A series of articles on five tools of quality management, seven methods, and quality control

Recently, many quality colleagues asked us if we could introduce the commonly used quality management system tools and methods. This is a big question, and each tool, method has a lot of content. But it doesn't matter, with this proposition, I made a sorting out, a simple version of the five tools and seven methods, decisively collected them, and shared them with the quality people around me. It is recommended to read on weekends and grow knowledge together.

five tools

APQP

APQP (Advanced Product Quality Planning) is a structured method for determining and formulating the steps required to ensure that a product satisfies customers.

The goal of Product Quality Planning is to facilitate contact with everyone involved to ensure that the required steps are completed on time. Effective product quality planning depends on the commitment of the company's top management to the goal of striving to achieve customer satisfaction.

Product quality planning has the following benefits:

to direct resources to satisfy customers;
to facilitate early identification of required changes;
to avoid late changes;
to provide quality products in a timely manner at the lowest cost.

FMEA

FMEA (Potential Failure Mode and Effects Analysis) is the analysis of potential failure modes and consequences. It is in the planning and design stage of products/processes/services. Analyze one by one to find potential failure modes, analyze their possible consequences, and assess their risks, so as to take measures in advance to reduce the serious procedures of failure modes and reduce the probability of their possible occurrence, so as to effectively improve quality and reliability, and ensure Systematic activities for customer satisfaction.

Types of FMEA:
According to its application field, common FMEA includes design FMEA (DFMEA) and process FMEA (PFMEA). Others include system FMEA, application FMEA, procurement FMEA, and service FMEA.

MSA

MSA (Measurement System Analysis) is MSA measurement system analysis, which uses mathematical statistics and chart methods to analyze the error of the measurement system to evaluate whether the measurement system is suitable for the measured parameters and determine the main components of the measurement system error .

PPAP

PPAP (Production part approval process) is the production part approval process, which is a control program for production parts and a management method for quality.

PPAP production part submission guarantee: mainly includes production part size inspection report, appearance inspection report, function inspection report, material inspection report, plus some parts control methods and supplier control methods; manufacturing enterprises require suppliers to do PPAP when submitting

products Documents and the first piece can only be submitted when all PPAP documents are qualified; when the project is changed, a report must be submitted.

SPC

SPC (Statistical Process Control) is statistical process control, which mainly refers to the application of statistical analysis technology to monitor the production process in a timely manner, scientifically distinguish random fluctuations and abnormal fluctuations of product quality in the production process, and thus provide early warning for abnormal trends in the production process. In order for the production management personnel to take measures in time to eliminate abnormalities and restore the stability of the process so as to achieve the purpose of improving and controlling quality.

SPC is very suitable for repetitive production processes. It can help organizations make reliable assessments of the process, determine the statistical control limits of the process, judge whether the process is out of control and whether the process is capable; provide an early warning system for the process, and monitor the process in time. , in order to prevent the production of waste products, reduce the dependence on routine inspection, regularly replace a large number of inspection and verification work with observation and systematic measurement methods.

Significance of SPC implementation:

It can enable enterprises to: reduce costs; reduce defective rate, reduce rework and waste; improve labor productivity; provide core competitiveness; win a wide range of customers.

There are two stages of SPC implementation:

Analysis stage: use control charts, histograms, process capability analysis, etc. to make the process in a statistical steady state, so that the process capability is sufficient.
Monitoring stage: Use control charts and other monitoring processes.

The emergence of SPC:

After the industrial revolution, with the further development of productivity and the formation of large-scale production, how to control the quality of large-scale products has become a prominent problem. Improve quality management methods. Therefore, the United Kingdom, the United States and other countries began to study the quality control method that replaced the post-mortem inspection with statistical methods.

In 1924, Dr. Shewhart of the United States proposed to apply the 3Sigma principle to the production process, and published the famous "control chart method" to control process variables, which laid the theoretical and methodological foundation for statistical quality management.

