Sampling in Textile Industry: Aims, Techniques and Factors

Sampling in Textile Industry

In the textile industry, sampling is a crucial process that bridges the gap between product development and bulk production. A sample is, “a relatively small fraction which is drawn from the population to represent the popu­lation”. In most situations, the evaluation or estimation of attributes or characteristics of the entire system, process or product is carried out through a representative sample. The whole population of a material or a product generally is so large that every item cannot be tested completely. Testing can be destructive or non- destructive, and not all the testing can be done without damaging or separating a fraction of material. For example, tensile testing cannot be performed without rupturing the fabric or yarn. The whole population can never be tested for tensile strength. Therefore, testing a represen­tative sample is done instead. It can be more efficient while still providing the required information. A well-executed sampling process helps ensure product quality, cost-effectiveness, and customer satisfaction.

It is not necessary to perform the testing of complete material and it is practically impossible as well, due to time and cost factor. The destructive type of testing will also increase the wastage in the process, which will ultimately increase the cost of testing, resulting in overall profit being decreased. Due to this we use model / symbolic samples from the bulk material for testing. Major objective of sampling is to make an unbiased sample representing the whole.

Assume that a spinning industry having a capacity of fifty thousand spindles, is producing cotton yarn. So, the industry is producing more than fifty thousand ring bobbins per day, which is a huge quantity to be tested. To test the whole population, staffing, machine, time, electricity, and many other expenses will be needed. It will be time and cost- consuming. To evaluate the quality of the produced bobbins, it is better to take a fraction of the population that can represent the entire popu­lation.

The amount of effort, time and money that can be saved using sample testing can be seen in the following examples.

If a sample of 100 g cotton is taken from 10 tons of cotton bales, the fraction is:

………………………100
Fraction = —————————–
………………10,000 x 1000

……1
= ———-
….105

It can be seen that only a 1/105 fraction of the total mass of cotton bales was tested.

If there is a container having 15 tons of yarn in the form of cones of 1.5 kg each. 10 random samples are taken out for testing, then the fraction is,

…..10 x 1.5
= ——————
….15 x 1000

…….1
= ————–
….1000

It can be seen that only a 1/1000 fraction of the total quantity of yarn was tested.

This fraction is not only a saving of the material that is tested, but also is directly related to the saving of staffing, time, and money.

The biggest problem with textile materials is their variability. It may be fiber, yarn, or fabric, invariably it will have variation in characteristics. Fibers will have varying diameter, length, maturity, and so forth, the yarns will have varying linear density, diameter, strength, hairiness, and so forth, and the fabrics will have varying strength, handle, and so forth, in different portions. So, sampling must be done in a very precise manner.

Aims / Objectives of Sampling

The main objective is to get an unbiased test sample that represents the whole population; each and every part of the possible length group is represented in the sample. The result of a sample may be different from the result of the whole population, but the sampling must be unbiased. For example, in cotton fiber sampling from bale, the proportion of fiber length in a sample should be almost the same as the proportion in the bulk. So, if random or unbiased sampling is done, then it can only be said with certain confidence that the proportion of length of fibers in the sample is similar to that of fibers in bulk or the bale. Each fiber in the bale should have equal chance of being chosen.

As an extreme example, suppose it is indicated in a standard that the sample should be taken from a bale from its surface, the fiber supplier has put shorter fibers in the center and longer fiber at surface as shown in Figure 1. That is how one can sell inferior quality fibers at a price of superior ones. In this situation it is important that the fibers are taken from everywhere from the bale, namely, fibers from the center, surface, top and bottom are collected. Therefore, all the fibers will have an equal opportunity of being tested.

Factors of Sampling in Textile Industry

The sampling method is not a rigid procedure. For a particular type of testing, it may differ based upon various factors as listed below.

1. Sampling method is decided based upon the form of the material. For example, the material could be in the form of fiber, yarn, or fabric. Fibers could be in loose form, in the form of a bale, sliver or roving form. According to the form, the sampling method may be entirely different.
2. Sampling is also influenced by the nature of the material. Material can be cotton, wool, or any other fiber. Different fibers have different natures. For example, it is easy to take fibers from a cotton bale because cotton fiber is physically compressed into the form of a bale. However wool fibers contain a large quantity of grease which makes it very compact and highly sticky on its own and it is very difficult to take out fiber tufts from a wool bale. So, fiber sampling of wool will be completely different from that of cotton.
3. Sampling method is also decided based on the amount of material available. For a larger amount of material, sampling method will be different from that of smaller size sample.
4. Sampling also depends upon the nature of the test. Testing techniques may require different amounts and orientations of samples. For example, for the testing of fiber length, the comb sorter method has different sampling technique to that of the high volume instrument (HVI) method.
5. The type of the testing instrument also affects the sampling method. The sampling techniques are different for the high volume instrument (HVI) method and the advanced fiber information system (AFIS) method, although both of these instruments test cotton fibre parameters.
6. For same material, the sampling process can be different for different types of required infor­mation. For example, the core sampling technique is used to take out fibers from the core of the wool bale. The coring tube penetrates inside the wool bale through cutting the wool fibers. For this reason the fiber diameter, fiber moisture content or grease content, and the like, can be tested from the core sampling, but the fiber length cannot be measured using core sampling because the fibers are cut during their extraction.
7. Sampling is also decided by the degree of accuracy required. A lower number of samples will result in lower accuracy in the test result than that of a higher number of samples.

