In quality control, inspection systems serve to examine and measure the characteristics of a product, as well as the components and materials from which it is made, or of a specific service or process, all using measuring instruments, comparison standards") or test and trial equipment, all in order to verify whether or not it meets the specified requirements. Therefore, they serve to confirm that the quality system works as expected. Usually, it is done by sampling "Sampling (statistics)"), and is only used 100 percent control for important safety features, functionality, or standards).
The definition of sampling in ISO 8402/94 says that activities such as measuring, examining, testing or patterning one or more characteristics of an entity and comparing the results with specified requirements serve to determine whether conformity has been achieved for each characteristic.
Inspection types
Contenido
En una primera clasificación, los tipos de inspección podrían diferenciarse entre la «inspección 100 %» y la «inspección por muestreo».
Inspection Process
The inspection process is the process that consists of verifying all the units of a lot. An inspection allows accepting only pieces of the specified quality, but when the inspection is carried out manually, two types of problems arise, one would be the expense involved and the other would be the precision of the inspection, considering an error or a way to perform the proper inspection.
Sampling inspection
In contrast, sampling inspection systems, also known as acceptance sampling or lot sampling, is a procedure in which one or more samples of the lot are checked to determine their quality. Sampling is used to reduce the need to inspect each item or product, thereby reducing inspection time and expense. Sample inspection has a number of advantages over 100% inspection. Inspector fatigue caused by repetitive operations can be a serious obstacle to a 100% good inspection, it is more economical and requires less time to carry out.
Automated quality inspection
Introduction
In quality control, inspection systems serve to examine and measure the characteristics of a product, as well as the components and materials from which it is made, or of a specific service or process, all using measuring instruments, comparison standards") or test and trial equipment, all in order to verify whether or not it meets the specified requirements. Therefore, they serve to confirm that the quality system works as expected. Usually, it is done by sampling "Sampling (statistics)"), and is only used 100 percent control for important safety features, functionality, or standards).
The definition of sampling in ISO 8402/94 says that activities such as measuring, examining, testing or patterning one or more characteristics of an entity and comparing the results with specified requirements serve to determine whether conformity has been achieved for each characteristic.
Inspection types
Contenido
En una primera clasificación, los tipos de inspección podrían diferenciarse entre la «inspección 100 %» y la «inspección por muestreo».
Inspection Process
The inspection process is the process that consists of verifying all the units of a lot. An inspection allows accepting only pieces of the specified quality, but when the inspection is carried out manually, two types of problems arise, one would be the expense involved and the other would be the precision of the inspection, considering an error or a way to perform the proper inspection.
Sampling inspection
In contrast, sampling inspection systems, also known as acceptance sampling or lot sampling, is a procedure in which one or more samples of the lot are checked to determine their quality. Sampling is used to reduce the need to inspect each item or product, thereby reducing inspection time and expense. Sample inspection has a number of advantages over 100% inspection. Inspector fatigue caused by repetitive operations can be a serious obstacle to a 100% good inspection, it is more economical and requires less time to carry out.
That is why research was carried out in the field of probability theories and statistics, reaching the conclusion that to make decisions about the quality of in-process and finished production, there is no need to carry out a 100% inspection on all items, but it is enough to inspect only a part of the lot, that is, a sample, through a sampling inspection.
Some of the factors to consider in sampling inspection will be the level of trust in suppliers, the cost incurred by accepting defective products, and the risk of sampling, which will always exist due to the statistical nature of the process. In general, there are two types of errors that are likely to occur, the first is called type I error, and occurs when we reject a lot that meets quality specifications, and the second is called type II error, and occurs when we accept a lot that does not meet quality specifications.
Within quality sampling inspection, two types of inspection are mainly distinguished to control production processes. These processes are called “Inspection by Attributes” and “Inspection by Variables”.
Inspection by attributes can be considered that type of inspection of random samples of units in which each article or product is classified according to certain attributes as acceptable or defective, that is, it consists of finding out whether the material under consideration meets or does not meet what is specified, regardless of the measurement of the characteristic.
For inspection by attributes, the size of the samples and the interval between them must be such that approximately 5% of the production is inspected. In very massive processes that do not present frequent difficulties or the percentage of defective production is not serious, this percentage can be reduced to less than 5% where it is recommended that there must be at least 25 defectives in each sample to establish adequate behavior of the process.
Inspection by variables is a type of inspection that consists of measuring and recording a unit of measurement in which a specific quality characteristic is measured with a continuous scale to review its variation.
The statistical methods applicable to inspection by variables are based on the assumption of a normal distribution and not on a distribution of proportions as is the case with inspection by attributes. For the applicable methods, and with the measurements obtained, a statistic will be calculated, which will generally be a function of the sample mean and standard deviation, and depending on the value of this statistic when compared with a permissible value, the entire batch will be accepted or rejected.