The role of SPC:

① Ensure that the process is continuously stable and predictable.
② Improve product quality, production capacity, and reduce costs.
③ Provide basis for process analysis.
④ Distinguish between special causes of variation and common causes, as a guide for taking local measures or taking measures for the system.

seven techniques

checklist

The checklist is a method of listing the contents or items that need to be checked one by one, and then checking them one by one on a regular or irregular basis, and recording the problem points, sometimes called a checklist or a spot checklist. For example: inspection form, diagnosis form, work improvement checklist, satisfaction survey form, assessment form, audit form, 5S activity checklist, engineering anomaly analysis form, etc.

Component elements:

① determine the items to be inspected;
② determine the frequency of inspection;
③ determine the personnel to be inspected.

Implementation steps:

① Determine the inspection object;
② Formulate the checklist;
③ Check and record according to the items in the checklist;
④ Request the responsible unit to improve the problems found in the inspection in a timely manner;
⑤ The inspectors confirm the improvement effect within the specified time;
⑥ Periodic summary and continuous improvement.

stratification method

The classification method is to classify a large number of viewpoints, opinions or ideas on a specific topic into groups, and to group a large amount of collected data or materials according to their mutual relationship, and to classify them. The stratification method is generally used in combination with other seven methods such as Plato and histogram, or it can be used alone. For example: sampling statistics table, bad category statistics table, ranking list, etc.

Implementation steps:

① Determine the subject of the research;
② Make tables and collect data;
③ Stratify the collected data;
④ Comparative analysis, analyze the data, find out the internal reasons, and determine the improvement items.

Plato

The use of Plato is based on the premise of the stratification method, and the items determined by the stratification method are arranged in descending order, and the graph of the cumulative value is added. It can help us identify key issues, grasp the important few and the useful majority, and is suitable for counting value statistics. Some people call it an ABC chart, and because Plato's sorting consciousness is from large to small, it is also called a permutation chart.

Classification:

① Analyzing phenomena using Plato: related to bad results, used to find major problems.
A Quality: non-conformity, failure, customer complaints, returns, repairs, etc.;
B Cost: total loss, expenses, etc.;
C delivery time: stock shortages, payment defaults, delivery delays, etc.;
D Safety: accidents, mistakes wait.

② Plato is used to analyze the reasons: it is related to the process factors and used to find the main problems.
A Operator: shift, group, age, experience, proficiency, etc.;
B Machine: equipment, tools, molds, instruments, etc.;
C Raw materials: manufacturer, factory, batch, type, etc.;
D Operation method: operating environment, Process sequence, job arrangement, etc.

The role of Plato:

① Reduce the bad basis;
② Determine the improvement target and find out the problem;
③ Can confirm the effect of improvement.

Implementation steps:

① Collect data, classify by layer method, and calculate the percentage of each layer item in the overall project;
② Summarize the classified data, arrange them from most to least, and calculate the cumulative percentage;
③ Draw the horizontal axis and vertical axis scale;
④ Draw a histogram;
⑤ Draw a cumulative curve;
⑥ Record necessary items
⑦ Analyze Plato
⑧ Key points:

A Plato has two ordinates, the left ordinate generally represents quantity or amount, and the right ordinate generally represents the cumulative percentage of quantity or amount; B Plato’s abscissa generally represents
inspection items, according to the degree of influence, from left to right Arranged in order;
C When drawing Plato, according to the frequency of the number of items or the amount of money, draw a rectangle corresponding to the vertical axis on the left, draw the points corresponding to the cumulative frequency of each item, and draw these points in order Connected into lines.