Sampling Errors

Errors can arise from a number of different sources. The properly calibrated instruments can still be a source of error if the test is not performed according to the specified protocol. Defects in machine parts, such as play due to wear and tear, slackness while mounting the specimen on the instrument, and vibrations (mechanical and electrical), can produce variation in test results.

The following types of errors might occur during sampling:

• Sampling only from the surface of a liquid at rest
• Sampling from edge of sheet
• Sampling from one segment of a lot
• Instrument calibration
• Improper reading
• Lack of accuracy of sample

Types / Techniques of Sampling

The sampling type is broadly divided into two categories, namely, statistical or probability sampling and non- statistical sampling.

Statistical Sampling or Probability Sampling

This sampling method is based upon statistical tools. The statistical method of sampling allows the random selection of samples and use probability theory to find out the results. It is further divided into four subcategories.

a) Random sampling:
Random sampling is based on a random number generation and then sampling is done by following the generated random number. In this method, each item in the population has an equal probability of being selected. The selection of the items can be done by allocation of a number to each item and selecting a random number with the help of a slip choosing method, or any random number generation tool, or by some software.

An advantage of random sampling is that a person can choose a sample both with and without replacement according to his wishes or according to the system he wants to adopt.

But the disadvantage is that it is possible that the higher proportion of samples are taken from a particular portion of the population and a major portion could be left untested.

b) Systematic sampling:
Unlike random sampling, systematic sampling follows a definite system during selection of a sample. The samples are selected at regular intervals after proper numbering of each item. The basic idea of this type of sampling is that the samples must be taken uniformly from the whole area of the population.

The biggest advantage of the systematic sampling is that the samples are taken from the whole population evenly. There are no chances of samples being taken from a small area. But it also has a disadvantage. Suppose a yarn has thin places at a regular interval. If the strength of the yarn is tested with systematic testing, it may be possible that all the samples have thin places, or no sample has a thin place. This may lead to false information about that yarn. So, it must be ensured before choosing systematic sampling that the actual population must not be in an order that may introduce any non- random factor in the sampling.

c) Stratified sampling:
Stratified sampling is used when the whole population has subgroups and at least one representative sample from each subgroup must be there in the final sample. It can be done in two ways. Either the proportion of a subgroup in the sample will be same as the proportion of that subgroup in the whole population or the number of samples from each subgroup will be the same in the sample irrespective of the proportion of subgroups in the population. For example, in a spinning mill, ring frames from four different manufacturers are available. 10% of the ring frames are Make- 1, 60% are of Make- 2, 25% are of Make- 3 and 5% are Make- 4. If 40 ring bobbins are to be selected, for any kind of testing, then 4 bobbins from Make- 1, 24 bobbins from Make- 2, 10 bobbins from Make- 3 and 2 bobbins from Make- 4 would be selected randomly. The larger sub­group will have a larger proportion in the samples. In the second method, 10 bobbins will be selected from each subgroup for testing. The population and sample combinations of the two methods are shown in Figure 2. The idea behind this strategy is that all the makes of ring frames should have equal and logical representation.

d) Cluster sampling:
Cluster sampling also has subgroups as stratified sampling, known as clusters, but where not all of the clusters are tested. It is also known as block sampling. This is suitable for a large population. Unlike the stratified sampling, cluster sampling is as heterogeneous as possible to match the population. Random samples are taken from one or more clusters.

Non- statistical Sampling

Non- statistical sampling is done based on the judgement of the person who is performing the tests. The person decides sample size, test groups and evaluation methods. Here, the information about the entire population cannot be extrapolated from the sample but some information that is quick and important can be taken. For example, a buyer picks up one bobbin and tests its strength. By doing this he can have an idea about the strength of the population but not with certain confidence. In tex­tile or garment industries, this type of quality audit is very common. An auditor may ask the testing person to pick one bobbin and test its strength to get a quick idea about the quality of the product. Non-statistical sampling can be classified into two types.

a) Haphazard Sampling:
In haphazard sampling, the samples are selected based on conveni­ence but preferably should still be chosen as randomly as possible. The example of hap­hazard sampling is the same as given earlier. A quality auditor asks the testing person to pick one bobbin and test its strength. The testing person may pick a bobbin from the top and nearest to him because of its convenience and test it. The advantage of this sampling is that it is usually quicker and uses smaller sample sizes than other sampling techniques. The main disadvantage of haphazard sampling is that since it is not statistically based, generalizations about the total population should be made with extreme caution.

b) Judgemental Sampling:
In this type of sampling, the person doing the sampling uses his/ her knowledge or experience to select the items to be sampled. As the name defines itself, it is judgemental. An experienced person may know which type of item will have a greater chance of having non- conformance and avoid that type of item. For example, a particular ring frame in the spinning section continuously produces faulty materials. It may be due to some humidity problem in that portion or vibration in the spindle. Even after various main­tenance cycles this could not be rectified. Knowing this, the testing person will try to avoid taking samples from that ring frame so that because of that particular ring frame the overall test results of the entire population would not be affected.

It is very much dependent upon the mentality of the person who is performing the sam­pling. So if a person can see a damaged bobbin or one that looks dirty, he might try to avoid that bobbin while sampling. He will prefer fresh looking items. Such types of sampling creates bias.

The sampling methods can be classified into more relative terms for textiles as follows.

Numeric Sampling

In numeric sampling, the fiber samples are extracted from the population in such a way that the proportion of the length of all the fibers in the sample is exactly equal to the proportion in the popu­lation or the fiber sample is not biased towards length. The sampling of the fiber is not affected by the fiber length. Separating the numeric sample from the population does not directly affect the rest of the population. For example, when the fiber samples are extracted from a sliver by gripping the tips of the fibers, the selection of fibers does not depend on the fiber length. The fiber distribution is assumed to be random along the length of a sliver. Therefore, it can be easily assumed that the characteristics of fibers in the sample are exactly same as that in the population.

Biased Sampling

In sampling, when the selection is influenced by factors other than chance then that is called biased sampling. The causes of bias can be as follows.

1. Bias due to physical characteristics: Bias depends upon the physical characteristics of the material. For example, if fibers are selected from a tuft then the longer fibers will have greater chances of being selected in the sample because the longer fibers are exposed in a larger area than the shorter fibers.
2. Bias due to position relative to the person: It is always easier to pick a sample that is nearest to the person. Like the example given earlier, to test a yarn in a bobbin, the lab assistant will have a natural tendency to pick the bobbins from the top of the heap or those nearest to him.
3. Subconscious bias: This is a psychology dependent bias. A person always likes to pick the best- looking sample subconsciously and avoids the one that looks damaged. It is similar to that of judgemental sampling.
4. Conscious bias: Conscious bias is governed by natural human tendencies and depends on the position of the producer or the customer. For example, as a producer a person will always be biased towards the positive side of his product whereas a buyer will always be biased towards the negative side and will try to suggest that the material is defective.

Core Sampling

Core sampling is a widely used technique in the textile industry, particularly for raw fibers like wool and cotton. This method is used to extract a small, representative sample from a larger bale or package of fiber to test and analyze its quality without damaging the entire lot. The core sampling technique is used for assessing the proportion of foreign matter, the waste percentage, and the moisture content in the compressed unopened bales of cotton or wool. A tube with a sharpened tip is forced into the bale and a core of wool or cotton is withdrawn. The technique was first used as core boring in which the tube was rotated by a transportable electric drill. It was then developed further to facilitate the cores to be cut by pressing the tube into the bale by hand. This enables samples to be taken in areas distant from sources of power (Figure 3).

Zoning Technique

Zoning is a widely used and important method for sampling raw fibers such as cotton, wool, or other loose natural fibers in the textile industry. The properties of these natural fibers may vary significantly from place to place. A small tuft of fibers is taken at random from each of at least 40 widely spaced places (zones) throughout the bulk of the consignment (Figure 4). Zoning technique ensures that the final sample is representative of the entire bulk, which is critical for accurate quality assessment.

Conclusion

Sampling is a fundamental process in the textile industry, ensuring quality control, cost-effectiveness, and reliable product evaluation across all stages of production—from raw fiber to finished garments. Effective sampling ensures quality, saves resources, and meets standards, which is vital for customer satisfaction and business reputation.

References

[1] Das, A. (2024). Testing of textile and fibrous materials. CRC Press.

[2] Ahmad, S., Rasheed, A., Afzal, A., & Ahmad, F. (2017). Advanced textile testing techniques.

[3] Amutha, K. (2016). A practical guide to textile testing. CRC Press.

[4] Saville, B. P. (1999). Physical testing of textiles. CRC Press.

Share This Article!