Lots of product entering inspection may have some proportion of nonconforming items. This proportion of nonconforming units is characterized by the level of quality. A quality level is any relative measure of quality obtained by comparing observed values ​​with specified requirements.

The quality level can be expressed as a percentage of nonconforming items (the ratio of the number of nonconforming items to the total number of items) or as the number of nonconformities per 100 items of product (the ratio of the number of nonconformities to the total number of items).

With selective control, it is impossible to establish the actual level of quality in the controlled batch of products, but only its assessment can be obtained. The accuracy of this estimate depends on how justified the control plan is. As such an assessment, when controlling by a quantitative attribute, the limit value of the controlled parameter in the sample is used, and when controlling by an alternative attribute, the quality level is used.

Under acceptable AQL quality level when considering the sequence of batches, the average level of quality is understood, which, for the purposes of product acceptance, is satisfactory.

An acceptable level of quality for a particular inspection plan corresponds to a high probability of acceptance, provided that the level of nonconformities in the inspected lot does not exceed the specified AQL value. However, a given AQL does not mean that no more than the specified percentage of nonconforming items is allowed in the lot. In any case, it is preferable to have no mismatched units than to have any percentage whatsoever, and the more it can be reduced from the AQL, the better. Reducing the percentage of nonconforming items increases the likelihood of acceptance for each lot.

The choice of the required AQL value is carried out by agreement between the supplier and the consumer and is negotiated in the contract.

In many cases, AQL is a trade-off between what the consumer wants and what the manufacturer can afford, as stringent requirements are more difficult to meet in manufacturing process and more control costs will be required to verify that they are met.

Choosing the correct AQL value is one of the most important tasks when using statistical acceptance testing. The question of lowering or raising AQL should be economically justified. Choosing an unreasonably low AQL value will cause the supplier to incur losses from rejecting a significant proportion of good products, and setting an unreasonably high AQL value will force the consumer to accept batches of products containing a large number of nonconforming products.

An acceptable quality level serves as the basis for determining control standards in the case of batch sequence control (reference indicator in the tables of STB GOST R 50779.71 and tables of Annex A of STB GOST R 50779.75).

The AQL value determines the severity of the sampling.

Assign different AQLs to groups of nonconformities or different types of nonconformities.

When establishing the value of an acceptable quality level for products that are controlled by several quality indicators, acceptable level quality is defined in two ways:

AQL is established for individual quality indicators, and then for products as a whole;

· AQL is established for products as a whole, and then for individual quality indicators.

AQL values ​​of no more than 10 are established for both the percentage of nonconforming product items and the number of nonconformities per 100 product items. An AQL value greater than 10 is set only for the number of nonconformities per 100 units of production.

It is recommended to use the preferred AQL values ​​(26 values ​​from 0.010 to 1000), however, if this is not possible, the STB GOST R 50779.71 system allows you to build control plans for other AQL values ​​(STB GOST R 50779.70).

In AQL-based sampling, inspected lots taken from the process with a quality equal to or better than the AQL will in most cases be accepted.

In a continuous series of sampling lots, AQL is the quality level corresponding to the average satisfactory process limit.

AQL is a chosen boundary between acceptable and unacceptable process average values. It does not describe a sampling plan, but is a requirement of what production should be and a convenient value for determining an acceptable process.

When assigning AQL, it must be taken into account that it is a quality indicator required in production. The manufacturer is recommended to produce batches of medium quality better than AQL. On the other hand, this quality must be realistically achievable and at the same time justified from the point of view of the consumer. With a properly designed and managed process, it is possible to produce products with a lower percentage of nonconforming units than AQL. By obtaining a better average process, the overall cost of producing and inspecting better quality products is reduced.

Taking into account the requirements of the customer, it is necessary to check that it is not too high, and also take into account the intended use of the controlled products and the consequences of failures. If, with a larger number of items, a failure can be seen as a signal to replace the nonconforming item, then a reasonably mild AQL is acceptable. If this failure results in damage to a costly and critical piece of equipment where replacement is not possible, a more stringent AQL is required.

The process average is the average quality level of a series of lots submitted for inspection (lots resubmitted for inspection are excluded). The process average refers to what is actually produced, regardless of the controls carried out.