Application points and precautions:

① Plato should be retained, and the improvement effect can be evaluated by arranging the pre-improved and improved Plato together; ② To
analyze Plato, you only need to grasp the first 2~3 items;
③ Plato’s classification items Don’t set too little, 5~9 items are suitable for teaching. If there are too many classification items, more than 9 items, they can be classified into other categories. If there are too few classification items, less than 4 items, it is meaningless to be a Plato; ④ If the Plato is
created When it is found that the allocation ratio of each item is similar, Plato will lose its meaning, which is inconsistent with Plato’s law. Data should be collected from other angles for analysis; ⑤
Y Plato is a means of management improvement rather than an end. If the data items are already clear, there is no need to waste Time to make Plato;
⑥ If other items are greater than the previous ones, they must be analyzed at different levels to check whether there are reasons
; When the existing conditions are difficult to solve, or even if it is solved, it will cost a lot and the gain is not worth the loss, then you can avoid the first item and start with the second item.

causal diagram

The so-called cause-and-effect diagram, also known as characteristic factor diagram, is mainly used to analyze the causal relationship between quality characteristics and possible causes affecting quality characteristics. A tool for quality characteristics (effects) and factors (causes) that may affect the characteristics. Also known as fishbone diagram.

Classification:

① Cause-pursuing type: It is to seek the cause of the problem and find its impact, and use the causal diagram to show the relationship between the result (characteristic) and the cause (cause); ② Pursue countermeasures: Pursue how to prevent the problem and how to achieve the goal
. And the causal diagram shows the relationship between the expected effect and the countermeasures.

Implementation steps:

① Set up a cause-and-effect diagram analysis team, preferably with 3 to 6 people, preferably representatives of each department; ②
Determine the problem points;
③ Draw the main bone, middle bone, and small bone of the main line and determine the major causes (generally from 5M1E, that is, people Man, Machine, Material, Method, Measure, and Environment to comprehensively find out the reasons);
④ Participants discussed enthusiastically, analyzed according to major reasons, found medium or small reasons, and drew them in the cause-effect diagram;
⑤ The cause-and-effect diagram group should form a consensus, and mark the items that are most likely to be the source of the problem with a red pen or a special mark; ⑥
Write down the necessary

items

. omissions;
②The more detailed the cause analysis, the better, and the more detailed the rules, the better the key reason or solution to the problem can be found; ③As
many quality characteristics as there are, how many cause-and-effect diagrams should be drawn;
④If no measures can be taken for the analyzed reasons, explain the problem It has not been solved yet. If the improvement is to be effective, the reasons must be subdivided until measures can be taken;
⑤ Objectively evaluate the importance of each factor on the basis of data;
⑥ Focus on solving problems, and list them item by item according to the 5W2H method. When drawing a cause-and-effect diagram, focus first on "why this cause and result occurred", and put forward countermeasures after analysis on "how to Only then can it be solved”;

Why——Why should we do it? (subject)

What - what to do? (Purpose)

Where - where to do it? (Place)

When - when to do it? (Sequence)

Who - who will do it? (people)

How - how to do it? (Means)

How much - how much does it cost? (Expenses)

⑦ The cause-and-effect diagram should be considered based on the problems that occur on the spot;
⑧ After the cause-and-effect diagram is drawn, a consensus must be reached before determining the key causes, and marked with a red pen or a special mark;
⑨ The use of the cause-and-effect diagram should be continuously improved.

Scatterplot

The data corresponding to the causal relationship is depicted on the XY axis coordinate system, so as to grasp whether the two variables are correlated and the degree of correlation. This kind of graphics is called "scatter diagram", also known as "correlation diagram".

Classification:

① Positive correlation: When variable X increases, another variable Y also increases;
② Negative correlation: When variable X increases, another variable Y decreases;
③ Irrelevant: Variable X (or Y) When it changes, the other variable does not change;
④ Curve correlation: When the variable X starts to increase, Y also increases, but after reaching a certain value, when the value of X increases, Y decreases instead. ;

Implementation steps:

① Determine the two variables to be investigated, collect the latest relevant data, at least 30 groups;
② Find the maximum and minimum values of the two variables, and draw the two variables on the X-axis and Y-axis;
③ Mark the corresponding two variables on the coordinate system in the form of dots;
④ Include items such as the title of the map, creator, and production time;
⑤ Interpret the correlation and degree of correlation of the scatter diagram.

