15 0.08 0.2 0.06 0.25 0.04 0.3 0.35 0.02 0.4

0 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4

AOQ (Fraction defective out)

0 0.046 0.074 0.082 0.075 0.061 0.045 0.03 0.019

Approximate AOQL = .082

Incoming fraction defective

Copyright Tapan P Bagchi

64

AOQ Curve

0.015

**AOQL Average Outgoing Quality
**

0.010

0.005

0.01 0.02

0.03

0.04

0.05

0.06

0.07

0.08

0.09

0.10

AQL

**LTPD (Incoming) Percent Defective
**

Copyright Tapan P Bagchi 65

Average outgoing quality AOQ

Copyright Tapan P Bagchi

66

**Double Sampling Plans
**

Take small initial sample

If # defectives < lower limit, accept If # defectives > upper limit, reject If # defectives between limits, take second sample

**Accept or reject lot based on 2 samples Less inspection than in single-sampling
**

Copyright Tapan P Bagchi 67

**Multiple Sampling Plans
**

Advantage: Uses smaller sample sizes Take initial sample If # defectives < lower limit, accept If # defectives > upper limit, reject If # defectives between limits, re-sample Continue sampling until accept or reject lot based on all sample data

Copyright Tapan P Bagchi 68

Sequential Sampling

The ultimate extension of multiple sampling Items are selected from a lot one at a time After inspection of each sample a decision is made to accept the lot, reject the lot, or to select another item In Skip Lot Sampling only a fraction of the lots submitted are inspected

Copyright Tapan P Bagchi 69

**Sequential Sampling Plans
**

Sequential Analysis is the technique by which we build up our sample one item at a time, and after inspecting each item, ask ourselves: "Can we be sure enough to accept or reject this batch on the information so far collected?“ Seq SP offers Its value is in enabling reliable conclusions to be wrung from a minimum of data. This was deemed sufficient to require that it be classified "Restricted " within the meaning of the Espionage Act during the war of 1939-45. Used for lot-by-lot acceptance plans for applications needing the minimum possible average sample numbers. The SPR type sequential sampling plan is also known as Columbia sequential sampling because the methodology was developed at Columbia University.

Copyright Tapan P Bagchi 70

Example of Sequential Decision Diagram

Results of sequential sampling data is plotted on the diagram sequentially as it is collected. The accept or reject decision is made when sample data plots on or over an acceptance line or rejection line. The stair step pattern is caused by the discrete nature of the sample size and the number of defectives in the n-table upon which this diagram is based.

Copyright Tapan P Bagchi 71

Sequential Sampling: Acceptance and Rejection Limit lines

Copyright Tapan P Bagchi

72

**Equations for the Limit Lines
**

The equations for the two limit lines are functions of the parameters p1, , p2, and .

where

Copyright Tapan P Bagchi

73

OC-Curve for Count Data in Sequential Sampling—similar to other sampling plans

Copyright Tapan P Bagchi

74

**Choosing A Sampling Method: Single, Double or Sequential?
**

It is an economic decision Single sampling plans high sampling costs Double/Multiple sampling plans

low sampling costs More complex to implement

Copyright Tapan P Bagchi

75

**Designing a Sampling Plan
**

The plan must protect both the consumer (the buyer) and the supplier (the seller)

Copyright Tapan P Bagchi

76

Designing The Single Sampling plan

Take a randomized sample of size n from the lot of unknown quality p

Inspect all items in the sample Defectives found = d

**d≤c? No Reject lot Return lot to supplier
**

Copyright Tapan P Bagchi

Yes Accept lot

**Do 100% inspection
**

77

Sampling Plan Design by Binomial Distribution Binomial distribution: P(x defectives in n) = [n!/(x!(n-x))!]px(1- p)n-x Recall n!/(x!(n-x))! = ways to choose x in n Q. If 4 samples (items) are chosen from a population with a defect rate = .1, what is the probability that

a) b)

exactly 1 out of 4 is defective? at most 1 out of 4 is defective?

Copyright Tapan P Bagchi

78

**Designing a single-sampling plan with a specified OC curve
**

To construct sampling plan such that

Probability of acceptance for lots with fraction defective p1

Two simultaneous conditions that the (n, c) plan should meet

**Probability of acceptance for lots with fraction defective p2
**

Copyright Tapan P Bagchi 79

Solving for (n, c)

To design a single sampling plan we need two points. Typically these are p1 = AQL, p2 = LTPD and , are the Producer's Risk (Type I error) and Consumer's Risk (Type II error), respectively. By binomial formulas, n and c are the solution to

These two simultaneous equations are nonlinear so there is no simple, direct solution. The Larson nomogram can help us here.