The evaluation of the average process is not an integral part of the control scheme, but it is important in itself. Both the inspector and the manufacturer are interested not only in decisions about successive lots, but also in the long-term picture of the quality of production.

With two-stage and multi-stage control, a number of special rules must be observed. Only the results of the first sample are used to evaluate the average process.

In some cases, it is recommended to exclude abnormal results, but this rule should be applied with great care. This can be done with confidence when the abnormal results are due to a special cause that has already been eliminated. Further, in this case, data are given, including and not including anomalous results, to show the presence of these inconsistencies.

If there are many characteristics or multiple AQL classes, the averages of the individual processes should be evaluated.

Ultimate quality LQ – the level of quality at which, for sampling purposes, the probability of acceptance is low when considering a single lot. Ultimate quality is a reference indicator in STB GOST Ρ 50779.72. When inspecting a single lot, this corresponds to a level of quality, expressed as a percentage of nonconforming items or a number of nonconformities per 100 items of product, at which a low probability of acceptance is required for sampling purposes. Marginal quality actually corresponds to undesirable quality. To ensure lot acceptance, the proportion of nonconforming items must be well below the LQ (typically less than a quarter of the LQ).

Average output quality AOQ is the expected average quality level of the outgoing product after control at a given value of the input quality level. Unless otherwise specified, the average output quality is calculated over all accepted lots plus all rejected lots after sweeping and replacing nonconforming units with appropriate ones. Often used approximation:

(average output quality) = (process quality before inspection) x (probability of acceptance)

AOQL average output quality limit- the maximum value of the average output quality among all possible values ​​​​of the quality level of the output products for a given plan for sampling and elimination of inconsistencies in all rejected lots

Similar to the concept of AQL, the concepts of average output quality (AOQ) and its limit (AOQL) are justified only with a large number of consecutive lots presented in a certain sampling system. The lot will be accepted if the number of nonconforming items in the sample is less than or equal to the acceptance number. If the number of nonconforming units is greater than or equal to the rejection, the lot will not be accepted. With an average process level close to AQL, most lots will be accepted. If the quality of the process does not change and rejected lots are rejected rather than corrected, sampling does not affect the quality.

In some cases, when the movement of products occurs between departments, and not enterprises, the rejected batch is checked by continuous control with the removal of non-conforming units from it (replacement with appropriate units is possible). This is called culling control.

During the inspection with sorting out, the batch is either accepted without further control, or in case of rejection, a complete inspection of each unit is carried out with the withdrawal or replacement of all non-conforming units with the appropriate ones. In the first case, the output quality practically corresponds to the input, in the second - all products correspond specifications. Even if the input quality p does not change, the output quality can change from batch to batch, taking on the values ​​p or 0 depending on whether the batch was accepted or submitted for inspection with culling. Nevertheless, it is possible to consider the average output quality over a long period, when the input quality does not change and is equal to p. This average quality will be no worse than p, and with complete control of a large proportion of batches, it can be much better.

The term "average output quality" can be thought of as the average percentage of nonconforming items in a large number of lots from a process that continuously produces quality p. All lots are tested and evaluated with the same sampling plan, which has the probability of accepting the lot P a. Rejected lots are (in theory) cleared of nonconforming units. As a result of control with sorting out, on average, 100 R a % of batches do not contain non-conforming units, and 100 (1 - R a) % of batches that have passed only selective control and accepted from the first presentation contain 100 p % of non-conforming units (minus a number of items withdrawn from samples under control). The average output quality, expressed as a percentage of nonconforming units, is approximately 100(P a x p)%. The approximation is acceptable provided that lot size N is ten or more times larger than sample size n.

The output quality can be good both because of the good input quality and because of the complete control of a number of batches. In addition, there is an intermediate input quality p for which the average output quality has a maximum value. This value is the AOQL average output quality limit. This is not the limit of the output quality of any single batch, nor is it the limit of the actual output quality averaged over a small number of consecutive batches. In the case of a large number of them, the actual average output quality in this sequence will differ slightly from this AOQL. With variations in the input quality p, the actual quality can be much better than AOQL. Therefore, it is more efficient to estimate the real average quality directly than to rely on AOQL as an upper bound.

Process mean is the average quality level of a series of batches supplied for inspection (lots repeatedly submitted for inspection are excluded). This concept is different from AQL, AOQL or LQ, which can be calculated or selected, and is not a characteristic of a particular sampling plan. The process average refers to what is actually produced, regardless of the controls carried out.