Application points and precautions:

① The corresponding number of two groups of variables should be at least 30 groups, preferably 50 to 100 groups. When the data is too small, it is easy to cause misjudgment; ② Usually, the
abscissa is used to indicate the cause or independent variable, and the ordinate Indicates the effect or dependent variable;
③ Since the data acquisition is often affected by the change of 5M1E, the correlation of the data is affected. In this case, it is necessary to classify the conditions of the data acquisition, otherwise the scatter diagram cannot truly reflect the two variables
④ When an abnormal point appears, the reason should be found immediately, and the abnormal point cannot be deleted; ⑤
When the correlation of the scatter diagram does not match the technical experience, it should be further examined whether there is any reason for the false appearance.

histogram

The histogram is aimed at the characteristic value of a certain product or process, using the principle of normal distribution (also called normal distribution), to group more than 50 data, and calculate the number of occurrences of each group, and then use similar histograms to depict them in the on the horizontal axis.

Implementation steps:

① Collect the same type of data;
② Calculate the range (full distance) R=Xmax-Xmin;
③ Set the number of groups K: K=1+3.23logN

④ Determine the minimum measurement unit, that is, when the number of decimal places is n , the minimum unit is 10-n;
⑤ Calculate the group distance h, group distance h=range R/group number K;
⑥ Find the upper and lower limits of each group
The first group lower limit = X min - the smallest unit of measurement 10-n/27
The lower limit of the second group (the upper limit of the first group) = the lower limit of the first group + the group distance h;
⑦ Calculate the center value of each group, the center value of the group = (the lower limit of the group + the group upper limit)/2;
⑧ Make a frequency table;
⑨ Draw a histogram according to the frequency table.

The common form and judgment of the histogram:

① Normal type: It is a normal distribution, obeys the statistical law, and the process is normal;
② Missing tooth type: It is not a normal distribution,
does not obey the statistical law; Obey statistical laws;
④ Outlying island type: not normal distribution, do not obey statistical laws;
⑤ Plateau type: not normal distribution, do not obey statistical laws;
⑥ Bimodal: not normal distribution, do not obey statistical laws;
⑦ Irregular Type: Not a normal distribution, does not obey statistical laws.

Control Charts

There are many factors affecting product quality, including static and dynamic factors. Is there a way to monitor the production process of the product in real time and find hidden quality risks in time, so as to improve the production process and reduce the output of waste and defective products? The control chart method

is Such a prevention-based quality control method uses the quality characteristic values collected on site to draw a control chart, and judges the quality status of the product in the production process by observing the graphics. Control charts can provide a lot of useful information and are one of the important methods of quality management.

The meaning of the control chart method:

the control chart is also called the management chart, which is a quality management chart with control boundaries. One of the purposes of using the control chart is to analyze and judge whether there is an abnormality in the production process by observing the distribution of product quality characteristic values on the control chart. Once an abnormality is found, it is necessary to take necessary measures to eliminate it in time to restore the stability of the production process. state. Control charts can also be applied to bring production processes into a state of statistical control. The distribution of product quality characteristic values is a statistical distribution, therefore, drawing control charts requires the application of relevant theories and knowledge of probability theory.

The control chart is a recording graphic of the quality of the production process. There are center lines, upper and lower control limits on the chart, and numerical points reflecting the statistics of each sample drawn in chronological order. The center line is the mean value of the statistic being controlled, and the upper and lower control limits are several standard deviations away from the center line. Most manufacturing industries apply three standard deviation control limits, but other control limits may be used if there is sufficient evidence.

Commonly used control charts include measurement value and count value, which are respectively applicable to different production processes; each type can be subdivided into specific control charts, such as measurement value control charts can be specifically divided into mean value-range control chart, individual value-moving range control chart, etc.

