3.

ACCEPTANCE SAMPLING

3.1 Introduction

Acceptance sampling is the process of evaluating a portion of the

product/material in a lot for the purpose of accepting or rejecting the lot as

either confirming or not confirming to a quality specification.

Generally inspection for acceptance purpose is carried out in two

ways. They are

• 100% inspection and

• Sampling inspection.

In this, “Sampling inspection” is more practical, quick and

economical.

However, the sampling inspection is having certain disadvantages

such as "producer risk" and "consumer risk". In scientific sampling plans,

these risks are quantified and the sampling criteria are adjusted to balance

these risks in the light of economic factors involved.

3.2 Definition

Inspection of raw materials, semi finished products, or finished

products are important part of quality assurance. When inspection is for

the purpose of acceptance or rejection of a product, based on adherence

23
 to standard, the type of inspection procedure employed is called as

acceptance sampling.

3.2.1 Characteristics

It is not a substitute for process control techniques, but it is an audit

tool to improve the output of the drain conferring to requirements.

Adequate process control techniques at the early stages of manufacturing

greatly reduce the use of sampling inspection.

3.2.2 Procedure

In acceptance sampling, a sample is taken from the lot and some

quality characteristic of the units in the sample is inspected. Based on the

information in this sample, a decision is made regarding lot disposition.

3.3 Quality Indices for Acceptance Sampling Plan

3.3.1 Acceptable Quality Level (AQL)

It represents the maximum proportion of defective that the

consumer finds definitely acceptable. It is the fraction defective that can be

tolerated without any serious effect upon further processing or customer

relations.

24
3.3.2 Rejectable Quality Level (RQL)

It represents the proportion of defectives that the consumer finds

definitely unacceptable. It is also called as "Lot Tolerance Percent

Defective". (L.T.P.D.).

3.3.3 Indifference Quality Level (IQL)

It is a quality level somewhere between the A.Q.L and R.Q.L. It is

frequently defined as the quality level having a probability of acceptance of

0.50 for a given sampling plan.

3.3.4 Average Outgoing Quality (AOQ)

It represents the average % defective in the outgoing products after

inspection, including all accepted and all rejected lots that have been 100%

inspected and defectives replaced by non-defectives. The AOQ gives the

excepted quality in the long run. Over a short period, the outgoing quality

may be better or worse than the long run average.

AOQ =Pa.Pi(N-n)/N

Where,

P-i = Fraction defective,

Pa = Probability of acceptance,

N = Lot size and n = Sample size.

25
3.3.5 Average Outgoing Quality Limit (AOQL)

For any given lot size, sample size and acceptance number there is

a maximum value of AOQ beyond which the average fraction of defective

passed forward will not rise, no matter how bad the quality of the lots when

they arrive at inspection, because when incoming quality is perfect

outgoing quality likewise be perfect. However when incoming quality is

very bad, outgoing quality will also be perfect, because sampling will

cause all lots to be rejected, which are subjected to 100% inspection and

rectification.

3.4 Applications of Acceptance Sampling

Acceptance sampling is applicable when

• Testing is destructive.

• The cost of 100% inspection is extremely high.

• 100% inspection is not technologically feasible or would require

much time that would seriously affect the production schedule.

• There are many items to be inspected; the inspection error rate is

sufficiently high that 100% inspection might cause higher

percentage of defective units.

• The vendor has an excellent quality history, and some reduction

from 100% inspection is desired, but the vendor’s process capability

26
 rates is sufficiently low to make no inspection an unsatisfactory

alternative.

• When there are potentially serious product liability risks, and

although the vendor’s process is satisfactory, a program for

continuously monitoring the product is necessary [6].

3.5 Advantages of Acceptance Sampling

It is usually less expensive because there is less inspection.

There is less handling of products, hence reduced damage.

It is applicable to destructive testing also.

Fewer personnel are involved in inspection activity.

It reduces the amount of inspection error.

The rejections of entire lots as opposed to the simple rejection of

defectives often provide a stronger motivation to the vendor for

quality improvements.

3.5.1 Disadvantages of Acceptance Sampling

• There are risks of accepting bad lots and rejecting good lots.

• Less information is generated about the product and the process

that manufactured the product.

27
 • Acceptance sampling requires planning and documentation of the

acceptance sampling procedure.

3.6 Sampling plans

3.6.1 Definition

It is a statement of the sample size and the associated acceptance

or rejection criteria to be used.

3.6.2 Sampling Scheme

A set of procedures consisting of acceptance sampling plans in

which lot sizes, sample sizes, and acceptance or rejection criteria, along

with 100% inspection and sampling are related.

3.6.3 Sampling System

It is a collection of one or more acceptance sampling schemes.

3.6.4 Classification of Sampling Schemes

The major classification is (i) Attributes and (ii) Variables.

The attributes are quality characteristic that are expressed on a “go”,

“no-go” basis. The variables are of quality characteristics that are

measured on a numerical scale.

28
 Another way of classification is (i) Single Sampling Plan (ii) Double

Sampling Plan (iii) Multiple Sampling Plan and (iv) Sequential Sampling

Plan.

3.6.4.1 Single Sampling Plan

It is a lot sentencing procedure in which one sample is selected at

random from the lot and the disposition of the lot is determined based on

the information contained in the sample.

3.6.4.2 Double Sampling Plan

In this procedure after the first sample, a decision based on the

information in that sample is made either to (i) accept the lot (ii) reject the

lot or (iii) take a second sample. If the second sample is taken, the

information from both the first and second sample is combined in order to

reach a decision whether to accept or reject the lot.

3.6.4.3 Multiple Sampling Plan

It is an extension of the double-sampling concept in that more than

two samples may be required in order to reach a decision regarding the

disposition of the lot. Sample sizes are usually smaller than they are in

either single or double sampling.

29
3.6.4.4 Sequential Sampling Plan

It is an extension of multiple sampling system. In this method units

are selected one at a time from a lot and following the inspection of this

unit, a decision is made either to accept the lot or to reject the lot or to take

the next sample.

3.7 Operating Characteristic Curve (OC Curve)

An important measure of the performance of the acceptance

sampling plan is the operating characteristic curve. This curve plots the

probability of accepting the lot versus the lot fraction defective. A sample

O.C. curve for n = 89, c = 2; is shown in Fig. 3.1. The ideal OC curve for

the single sampling plan n=89, c=2; is shown in Fig 3.2.

3.7.1 Specific Points on OC Curve

Acceptable quality level represents the poorest level of quality for

the vendor’s process that the consumer would consider to be acceptable

as a process average.

Lot tolerance percent defective is the poorest level of quality that

the consumer is willing to accept in an individual lot.

30
 P robability O f Acceptance

0.02 0 04 0.06 0 08 0.1 0.12

Fig.3.1: OC Curve
P robability Of Acceptance

0 01 0 015 0.02 0 025

Lot Fraction Defective

Fig.3.2: Ideal OC Curve

This O.C. curve is used to calculate probabilities of acceptance for

an isolated lot of finite lot size. The exact sampling distribution of the

number of defective items in this sample is the hypergeometric distribution.

3.7.3 Type - B OC curve

This O.C. curve is used to calculate probabilities of acceptance for

a large lot or stream of lots. The exact probability distribution of the

number of defective items in this sample is the binomial geometric

distribution.

3.8 Designing a Single Sampling Plan with a specified OC curve

A common approach to the design of an acceptance-sampling plan

is to require that the O.C. curve pass through two designated points.

In general, it does not matter which two points are specified. If we wish to

construct a sampling plan such that the probability of acceptance is (1- a)

for lots with fraction defective P-i and probability of acceptance is (3 for lots

with fraction defective P2.

Assuming the binomial sampling, it is seen that the sample size n

and acceptance number c are the solution to

1-a = £ n1/d1(n-d)1 P^ (1-P1)n’d

32
 p = n1/d1(n-d)1 P2d (1-P2)n'd

These equations are obtained by writing at the two points on the

OC curve using binominal distribution. These equations are non-linear and

there is no simple direct solution.