The evaluation of the average process is not an integral part of the control scheme, but it is important in itself. Both the inspector and the manufacturer are not only interested in successive batch decisions, but also in the long-term picture of production quality. It is recommended to keep records of process average data, which is an effective measure to improve product quality and necessary information when choosing a sampling plan.

Operational characteristics of the sampling plan

When applying control plans, batches of products are accepted or rejected with some probability less than one. The probability of accepting an inspected lot depends on the proportion of nonconforming items in that lot. If there are no nonconforming items in the lot, then there cannot be any in the sample, and such a lot will in all cases be accepted with probability equal to 1. As the proportion of nonconforming items in the lot increases, the probability of accepting the lot decreases. If the entire batch consists of nonconforming units of production, then such a batch will in all cases be rejected with a probability equal to 1.

The function that sets the probability of acceptance of a controlled batch of products, depending on the input quality level, is called the operational characteristic.

The operational characteristic curve shows the mathematical expectation of the percentage of accepted batches of products. These values ​​are average values ​​that correspond to actual values only with a large number of considered batches of products.

The probability of accepting a lot of products depends on the sample size, control standard and quality level in the lot.

With an increase in the sample size (with the other two initial data unchanged), the probability of accepting a batch of products decreases.

For the supplier, increasing the sample size is unprofitable, as it increases the risk of rejecting a good batch of products; on the contrary, it is beneficial for the consumer, since his risk of accepting defective products is reduced. With the weakening of the requirements for the rigidity of the control standard (also with unchanged initial data), the probability of accepting a batch of products increases, which is beneficial for the supplier and disadvantageous for the consumer.

To meet the requirements of the supplier and the consumer at the same time, a compromise is necessary. As such a compromise, there should be an acceptable level of quality agreed between the supplier and the consumer.

Control Plans

The control plan is understood as a set of requirements and rules that should be observed when controlling a batch of products. The set of requirements and rules is understood as the volume of the controlled lot, the level and type of control, the type of sampling plan, the sample size, control standards, etc.

Sampling plan - a set of data on sample sizes and control standards - acceptance and rejection numbers (by an alternative attribute) or limit values ​​of a controlled parameter in a sample (by a quantitative attribute).

Statistical acceptance control scheme - a complete set of sampling plans in combination with a set of rules for applying these plans.

Depending on the number of samples selected for control, single-stage, two-stage, multi-stage and sequential control plans are distinguished.

A single-stage control plan is characterized by the fact that the decision on the acceptance of a batch of products is made based on the results of control of only one sample.

From a batch of products with a volume N the sample size is extracted n z z less than or equal to acceptance number A C z greater than or equal to the rejection number R E, the batch is rejected.

This plan is applied in the following cases: the cost of the inspection is low, the duration of the inspection is too long and the lot cannot be held until the end of the inspection. The plan is characterized by the largest sample size. A diagram of a single-stage plan is shown in the figure.

Figure - Scheme of a single-stage statistical acceptance control.

A two-stage control plan is characterized by the fact that the decision to accept a batch of products is made based on the results of control of no more than two samples, and the need to select a second sample depends on the results of the control of the first sample.

From a batch of products with a volume N the sample size is extracted n 1, which determines the number of nonconforming units z1. If the number of nonconforming units z1 less than or equal to the acceptance number of the first stage of control A 1, the party is accepted; if the number of nonconforming units z1 greater than or equal to the rejection number of the first stage R1, the batch is rejected. If the number of nonconforming units z1 is within the limits between the acceptance number of the first stage of control A 1 and marriage number of the first stage R1, move on to the next level of control. From a batch of products, a sample is taken with a volume n 2, which determines the number of nonconforming units z2. If the sum of the number of nonconforming units at the first and second stages of control ( z1 + z2) less than or equal to the acceptance number of the second stage of control A2, the party is accepted; if the sum of the number of nonconforming units at the first and second stages of control ( z1 + z2) greater than or equal to the rejection number of the second stage R2, the batch is rejected.

This plan should be used if a single-stage design cannot be accepted due to the large sample size and a multi-stage design due to long duration. The scheme of the plan is shown in the figure.