Copyright Tapan P Bagchi 80

**SSP Design by Larson Nomogram
**

●

●

●

**Applies to single sampling plan Based on binomial distribution Uses
**

1-α = Pa at AQL β = Pa at RQL

●

**Can produce OC curve
**

81

Copyright Tapan P Bagchi

Two simultaneous conditions that the (n, c) plan will meet

Designing a single sampling plan by Larson’s Binomial Nomogram Inputs are AQL, α, RQL and β. Outputs are n and c

Copyright Tapan P Bagchi

82

**The MIL-STD-105E approach
**

A Query from a Practitioner: Selecting AQL (acceptable quality levels) “I'd like some guidance on selecting an acceptable quality level and inspection levels when using sampling procedures and tables. For example, when I use MIL-STD-105E, how do I to decide when I should use GI, GII or S2, S4?” -- Confused in Columbus, Ohio W. Edwards Deming observed that the main purpose of MIL-STD-105 was to beat the vendor over the head. "You cannot improve the quality in the process stream using this approach," cautions Don Wheeler, author of Understanding Statistical Process Control (SPC Press, 1992). "Neither can you successfully filter out the bad stuff. About the only place that this procedure will help is in trying to determine which batches have already been screened and which batches are raw, unscreened, run-of-the-mill bad stuff from your supplier. I taught these techniques for years but have repented of this error in judgment. The only appropriate levels of inspection are all or none. Anything else is just playing roulette with the product."

Copyright Tapan P Bagchi 83

**MIL-STD-105E
**

Original version (MIL STD 105A) issued in 1950 as tables; Last version (MIL STD 105E) in 1989; ISO adopted it as ISO 2859 Plan covers sampling by attributes for given lot size (N) and acceptable quality level (AQL). Prescribes sample size n, acceptance number c, and rejection number r Standard included three types of inspection—normal, tightened and reduced and gives switching rules Plans assure producer’s risk (α) of 0.01 – 0.1. The only way to control the consumer’s risk (β) is to change inspection level

Copyright Tapan P Bagchi

84

**Military Standard 105E
**

Sampling procedure for inspection by attributes developed during World War II and is the most widely used acceptance-sampling system for attributes in the world today A collection of sampling schemes; therefore an acceptancesampling system Provides for three types of sampling: single, double, and multiple Primary focal point is the acceptable quality level (AQL)

Different AQLs may be designated for different types of defects: critical, major, and minor Generally specified in contract or by authority responsible for sampling

**Sample size is determined by lot size and by choice of inspection level
**

Copyright Tapan P Bagchi 85

**AQL Acceptance Sampling by Attributes by MILSTD 105E
**

Determine lot size N and AQL for the task at hand Decide the type of sampling—single, double, etc. Decide the state of inspection (e.g. normal) Decide the type of inspection level (usually II) Look at Table K for sample sizes Look at the sampling plans tables (e.g. Table IIA) Read n, Ac and Re numbers

Copyright Tapan P Bagchi

86

**Military Standard 105E Procedure
**

1. 2. 3. 4. 5. 6. 7.

Choose the AQL Choose the inspection level Determine the lot size Fine the appropriate sample size code letter (from the MIL-STD 105E Table) Determine the appropriate type of sampling plan to use (single) Enter the appropriate table to find the type of plan to be used Determine the corresponding normal and reduced inspection plans to be used when required

Copyright Tapan P Bagchi 87

MIL STD 105E Inspection levels and Sample Size codes

Copyright Tapan P Bagchi

88

OC Curves under different Inspection Conditions

Copyright Tapan P Bagchi

89

MIL STD 105E “Normal” Sample Size n and defective limit c

Copyright Tapan P Bagchi

90

The OC curve of a MIL STD 105E Sampling Plan

Copyright Tapan P Bagchi

91

MIL-STD-105E Normal Inspection

Copyright Tapan P Bagchi

92

MIL-STD-105E Reduced Inspection

Copyright Tapan P Bagchi

93

MIL-STD-105E Tightened Inspection

Copyright Tapan P Bagchi

94

**Military Standard 105E
**

Sampling rules for switching between normal, tightened and reduced inspection

Copyright Tapan P Bagchi

95

OC Curves under different MIL Inspection Conditions

Copyright Tapan P Bagchi

96

**Discussion of Military Standard
**

Several points about MIL STD 105E should be emphasized: MIL STD 105E is AQL-oriented, no concern for β Not all possible sample sizes are possible (2,3,5,8,13,20,32,50, etc.) Sample sizes are related to lot sizes Switching rules are subject to criticism for both misswitching between inspection plans and discontinuation even though there has been no actual quality deterioration MIL STD 105E has been replaced by MIL STD 1916 by the US Department of Defense as of October 2008 http://assist.daps.dla.mil/

Copyright Tapan P Bagchi 97

**How/When would you use Acceptance Sampling?
**

Advantages of acceptance sampling

Less handling damages Fewer inspectors to put on payroll 100% inspection costs are to high 100% testing would take to long

Acceptance sampling has some disadvantages

Risk included in chance of bad lot “acceptance” and good lot “rejection” Sample taken provides less information than 100% inspection

Copyright Tapan P Bagchi 98

Summary

Many basic terms you need to know to be able to understand acceptance sampling

SPC, Accept a lot, Reject a lot, Complete Inspection, AQL, LTPD, Sampling Plans, Producer’s Risk, Consumer’s Risk, Alpha, Beta, Defect, Defectives, Attributes, Variables, ASN, ATI.

Copyright Tapan P Bagchi

99

**Some Key Definitions and Terms
**

Reference: NIST Engineering Statistics Handbook

Acceptable Quality Level (AQL): A percent defective that is the base line requirement for the quality of the producer's product. The producer would like the buyer to design a sampling plan such that there is a high probability of accepting a lot that has a defect level less than or equal to the AQL. Lot Tolerance Percent Defective (LTPD) also called RQL (Rejection Quality Level): A designated high defect level that would be unacceptable to the consumer. The consumer would like the sampling plan to have a low probability of accepting a lot with a defect level as high as the LTPD.

Copyright Tapan P Bagchi 100

Type I Error (Producer's Risk): The probability for a given (n, c) sampling plan of rejecting a lot that has a defect level equal to AQL. The producer suffers when this occurs, because α lot with acceptable quality was rejected. The symbol a is commonly used for the Type I error and typical values for α to range from 0.2 to 0.01. Type II Error (Consumer's Risk): This is the probability, for a given (n, c) sampling plan, of accepting a lot with a defect level equal to the LTPD. The consumer suffers when this occurs, because a lot with unacceptable quality was accepted. The symbol β is commonly used for the Type II error and typical values range from 0.2 to 0.01.

Copyright Tapan P Bagchi 101

Average Outgoing Quality (AOQ): A common procedure, when sampling and testing is nondestructive, is to 100% inspect rejected lots and replace all defectives with good units. In this case, all rejected lots are made perfect and the only defects left are those in lots that were accepted. AOQ's refer to the long term defect level for this combined LASP and 100% inspection of rejected lots process. If all lots come in with a defect level of exactly p, and the OC curve for the chosen (n, c) LASP indicates a probability Pa of accepting such a lot, over the long run the AOQ can easily be shown to be: where N is the lot size. Copyright Tapan P Bagchi

102

Average Outgoing Quality Level (AOQL): A plot of the AOQ (Y-axis) versus the incoming lot p (X-axis) will start at 0 for p = 0, and return to 0 for p = 1 (where every lot is 100% inspected and rectified). In between, it will rise to a maximum. This maximum, which is the worst possible long term AOQ, is called the AOQL. Average Total Inspection (ATI): When rejected lots are 100% inspected (rectifying inspection), it is easy to calculate the ATI if lots come consistently with a defect level of p. For a sampling plan (n, c) with a probability Pa of accepting a lot with defect level p, we have ATI = n + (1 - Pa) (N - n) where N is the lot size.

Copyright Tapan P Bagchi 103

Average Sample Number (ASN): For a single sampling plan (n, c) we know each and every lot has a sample of size n taken and inspected or tested. For double, multiple and sequential plans, the amount of sampling varies depending on the number of defects observed. For any given double, multiple or sequential plan, a long term ASN can be calculated assuming all lots come in with a defect level of p. A plot of the ASN, versus the incoming defect level p, describes the sampling efficiency of a given lot sampling scheme.

Copyright Tapan P Bagchi 104

Useful links

http://www.bioss.sari.ac.uk/smart/unix/mseqacc/slides/frames.htm Acceptance Sampling Overview Text and Audio http://iew3.technion.ac.il/sqconline/milstd105.html Online calculator for acceptance sampling plans http://www.stats.uwo.ca/courses/ss316b/2002/accept_02red.pdf Acceptance sampling mathematical background http://assist.daps.dla.mil/ The new US military sampling procedure

Copyright Tapan P Bagchi

105

- Quality Control
- Ashutosh 2
- Synopsis P P BANK
- Algorithm 368
- PG16 Sampling Plan - 1.0
- Report on Atm
- acceptance sampling
- Sqc New Data
- Chapter 2 Sad
- Work Sampling
- Case 1 Problem on Defect Sampling
- lecture4
- Iso-2859-Sampling-Standard
- Asistensi StatProb 2014 PraUTS
- etetst
- T-test
- 2 Marks
- 19-1-90-1-10-20151112
- MB0040 û STATISTICS FOR MANAGEMENT Assign Set-2
- Journal.pone.0174450
- Setting Standard Times
- Private Provision of Infrastructure
- C 02 Patton1990 Esantionarea Calitativa
- Unit 1_Into to Business Research
- MB0040 – STATISTICS FOR MANAGEMENT Assign Set-2
- Untitled
- SEM 1 MB0040 2 Statistics for Management
- researchproposal-customerstisfaction-150915151446-lva1-app6892.pdf
- Probability Vs Non-Probability Sampling
- CONSORT 2001 Checklist

Sign up to vote on this title

UsefulNot usefulClose Dialog## Are you sure?

This action might not be possible to undo. Are you sure you want to continue?

Loading