The advantages that this method has with respect to the inspection method by attributes would be that the operating characteristic curve can be obtained with a smaller sample size than that required by a sampling plan by attributes. Furthermore, when destructive tests are used, sampling by variables is particularly useful to reduce inspection costs. On the other hand, measurement data typically provides more information about the lot than attribute data.
On the contrary, the distribution of the quality characteristic must be known, a plan must be used for each quality characteristic that must be inspected, and it is possible that the use of a variable sampling plan will lead to the rejection of a lot even though the sample being inspected does not actually have any defective items.
In the case of process inspection by variables, the most commonly used sample sizes are between 1 and 25 units. Samples of two or three units are rarely used, due to their low sensitivity, and are used only when the cost of measurements is very high. Therefore, when greater sensitivity is desired in the graph, sample sizes should be larger.
Types of errors made
The types of errors that can be made during a quality inspection of a process are:
Thus, these errors generate errors in the classification of the products derived from the process, and lead to undesirable consequences in the quality control of a company.
Error production factors
Entre los tipos de factores que pueden producir errores en la inspección de los productos, resaltan:.
An incorrect and/or incomplete/insufficient choice of the characteristics to be inspected
Not giving enough importance to some characteristics that directly affect the quality of the final products or giving it on the other hand to those that really have no impact on the final result, allocating unnecessary personal and economic resources. Choose the characteristics to inspect that give us confidence that a process is within the quality parameters that we determine, maintaining a relationship with a reasonable implicit economic cost. Since a disproportionate choice of characteristics would slow down the inspection process and considerably increase its cost, being far from any business justification.
Inadequate location of inspection point locations
Knowing where the different inspection points should be located is very important, since the resources available to a company are neither unlimited nor free, all of this has an associated cost that is not always feasible to pass on to the value of the product and its consequent benefit to be marked. To carry out the described task, first of all, an exhaustive study of the technological process or processes carried out in the company must be carried out, and from there decide which interlocks are necessary for inspection. Typically, about five locations are established for inspection points:
Do not leave the tasks of: handling, storage, packaging and shipping of materials without inspection points.
Therefore, it is advisable, once production has been completed, to assess and make a balance between the means used in the inspection and the cost involved, seeing how many defective parts can be assumed, without this being counterproductive.
Inappropriate determination of the form and typology of inspection according to each previously determined inspection point
Not all inspection points have to proceed in the same way, since in each of them the operators will review, manually or automated, the products for confirmation or rejection, according to the determined quality parameters.
When determining the form of inspection, one usually proceeds by:.
Ignorance of the characteristics of the object to be inspected, as well as those of each of these shapes, can generate errors, and it is very important to select the most suitable one.
Inappropriate choice of sample size composition to be taken in the inspection
As is known, sample size is essential to be able to safely determine whether a process is under control or not. In order to make decisions about a process that directly affect the quality derived from and impacted on its products, it is essential to correctly choose the composition of the sample to be extracted for study, without this entailing a disproportionate extra cost. Generally, a 100% inspection is relegated to the final testing of special or complex products, seeking to offer the consumer products free of defects. 100% inspection is usually carried out in automated inspections, while manual inspections involve taking production samples.
Carrying out a sample inspection has a number of advantages over 100% inspection. Performing routine, repetitive and monotonous tasks produces physical and psychological fatigue in inspectors, which can sometimes lead them to make errors in judgment, decision and classification of suitable and unsuitable pieces. It is sensible to apply sampling inspection in the following situations, as pointed out by Kaoru Ishikawa, in his "Ishikawa Quality Principles"):
Other situations where a sampling inspection can be established may be:
Finally, an aspect that can help decide the usefulness of sampling inspection is the balance that must be made between the cost of inspecting an item or a quality characteristic and the cost of conforming to that defective item.
Determination of the composition of the units that make up the sample to be taken to carry out the inspection
The most commonly used sample sizes are n = 25 units, assuming 5% of defects in production and up to 10% if it is a process with constant variations in production.
Mark the frequency with which the sample to be inspected must be extracted at each of the inspection points.
When setting the inspection frequency, the following aspects must be kept in mind:.
Quality characteristics at each inspection point
Uniformity in measurements will be determined by the following factors:.
Determination of the forms of registration and processing of resulting information, according to the various assigned inspection points
A set of specific models will be designed and implemented in accordance with the purpose of the inspection; That is, if the purpose is preventive, control charts will be established, so the model will respond to their demands, depending on the type of chart to be used. If, on the other hand, the purpose is to accept or discard a product, the model will be established in correspondence with the sampling plan to be used. When designing these models, the modality in which the information is processed and treated must be kept in mind, whether it is manual or automated, as well as establishing its flow, course, sharing and storage.