Figure - Diagram of a two-stage statistical acceptance control.

A multi-stage control plan is characterized by the fact that the decision to accept a batch of products is made based on the results of the control of several samples, the maximum number of which is set in advance, and the need to select a subsequent sample depends on the results of the control of previous samples.

The scheme of multi-stage statistical acceptance control resembles the scheme of two-stage statistical acceptance control. At each stage of control, a sample of size n is taken from a batch of products with a volume of N, in which the number of nonconforming units z is determined. If the sum of the number of nonconforming items at all inspection levels is less than or equal to the acceptance number for that inspection level, the lot is accepted; if the sum of the number of nonconforming items at all stages of control is greater than or equal to the rejection number of this stage, the lot is rejected. If the sum of the number of nonconforming items at all stages of control is between the acceptance number of this stage of control and the rejection number of this stage , move on to the next level of control. Control stages (according to STB GOST Ρ 50779.71) cannot be more than seven.

This plan should be used when the time required to select and control items is short and the cost of testing is high.

Medical gloves are one of the main means of ensuring infectious safety in a medical institution. Wide variety of medical gloves Russian market currently, as well as aggressive marketing policy a number of manufacturers do not always allow healthcare facility staff to adequately understand the types of gloves and choose the really right products to ensure safe work with patients.

All properties (characteristics) of medical gloves can be divided into 3 main groups: the main ones are present in any glove, additional ones - specialized gloves may have, dubious ones - various marketing tricks of manufacturers and suppliers, designed, among other things, to reduce possible competition.

The main properties of a medical glove include:

1) material of manufacture.

The main manufacturing material is latex - an emulsion of rubber particles in an aqueous solution. There are natural or synthetic rubbers, depending on the type of rubber, natural latex is distinguished, as well as nitrile, polyisoprene, polychloroprene latex and vinyl (polyvinyl chloride, "plastic" latex).

Natural latex consists of more than 60% polyisoprene particles, which allows us to consider polyisoprene gloves as the closest in properties to conventional latex. The advantages of natural latex are widely known: gloves made of this material stretch well, fit the hand, are soft and elastic. At the same time, they contain a significant amount of proteins that cause allergic reactions and are not resistant to alcohols, oils, and esters. The world standard for the presence of proteins in a natural latex glove is less than 50 µg/g, determined by the Lowry method (colorimetric method). A number of manufacturers offer gloves with a protein level of less than 20 µg/g, and in early 2011, gloves made of natural latex, completely free of proteins (MPXX technology - maximum protection), appeared.

Nitrile latex perfectly resists the action of alcohols, aldehydes, phenols and acids, which makes it possible to use nitrile gloves in laboratories, when working with aggressive environments, and when cleaning rooms in healthcare facilities. In addition, synthetic gloves are absolutely non-allergenic, as they do not contain proteins, but, nevertheless, very often cause contact dermatitis during prolonged work. It is impossible to call nitrile gloves hypoallergenic. The disadvantages of nitrile gloves include low elasticity and extensibility, which prevents their widespread use in surgery.

Polychloroprene (neoprene) and polyisoprene gloves are quite expensive, so they are mainly used as surgical gloves. There is no reasonable need to use examination gloves made of neoprene and polychloroprene, since they are similar to nitrile in terms of resistance to chemicals. At the same time, the use of such gloves during surgical operations allows you to provide highest level protection for all members of the surgical team.

Vinyl gloves made from PVC - polyvinyl chloride, a short-lived and rather harmful substance. These gloves are cheap, but have one important drawback: slight permeability to any proteins (including blood proteins) and microorganisms, which does not allow them to be used even for a short-term examination of patients.

Currently there are gloves having 2 different layers. for example, DermaGEL gloves from the Polish company Mercator Medical have an inner layer of nitrile or polyurethane, which gives the glove extra strength, insulates the skin from the action of natural latex proteins and greatly facilitates donning. The start of production of such hybrid gloves, combining the properties of different materials, is an important step towards creating a new level of infectious protection.

2) the presence or absence of powder

The negative properties of the powder used in the manufacture of gloves are widely known. Initially, the use of powder was due to the technological process of production, to prevent sticking of the walls after removal from the mold. In this case, multiple cleaning and rinsing steps were used to make powder-free gloves, resulting in a significant price difference between powdered and powder-free gloves. At present, such cleaning is used only by manufacturers with outdated equipment; at most enterprises, powder is excluded from the technological process, which has significantly reduced the difference in price. In order to keep the gloves from sticking together after demolding, most factories use surface modification - this process will be discussed later.