Drawing of the control chart:

① The basic style of the control chart is shown in the figure, and the production of the control chart generally requires the following steps:

A. Take samples according to the specified sampling interval and sample size;
B. Measure the quality characteristic values of the samples and calculate their statistics. Quantitative value;
C draw points on the control chart;
D judge whether there is parallelism in the production process.

② When the control chart provides managers with a lot of useful production process information, they should pay attention to the following issues:

A. According to the quality of the process, choose the management point reasonably. Management points generally refer to key parts, key dimensions, special requirements of the process itself, and key points that have an impact on workers. For example, parts with unstable quality and many defective products can be selected as management points; B According to the management
point Reasonable selection of the type of control chart for the above quality problems:
C. When using the control chart for process management, a reasonable control limit should be determined first.
D. If the points on the control chart are abnormal, the cause should be found out immediately, and measures should be taken before production , which is the primary prerequisite for the control chart to function;
E The control line is not equal to the tolerance line, the tolerance line is used to judge whether the product is qualified, and the control line is used to judge whether the process quality has changed; F The control chart is abnormal
, Responsibilities must be clarified, resolved or reported in a timely manner.

Field sampling method:

Control limits are not calculated every time when making control charts, so how is the initial control line determined? If the current production conditions are similar to those in the past, past experience data can be followed, that is, the control limits of stable production in the past can be extended. The following introduces a method to determine the control limit, that is, the field sampling method.

The steps are as follows:

① Randomly select more than 50 samples, measure the data of the samples, calculate the control limit, and make a control chart;
② Observe whether the control chart is in the control state, that is, stable, if all the points are within the control limit. And if there is no abnormality in the arrangement of the points, you can go to the next step;
③ If there is an abnormal state, or if there is an abnormality in the arrangement although it has not exceeded the control limit, it is necessary to find out the cause of the abnormality and take appropriate measures to make it under control. state, and then re-take the data to calculate the control limit, and go to the next step;
④ Make a cubic diagram of the above-mentioned data, and compare the cubic diagram with the standard limit (tolerance upper limit and lower limit) to see if it is in an ideal state or a more ideal state , if the requirements are not met, measures must be taken to reduce the average shift or standard deviation. After taking measures, repeat the above steps to retake the data and make control limits until the standard is met.

How to use the control chart to judge abnormal phenomena:

use the control chart to identify the state of the production process, mainly based on the analysis and judgment based on the position of the sample point and the change trend formed by the sample data.

The out-of-control state mainly manifests in the following two situations:

the sample point exceeds the control limit,

and the sample point is within the control limit, but the arrangement is abnormal.

When the data points exceed the management limit, it is generally considered that there is an anomaly in the production process. At this time, the cause should be investigated and countermeasures should be taken. Arrangement abnormality mainly refers to the following situations:

A. More than seven consecutive points all deviate from above or below the center line. At this time, it should be checked whether the production conditions have changed.
B Two of the three consecutive points enter the vicinity of the management limit (referring to the area from the center line to more than two-thirds of the management limit), at this time, attention should be paid to whether the fluctuation of production is too large.
Point C successively shows an upward or downward trend, indicating that the process characteristics are changing upward or downward.
The arrangement state of point D changes periodically. At this time, the operating time can be hierarchically processed, and the control chart can be recreated to find out the cause of the problem.

The ability of the control chart to reveal abnormal phenomena will vary according to the amount of data in each group, the method of sample collection, and the division of strata. We should not be satisfied with the use of only one control chart, but should change various data collection methods and usage methods to make various types of charts, so as to receive better results.

It is worth noting that if anomalies beyond the boundaries of management are discovered, and no efforts are made to investigate the causes and take countermeasures, then the control chart, although useful, is just a piece of paper.

■ END ■

The above content comes from the Internet, and the copyright belongs to the original author. If it involves copyright issues, please contact us in time, and we will delete the content to protect your rights!