Manual inspection
Generally, this type of inspection is usually used when a sampling inspection of a process is carried out.
These are, in many cases, expensive and laborious processes, although the introduction of an automated inspection makes said action even more expensive above the labor cost of the operators in charge of supervising and checking a specific process. Manual inspections are usually carried out after the production chain has finished, resulting in a delay in carrying out said inspection once the product is completely finished. This delay (delay) usually entails an additional cost.
Not all technologically based parts and components of a process can be inspected in an automated manner, hence the still relevance of manual inspection depending on certain processes. Human mental interpretation and processing in certain cases continues to be irreplaceable, with the fatigue and possible errors committed that it implies.
Due to human errors, through 100% manual inspection, it is not possible to ensure that the quality of the products is also 100%.
Automated inspection
In cases where a company is equipped with appropriate infrastructure and the process carried out allows it, automated inspection facilitates the inspection of 100% of the manufactured products. This type of inspection avoids delays, as it can be carried out in a more direct, localized and appropriate manner, thus reducing possible increases in the manufacturing costs of a product, thanks to the timely detection of possible defects in it. As mentioned in the point referring to error production factors, correct location of the various inspection points is of vital importance. It is an important advantage, in a 100% inspection, that the probability of the lot being accepted is 1; in case the quality is equal to or less than the AQL and 0 if the quality is less than the AQL. However, the AQL should be set to q=0.
However, this type of inspection is not free from certain disadvantages that sometimes prevent proper supervision, for example:
That is why research was carried out in the field of probability theories and statistics, reaching the conclusion that to make decisions about the quality of in-process and finished production, there is no need to carry out a 100% inspection on all items, but it is enough to inspect only a part of the lot, that is, a sample, through a sampling inspection.
Some of the factors to consider in sampling inspection will be the level of trust in suppliers, the cost incurred by accepting defective products, and the risk of sampling, which will always exist due to the statistical nature of the process. In general, there are two types of errors that are likely to occur, the first is called type I error, and occurs when we reject a lot that meets quality specifications, and the second is called type II error, and occurs when we accept a lot that does not meet quality specifications.
Within quality sampling inspection, two types of inspection are mainly distinguished to control production processes. These processes are called “Inspection by Attributes” and “Inspection by Variables”.
Inspection by attributes can be considered that type of inspection of random samples of units in which each article or product is classified according to certain attributes as acceptable or defective, that is, it consists of finding out whether the material under consideration meets or does not meet what is specified, regardless of the measurement of the characteristic.
For inspection by attributes, the size of the samples and the interval between them must be such that approximately 5% of the production is inspected. In very massive processes that do not present frequent difficulties or the percentage of defective production is not serious, this percentage can be reduced to less than 5% where it is recommended that there must be at least 25 defectives in each sample to establish adequate behavior of the process.
Inspection by variables is a type of inspection that consists of measuring and recording a unit of measurement in which a specific quality characteristic is measured with a continuous scale to review its variation.
The statistical methods applicable to inspection by variables are based on the assumption of a normal distribution and not on a distribution of proportions as is the case with inspection by attributes. For the applicable methods, and with the measurements obtained, a statistic will be calculated, which will generally be a function of the sample mean and standard deviation, and depending on the value of this statistic when compared with a permissible value, the entire batch will be accepted or rejected.
The advantages that this method has with respect to the inspection method by attributes would be that the operating characteristic curve can be obtained with a smaller sample size than that required by a sampling plan by attributes. Furthermore, when destructive tests are used, sampling by variables is particularly useful to reduce inspection costs. On the other hand, measurement data typically provides more information about the lot than attribute data.
On the contrary, the distribution of the quality characteristic must be known, a plan must be used for each quality characteristic that must be inspected, and it is possible that the use of a variable sampling plan will lead to the rejection of a lot even though the sample being inspected does not actually have any defective items.
In the case of process inspection by variables, the most commonly used sample sizes are between 1 and 25 units. Samples of two or three units are rarely used, due to their low sensitivity, and are used only when the cost of measurements is very high. Therefore, when greater sensitivity is desired in the graph, sample sizes should be larger.
Types of errors made
The types of errors that can be made during a quality inspection of a process are:
Thus, these errors generate errors in the classification of the products derived from the process, and lead to undesirable consequences in the quality control of a company.
Error production factors
Entre los tipos de factores que pueden producir errores en la inspección de los productos, resaltan:.
An incorrect and/or incomplete/insufficient choice of the characteristics to be inspected
Not giving enough importance to some characteristics that directly affect the quality of the final products or giving it on the other hand to those that really have no impact on the final result, allocating unnecessary personal and economic resources. Choose the characteristics to inspect that give us confidence that a process is within the quality parameters that we determine, maintaining a relationship with a reasonable implicit economic cost. Since a disproportionate choice of characteristics would slow down the inspection process and considerably increase its cost, being far from any business justification.