Powder is the strongest absorbent, attracting and holding proteins and microorganisms. Thus, it is a source of infectious danger, and enhances the allergic effect of natural latex on the skin. Numerous studies have shown a significant role of powder in the occurrence of postoperative complications, adhesions and scars, the occurrence of allergic reactions in both patients and healthcare facility staff, and the spread of nosocomial infections.

3) textured surface

The texture of the outer surface improves the grip of the medical instrument. In production, a textured surface is formed using molding solutions - chemical substances acting on the outer surface of a not yet finished, not completely dried glove. In general, it is necessary to separate the concepts of a textured (textured) and micro-rough (micro rough) surface. In the second case, the surface of the glove changes very little and, in general, in terms of its contact properties, the microrough surface is closer to smooth than to textured. In addition, it is important to note that Russian GOSTs require the manufacturer to provide a minimum thickness of 0.08 mm for smooth sections of examination gloves, and 0.11 mm for textured ones. For surgical gloves, these values ​​are 0.10 and 0.13 mm, respectively. Neither the 0.07mm smooth glove nor the 0.10mm textured glove can be used for medical applications.

4) presence / absence of a roller

The roller is a structural element of the glove, with the help of which the glove is fixed on the wrist. The main requirement for gloves with a roll is that the roll must be rolled up inside the glove, if it is rolled outward, the space between the roll and the outer surface of the glove is a source of significant infectious hazard. During a surgical operation, this space cannot be treated with a skin antiseptic and is a kind of storage for carrying bacteria. In the absence of a roller, a mandatory component of the glove should be a reinforced or reinforced cuff that fits snugly around the wrist. Gloves without a roll without a reinforced cuff are made from ordinary gloves by simply cutting off the roll, they do not fit and do not stick to the wrist and their use is not justified.

5) AQL

AQL (Eng. Acceptable Quality Level, Guaranteed Quality Level) - maximum allowable amount defects in a batch of samples of a certain size. AQL is one of key indicators quality for mass production. In AQL testing - an acceptable quality level - a certain number of samples of manufactured products are selected according to a carefully defined procedure for random testing. These randomly selected samples are then tested in accordance with approved state standards and specifications. Based on the results obtained, a conclusion can be drawn about the quality of the entire batch of products. The higher the product quality requirements, the stricter the testing requirements.

Basically, they test medical gloves for water resistance. This is a procedure for determining the ability to retain water. 1000 ml is poured into the glove. water, while the glove should not leak for a certain period of time.

Thus, AQL is a statistical procedure for determining the quality of a glove.

The lowest AQL level allowed for a medical glove according to Russian GOST is 2.5, according to the European standard EN 455 - 1.5. There are medical gloves with an AQL of 1.0 or 0.65. Visually, this indicator can be assessed as the probability of having defective copies in a batch of 1000 boxes of gloves, 50 pairs each. With an AQL of 2.5, the probability that there will be no defective gloves in the box is only 3%, i.e. virtually every box in a batch will contain one, two, or more defective items. With AQL 1.5, the probability of having defective products in the box can be estimated at 22%, and with AQL 1.0 - at 6-8%.

Thus, a slight increase in the AQL level leads to a significant guaranteed reduction in the number of defective gloves in the lot.

6) Length and thickness

As mentioned above, Russian GOSTs (52238 - 2004 and 52239 - 2004) require the manufacturer to provide a minimum thickness of 0.08 mm for smooth sections of examination gloves, and 0.11 mm for textured ones. For surgical gloves, these values ​​are 0.10 and 0.13 mm, respectively.

The length of an examination glove should not be less than 220mm, for a surgical glove - 255mm. Moreover, European standards EN 455 are even more stringent and do not allow lengths less than 240 and 280 mm for examination and surgical gloves, respectively.

At the same time, I would like to note that more and more diagnostic gloves are produced with a length of at least 290mm. Long gloves are absolutely essential when used as protective against harmful chemicals - in laboratories, when cleaning, when working with cytostatics or to protect against viral infections.

Additional properties of a medical glove include the following:

1) Anatomical shape.

GOST 52238 - 2004 indicates the presence of an anatomical shape as a prerequisite for classifying this glove as a surgical glove. In general, the shape of a glove with the thumb extended forward is called anatomical, which significantly reduces hand fatigue during work and long-term surgical operations. Anatomically shaped gloves are more expensive to manufacture than a regular (flat) shape and can only be worn on the corresponding - right or left - hand. For a more accurate selection of sizes for such gloves, digital (from 5.5 to 9) designations are used instead of the usual alphabetic ones for diagnostic gloves (XS, S, M, L, XL). Size XS corresponds to sizes 5.5 and 6, S - 6, 6.5 and 7, M - 7, 7.5 and 8, L - 8 and 8.5.

There is the concept of “improved anatomical shape”, which is a shape with fingers bent towards the palmar side, which reduces the load not only on the thumb, but also on all the others.

2) Sterility

The sterility of the glove is ensured by sterilization, i.e. complete release from all types of microorganisms, including bacteria and their spores, fungi, virions, as well as from prion proteins. Sterilization can be carried out by thermal, chemical, radiation, filtration methods; in industrial volumes, surgical gloves are sterilized by chemical (ethylene oxide gas sterilization) or radiation (gamma radiation) methods. Thermal sterilization of non-sterile gloves is carried out in the healthcare facility. Both radiation and gas sterilization are absolutely safe for the consumer and equally effective in removing microorganisms. Often the same manufacturer of gloves can be sterilized with both gamma radiation and ethylene oxide. As a rule, to confirm sterilization, an indicator is applied to the box, which changes its color with sufficient exposure intensity and confirms sterility.

3) Surface modification

Modification of the inner or outer surface of a medical glove is a common process designed to make the surface smoother. The modification of the inner surface is used to facilitate donning, including on wet hands, the modification of the outer surface is intended for more convenient operation with small tools so that the contact surfaces of the fingers do not stick together.

There are 2 main methods of surface modification: chlorination (chlorination) and polymer treatment.

Chlorination- this is the treatment of a glove with perchloric acid, it can be single or double (on both sides). As a result of chlorination, latex particles are destroyed and form a smooth film on the surface. There is a process of partial dehydration, ie. removing water, making the glove feel drier to the touch. Chlorination can be carried out both on the production line (online chlorination) and by soaking gloves in a perchloric acid solution for a long time. Online chlorination is one of the main ways to prevent gloves from sticking together after removal from the mold, this step has replaced technological process dusting. The concentration of chlorine in this case is extremely low and the process practically does not affect the properties of the glove. Double chlorination, carried out for a long time, significantly changes the properties of the glove. The destruction of the latex particles leads to a decrease in the elasticity and extensibility of the glove. Uncontrolled long-term chlorination leads to the appearance of an intense yellow color, makes the latex permeable to proteins and microorganisms, traces of chlorine on the surface of the glove can adversely affect both the doctor and the patient.

At the same time, an increase in the intensity of exposure has practically no effect on the properties of the surface - the surface becomes smooth even with not very strong chlorination.

Another method of surface modification is polymer treatment. The function of the polymer coating is to improve donning and prevent sticking of the glove. Almost any polymer can be used for surface treatment, polyurethane or silicone are most commonly used. When processing with polymers, it is not the formation of an additional functional layer, but a kind of smoothing of surface irregularities. The polymer-treated surface does not protect the skin from the action of latex proteins and does not interfere with the contact of natural latex with the skin, therefore it cannot serve as a protection against allergic reactions or contact dermatitis.

The questionable properties of medical gloves include the presence of a moisturizer (glycerin, tocopherol), extracts of aloe vera, chamomile, various compounds to improve skin trophism without confirming clinical efficacy and indicating the quantitative composition.

/ / / /

What does ACL mean?

When using the services of an independent inspection company, it is very important to understand the standard used to take random samples for cosmetic inspection.

The most common standard for product inspection is ISO standard 2859-1 (ANSI/ASQC Z1.4-2003). It uses the AQL (Acceptable Quality Limit) concept.

What does ACL mean? The standard definition of AQL is "the maximum percentage of defects in a lot (or the maximum number of defects per hundred units of production) that can be considered as meeting average standards and acceptable for the purposes of lot acceptance for quality."

Sample size, based on the AQL tables, will be selected and then checked for defects.

Defects are divided into three categories: minor, major and critical. Although different clients categorize differently, typical definitions are as follows:

• Minor defect- this is a non-compliance with standards, but something that most likely will not affect the use of the product.
• Major Defect is one that is most likely to result in the inability to use the product for its intended purpose.
• Critical Defect is one that is considered dangerous or unsafe.

According to the number of defects found and according to the number of defects allowed (figures are given in the AQL tables), your inspection company can advise you, accept or refuse from the party.

How to use AQL tables?

The AQL tables will help you determine the inspection sample size we need based on the quantity ordered and your inspection severity level. You can select Levels I, II, or III, with Level III being the most stringent and Level I being the least stringent. The standard level that is used by default and 98% of people is level II. The choice is up to the client, but this is the recommended level.

In order to find the required inspection sample size, you must first refer to the first table and on the left find the total number of products produced. For example, if you produce 8,000 pieces, at level II you see the letter L, which in the second table represents a sample size of 200.

At the top of the second table are defect levels from 0 to 6.5 (we cut out the higher numbers because they don't apply to consumer goods buyers).

You can choose which level to apply to your types of defects: critical, major, and minor. As a rule, most importers choose standard defect levels: 0/2.5/4.0 , but someone also chooses 0/1.5/2.5, this is at the request of the client.

Using standard defect levels of 0/2.5/4.0 and a sample size of 200, we see that if you have more than 0 critical defects, 10 major defects, and 14 minor defects, you should discard that lot.

Of course, the decision what to do after receiving the results of the inspection, remains only with you. Most importers prefer to discuss what was found during the inspection with the supplier/manufacturer in order to improve anything that can be improved. In the event that the inspection results are very close to the AQL limits, it is important to double-check whether the level of defects found is acceptable to you or not.

Checking a batch of goods is an important and complex process. In any batch there is a marriage, and with large volumes, you cannot manually check each unit. How to quickly inspect products and make sure that the quality level is acceptable? How to more or less accurately find your boundary between rejection and acceptance of a batch if a defective product is found?

How many units need to be checked?

There are precise answers to these difficult questions, and you can find them in the Acceptance Quality Limits or AQL tables. As the name implies, they define the limiting level of quality at which a batch of products will not upset the consumer and you. So, after checking a relatively small number of samples, the batch can be accepted.

The authors of the AQL tables were guided by the ISO 2859 standard, but there are no general limits for each market and product type. In each segment they are different, and for small consumer goods this level is higher than for cars.

The level is determined as a percentage, that is, you have researched the market, your consumer and decided that the maximum allowable percentage of defective products is 2.5% of the entire batch. So your AQL will be 2.5%.

AQL tables already have metrics for your product group, including lot size, inspection level, and AQL itself.

And if you need to check a batch of high-end chairs with a volume of 3,500 pieces with a medium (usually choose this level) second-level check, then your sample size code is L.

We look at the second table and see that L = 200 units of goods from the entire batch. This is our sample size. We keep in mind that for high value products the AQL is lower and we choose levels of 0%, 1% and 2.5% for each defect category.

But first you need to divide the potential defects of your product into three categories.

The first is critical defects. The product cannot be used, does not meet safety requirements, or does not perform its functions.

The second is significant defects. The product can be used, but it is unlikely to be bought.

And the third-- Defects are minor, when deviations from the specification do not affect the popularity of the product.

We also chose an AQL of 4 for minor defects and 2.5 for major defects. It turns out that the batch can be accepted if during the inspection no more than ten significant and fourteen minor defects are found. If we find seven minor and eleven major defects, then such a batch does not meet the standard. If seven major defects and 10 minor defects are found, then the lot is accepted.

Do not forget that a separate check is needed for each group of goods in one segment, and defective products can be bought and sold at a reduced price. And in case of difficulties, Asianinspector experts will always help you. Write and call us!

Reference indicator in the tables of acceptance statistical control. In AQL-based sampling, inspected lots taken from the process with a quality equal to or better than the AQL will, in most cases, be accepted.

In a continuous series of sampling lots, AQL is the quality level corresponding to the average satisfactory process limit. AQL is a chosen boundary between acceptable and unacceptable process average values. It is a requirement of what the production should be and a convenient value for determining an acceptable process.

A given AQL does not mean that a percentage of non-conforming items is allowed in the lot, not more than the specified one. In any case, it is preferable to have no mismatched units than to have any percentage whatsoever, and the more it can be reduced from the AQL, the better. Reducing the percentage of nonconforming units increases the likelihood of acceptance for each lot.

With a properly designed and managed process, it is possible to produce products with a lower percentage of nonconforming items than AQL. By obtaining an average process, the total cost of producing and inspecting better quality products is reduced.

AQL must be specified in the contract or appointed by an authorized party (supervising organization). Various meanings AQLs can be assigned to different classes of nonconformities based on their significance.

According to the degree of significance, the following classes of non-compliance are distinguished:

• A - discrepancies that are of the greatest importance for products or services. In acceptance sampling, this type of nonconformity has small AQL values;
• B - inconsistencies of lesser significance; for them, the acceptable quality level AQL values ​​​​are set higher than for nonconformities of the first type, and less than for the third type, if any, for example C, etc.

Small values ​​set for class A nonconformities form a tighter control than in the case of class B.

Types of selective statistical control

Sampling applied to consumer goods can be classified in two ways:

• 1) by type of control plan;
• 2) rigidity of control.

By type of plan A distinction is made between single-stage control and two-stage control.

Single stage control- Sampling, in which the decision to accept or reject a lot in accordance with certain rules is made on the basis of the results of the control obtained from one sample(Table 4.5).

Table 45

Statistical sampling by attribute of garments

Lot volume, pcs.	Sample size, pcs.	Normal control		Sample size, pcs.	Enhanced control
		acceptance number	rejection number		acceptance number	rejection number
					Solid control

Two stage control - selective control, in which, after the control of the first sample, a decision is made to accept, reject a batch or select a second sample for a final decision on acceptance or rejection in accordance with the rules defined by the product standard (Table 4.6).

Single-stage plans are easier to control. However, two-stage ones provide, with the same sample size, greater accuracy of the decisions made, but are more complex in organization.

Table 4.6

Two stage control

Lot volume, pcs.		Sample size, pcs.	Total sample size, pcs.	acceptance number	Rejection number
Up to 50 inclusive

Consider an example with selective control of household utensils made of special household glass.

The batch is accepted if the number of products that do not meet the requirements of the standard in the first sample is less than or equal to the acceptance number.

The batch is rejected if the number of products that do not meet the requirements of the standard is equal to or greater than the rejection number for one-stage control or more than the rejection number for two-stage control.

If the number of products that do not meet the requirements of the standard in the first sample is equal to the rejection number, then a second sample is selected.

The batch is accepted if the total number of products that do not meet the requirements of the standard, the first and second samples is less than or equal to the acceptance number and rejected if the total number of products that do not meet the requirements of the standard is equal to the rejection number or more than the rejection number.

Based hardness, There are: normal control, enhanced control, weakened control.

Normal control- selective control, which is carried out when there is no reason to believe that the actual level of production quality differs from the acceptable level.

Enhanced control- selective control, more stringent than normal, to which they pass from the last one, if the results of the control of a given number of consecutive lots show that the quality level is worse than the established one.

Weakened control- selective control, less stringent than normal, to which they pass from the last one, if the results of the control of a given number of consecutive lots show that the quality level is better than the established one.

Normal control is the main type of control. Normal control is designed to protect the manufacturer from a higher proportion of rejected lots that are better than AQL. If the control results indicate a worse process average than the AQL, then the consumer introduces enhanced control.

If the quality is systematically better than AQL, then the control is not abandoned at all, since a signal of a possible deterioration in product quality is needed. In this case, it is much more economical to organize a weakened control.

From Table. 4.5, devoted to the selective control of garments, it is clear what the rigidity of enhanced control is manifested in: for example, when sampling 50 pieces. within normal inspection, you can accept a lot with a number of defective items equal to 3 and reject it with a number of 4; within the framework of enhanced control, a smaller number of defective units is allowed: 2 and 3, respectively.

The procedures and rules for switching to different types of control in terms of severity are considered in (2.25) and in fig. 4.7.

Rice. 4.7.

• AQL - from English, acceptable quality level - acceptable quality level.