The binominal nomograph can be used for solving these equations.

In this method, two lines are drawn on the nomograph, one connecting pi

and 1-a, and the other connecting p2 and p. The intersection of these two

lines gives the region of the nomograph, in which desired sampling plan is

located. The main disadvantage of this procedure is it gives several plans

that pass close to the desired points. In addition to the procedure

described above, tabular procedures are available for the same purpose.

3.9 The Dodge - Romig Sampling Plans

H.F.Dodge and the H.G.Romig have developed a set of sampling

inspection tables for lot-by-lot inspection of product by attributes.

These plans were developed in the year 1920 and enjoy extensive

industrial use. Two types of sampling plans are presented in the tables.

One set of tables for LTPD protection and another set of tables that

provide a specified AOQL protection.

33
3.9.1 AOQL Plans

This table give AOQL sampling plans for AOQL values of 0.1%,

0.25%, 0.50%, 0.75%, 1%, 1.5%, 2%, 3%, 4%, 5%, 7% and 10%.

For each of these AOQL values, six classes of values for the process

average are specified. Tables are provided for both single and double

sampling. These plans have been designed so that the average total

inspection (ATI) at the given AOQL and process average is approximately

minimum [5].

3.9.2 LTPD Plans

The Dodge-Romig LTPD tables are designed so that the probability

of lot acceptance at the LTPD is 0.1. Tables are provided for LTPD values

of 0.5%, 1%, 2%, 3%, 4%, 5%, 7% and 10%. Tables are provided

separately for both single and double sampling [5].

3.10 Military Standard 105 D

This standard sampling procedure, for inspection by attributes was

developed during World War II. MILSTD 105D is the most widely used

acceptance-sampling system for attributes in the world today. The original

version of the standard, MIL STD 105 A was released in 1950. Since than,

34
there have been three revisions, the latest version MIL STD 105 D, was

issued in 1963.

MIL STD 105 D is a collection of sampling schemes; therefore, it is

an acceptance sampling system. The standard provides for three types of

sampling: single, double and multiple sampling. For each type of sampling,

a provision is made for using normal, tightened or reduced inspections.

The primary focal point of this acceptance sampling is the AQL.

The sample size used in MIL STD 105 D is determined by the lot

size and by the choice of inspection levels. There are three general

inspection levels and four special inspection levels. For a specified AQL,

inspection levels and lot size, MIL STD 105 D provides a sampling plan.

The procedure is started with normal inspection. It provides a procedure

for switching to tightened and reduced inspection whenever there is an

indication that the incoming material quality has been changed.

3.10.1 Features of MIL STD 105 D

• MIL STD 105 D is a “AQL” oriented. It focuses attention on the

producer’s risk end of the OC curve. The only control over the

discriminatory power of the sampling plan is through the choice of

inspection levels.

35
 • The sample sizes used in MIL STD 105D are 2,3,5,8,13,20,

32,50,80,125, 200,315, 500,800, 1250 and 2000. This indicates

that all sample sizes are not available for use. There are also some

significant gaps such as between 125 and 200 and between200

and 315.

• The sample sizes are related to lot sizes. The sample size will

increase when the lot size increases. But the ratio of sample size to

lot size will decrease rapidly.

• Switching rules from normal to tightened and tightened to normal is

still subject to criticism [6].

3.10.2 Disadvantages of Tabled Sampling Schemes

Apart from certain draw backs particular for the respective tabled

schemes, all the tabled sampling schemes does not offer the flexibility

required for the quality control engineers in designing specific sampling

plans for some specific requirements.

3.11 Conclusion

Acceptance sampling is studied in detail. The statistical method of

determination of sampling plans is explained. Several standard tabled

sampling plans are discussed. Their disadvantages are also presented.

36
In Chapter 5 problems are formulated based upon these disadvantages.

These problems are solved using ANNs. In the next chapter, ANNs are

studied in detail.

37