Inadequate location of inspection point locations
Knowing where the different inspection points should be located is very important, since the resources available to a company are neither unlimited nor free, all of this has an associated cost that is not always feasible to pass on to the value of the product and its consequent benefit to be marked. To carry out the described task, first of all, an exhaustive study of the technological process or processes carried out in the company must be carried out, and from there decide which interlocks are necessary for inspection. Typically, about five locations are established for inspection points:
Do not leave the tasks of: handling, storage, packaging and shipping of materials without inspection points.
Therefore, it is advisable, once production has been completed, to assess and make a balance between the means used in the inspection and the cost involved, seeing how many defective parts can be assumed, without this being counterproductive.
Inappropriate determination of the form and typology of inspection according to each previously determined inspection point
Not all inspection points have to proceed in the same way, since in each of them the operators will review, manually or automated, the products for confirmation or rejection, according to the determined quality parameters.
When determining the form of inspection, one usually proceeds by:.
Ignorance of the characteristics of the object to be inspected, as well as those of each of these shapes, can generate errors, and it is very important to select the most suitable one.
Inappropriate choice of sample size composition to be taken in the inspection
As is known, sample size is essential to be able to safely determine whether a process is under control or not. In order to make decisions about a process that directly affect the quality derived from and impacted on its products, it is essential to correctly choose the composition of the sample to be extracted for study, without this entailing a disproportionate extra cost. Generally, a 100% inspection is relegated to the final testing of special or complex products, seeking to offer the consumer products free of defects. 100% inspection is usually carried out in automated inspections, while manual inspections involve taking production samples.
Carrying out a sample inspection has a number of advantages over 100% inspection. Performing routine, repetitive and monotonous tasks produces physical and psychological fatigue in inspectors, which can sometimes lead them to make errors in judgment, decision and classification of suitable and unsuitable pieces. It is sensible to apply sampling inspection in the following situations, as pointed out by Kaoru Ishikawa, in his "Ishikawa Quality Principles"):
Other situations where a sampling inspection can be established may be:
Finally, an aspect that can help decide the usefulness of sampling inspection is the balance that must be made between the cost of inspecting an item or a quality characteristic and the cost of conforming to that defective item.
Determination of the composition of the units that make up the sample to be taken to carry out the inspection
The most commonly used sample sizes are n = 25 units, assuming 5% of defects in production and up to 10% if it is a process with constant variations in production.
Mark the frequency with which the sample to be inspected must be extracted at each of the inspection points.
When setting the inspection frequency, the following aspects must be kept in mind:.
Quality characteristics at each inspection point
Uniformity in measurements will be determined by the following factors:.
Determination of the forms of registration and processing of resulting information, according to the various assigned inspection points
A set of specific models will be designed and implemented in accordance with the purpose of the inspection; That is, if the purpose is preventive, control charts will be established, so the model will respond to their demands, depending on the type of chart to be used. If, on the other hand, the purpose is to accept or discard a product, the model will be established in correspondence with the sampling plan to be used. When designing these models, the modality in which the information is processed and treated must be kept in mind, whether it is manual or automated, as well as establishing its flow, course, sharing and storage.
Manual inspection
Generally, this type of inspection is usually used when a sampling inspection of a process is carried out.
These are, in many cases, expensive and laborious processes, although the introduction of an automated inspection makes said action even more expensive above the labor cost of the operators in charge of supervising and checking a specific process. Manual inspections are usually carried out after the production chain has finished, resulting in a delay in carrying out said inspection once the product is completely finished. This delay (delay) usually entails an additional cost.
Not all technologically based parts and components of a process can be inspected in an automated manner, hence the still relevance of manual inspection depending on certain processes. Human mental interpretation and processing in certain cases continues to be irreplaceable, with the fatigue and possible errors committed that it implies.
Due to human errors, through 100% manual inspection, it is not possible to ensure that the quality of the products is also 100%.
Automated inspection
In cases where a company is equipped with appropriate infrastructure and the process carried out allows it, automated inspection facilitates the inspection of 100% of the manufactured products. This type of inspection avoids delays, as it can be carried out in a more direct, localized and appropriate manner, thus reducing possible increases in the manufacturing costs of a product, thanks to the timely detection of possible defects in it. As mentioned in the point referring to error production factors, correct location of the various inspection points is of vital importance. It is an important advantage, in a 100% inspection, that the probability of the lot being accepted is 1; in case the quality is equal to or less than the AQL and 0 if the quality is less than the AQL. However, the AQL should be set to q=0.
However, this type of inspection is not free from certain disadvantages that sometimes prevent proper supervision, for example: