A complete set of quality management tools and quality control methods

Thread Content

APQP: APQP (Advanced Product Quality Planning) refers to advanced planning for product quality; it is a structured approach used to identify and establish the steps necessary to ensure that a product meets customer expectations. The goal of product quality planning is to foster connections with everyone involved, to ensure that the required steps are completed on time. Effective product quality planning relies on the commitment of top management to strive for customer satisfaction. Product quality planning has the following benefits: ◆ Directs resources to ensure customer satisfaction ； ◆ Facilitate early identification of required changes ； ◆ Avoid late changes ； ◆ Provide high-quality products in a timely manner at the lowest cost. SPC, or Statistical Process Control, refers to the use of statistical analysis techniques to monitor production processes in real time. It enables the accurate differentiation between random fluctuations and abnormal variations in product quality during production, thereby providing early warnings about any abnormal trends. This allows production managers to take timely actions to eliminate these abnormalities and restore process stability, with the goal of improving and controlling quality. SPC is highly suitable for repetitive production processes. It helps organizations make reliable evaluations of these processes, determine their statistical control limits, and assess whether the processes are out of control or capable ； Provide an early warning system for the process to monitor its conditions in a timely manner, thereby preventing the generation of waste and reducing reliance on routine inspections; regular observations and systematic measurement methods can replace extensive testing and verification efforts. (1) Significance of implementing SPC: It enables enterprises to: ◆ Reduce costs ◆ Lower defect rates, thereby reducing rework and waste ◆ Improve labor productivity ◆ Develop core competitiveness ◆ Attract a wide range of customers. (2) Two phases of SPC implementation: Analysis phase: Utilize control charts, histograms, process capability analysis, etc., to bring the process into a state of statistical stability and ensure sufficient process capability. Monitoring phase: Using control charts and similar tools to monitor the process. (3) The emergence of SPC: After the Industrial Revolution, as productivity continued to develop and mass production came into being, how to control the quality of large quantities of products became a significant issue. Quality control methods that relied solely on post-production inspection were no longer sufficient to meet the demands of economic development at that time, so it was necessary to improve quality management approaches. Thus, countries such as the UK and the US began to research quality control methods that use statistical methods to replace post-hoc testing. In 1924, Dr. Shewhart in the United States proposed applying the 3Sigma principle to production processes, and introduced the famous \"control chart method\" for controlling process variables, thereby laying the theoretical and methodological foundation for statistical quality control. (4) Role of SPC: 1. Ensure the process remains continuous, stable, and predictable. 2. Improve product quality and production capacity, and reduce costs. 3. Provide a basis for process analysis. 4. Distinguish between special and common causes of variation to serve as a guide for taking localized actions or system-wide measures. FMEA, or Potential Failure Mode and Effects Analysis, is a systematic approach used during the planning and design phase of products/processes/services. It involves analyzing each subsystem and component that makes up a product, as well as various procedures involved in a process or service, in order to identify potential failure modes, assess their possible consequences, and evaluate the associated risks. By taking preventive measures, it is possible to reduce the severity of these failure modes and lower their likelihood of occurrence, thereby effectively improving quality and reliability and ensuring customer satisfaction. Types of FMEA: The common types of FMEA based on their areas of application are Design FMEA (DFMEA) and Process FMEA (PFMEA); other types include System FMEA, Application FMEA, Procurement FMEA, and Service FMEA. MSA: Short for Measurement System Analysis. MSA involves using mathematical statistics and graphical methods to analyze the errors in a measurement system, in order to determine whether such a system is suitable for measuring the desired parameters, as well as to identify the main sources of error in the measurement system. PPAP: Production Part Approval Process. It is a control procedure for manufactured parts, as well as a quality management method. PPAP – Production Part Approval Process: It mainly includes reports on the dimensions of the produced parts, visual inspection reports, functional inspection reports, and material inspection reports; in addition, there are also methods for controlling the parts and the suppliers ； Mainly, manufacturing companies require suppliers to submit PPAP documents and first articles along with the products; the submission is only allowed once all the PPAP documents meet the requirements ； A report must also be submitted after the engineering change. 2. Seven methods: 1. Checklist – A checklist is a method in which the items or aspects that need to be checked are listed one by one; thereafter, these items are checked regularly or irregularly, and any issues identified are recorded. It is sometimes also referred to as an inspection sheet or a check sheet. For example: inspection sheets, diagnostic forms, work improvement checklists, satisfaction surveys, evaluation forms, audit forms, 5S activity checklists, engineering anomaly analysis sheets, etc. 1. Components ① Determine the items to be inspected ； ②Determine the frequency of inspections ； ③Identify the personnel responsible for the inspection. 2. Implementation steps ① Determine the objects to be inspected ； ②Create a checklist ； ③Check according to the checklist items and record them ； ④The units responsible are required to make timely improvements to the issues identified during the inspection ； ⑤Inspectors verify the effectiveness of the improvements within the specified time frame ； ⑥Conduct regular summaries for continuous improvement. The 2-level stratification method: Stratification involves grouping a large number of opinions, views, or ideas related to a specific topic into categories, and organizing the extensive amount of data or information collected by grouping them based on their interrelationships, thereby creating layers. The stratification method is generally used in combination with other seven techniques such as Pareto charts and histograms, but it can also be used alone. For example: sampling statistics tables, defect category statistics tables, rankings, etc. Implementation steps: ① Determine the research topic ； ②Create a table and collect data ； ③Stratify the collected data ； ④Comparative analysis: Analyze these data to identify their underlying causes and determine areas for improvement. 3 Plato: The use of Plato relies on the stratification method; the items identified by this method are arranged from largest to smallest, along with a graph showing the cumulative values. It helps us identify key issues, pinpoint the vital few and the useful many. It is applicable to quantitative data statistics. Some call it an ABC chart; since the items are ranked from largest to smallest according to Pareto’s principle, it is also known as a Pareto chart. 1. Classification 1) Using Plato to analyze phenomena: Related to adverse outcomes, it is used to identify the main problems. Quality A: Unqualified, faulty, customer complaints, returns, repairs, etc ； Cost B: Total losses, expenses, etc ； C Delivery timeline: stock shortages, payment defaults, delivery delays, etc ； D Safety: Accidents, errors, etc. occur. 2) Use Plato to analyze the reasons: It is related to process factors and is used to identify the main problems. Operator A: Shift, group, age, experience, proficiency level, etc ； Machine B: Equipment, tools, molds, instruments, etc ； C Raw materials: manufacturer, factory, batch, type, etc ； D Work methods: working environment, sequence of processes, work scheduling, etc. 2. The role of Plato: ① Reducing the basis for negative outcomes ； ② Decide on improvement goals and identify the problem areas ； ③The effect of the improvement can be confirmed. 3. Implementation steps: ① Collect data, classify it using the stratification method, and calculate the percentage of items in each stratum relative to the total number of items ； ②Aggregate the categorized data, arrange them from highest to lowest, and calculate cumulative percentages ； ③Draw horizontal and vertical axis scales ； ④Draw a bar chart ； ⑤ Draw the cumulative curve ； ⑥Items to record ⑦ Analyzing Plato 【Key points】 A. Plato has two vertical axes; the left vertical axis generally represents quantity or amount, while the right vertical axis represents the cumulative percentage of quantity or amount ； B. In Plato’s diagram, the horizontal axis generally represents the inspection items, which are arranged from left to right in order of their degree of impact ； C. When creating a Pareto chart, rectangular bars are drawn on the left vertical axis corresponding to the frequency of occurrence of each item in terms of quantity or amount, while points are marked on the right vertical axis representing the cumulative frequency of occurrences for each item; these points are then connected in sequence to form a line. 4. Key application points and precautions: ① It is necessary to retain the Pareto charts; by placing those from before and after the improvement side by side, it is possible to assess the effect of the improvements ； ②To analyze Plato, it’s sufficient to focus on the first 2–3 items ； ③Don’t set too few categories in a Pareto chart; 5 to 9 categories is appropriate. If there are too many categories—more than 9—they can be grouped under “Other”. Conversely, if there are too few categories—less than 4—creating a Pareto chart would be pointless ； ④ If the resulting Pareto chart shows that the distribution ratios of various items are roughly similar, then the Pareto chart loses its significance; it does not conform to the Pareto principle. In such cases, data should be collected from other perspectives for further analysis ； ⑤ A Pareto chart is a means to improve management, not an end in itself; if the data items are already clear, there is no need to waste time creating a Pareto chart ； ⑥If other items are larger than the previous ones, it is necessary to conduct a hierarchical analysis to examine whether there are any underlying reasons for this ； ⑦ The main purpose of Plato’s analysis is to identify the key issues based on the information gathered in order to devise countermeasures. However, if the first item is difficult to address given the existing conditions, or if resolving it would require excessive effort and not be worth it, then it is possible to bypass that first item and start with the second one. 4 Cause-and-Effect Diagram. Also known as a characteristic-cause diagram, it is primarily used to analyze the causal relationships between quality characteristics and the potential causes affecting those characteristics. By understanding the current situation, analyzing the causes, and identifying appropriate measures, it helps in resolving problems. It serves as a tool for analyzing quality characteristics (results) and the factors (causes) that may influence them. Also known as a fishbone diagram. 1. Classification 1) Cause-seeking type: It focuses on identifying the causes of a problem and examining their impacts, using cause-and-effect diagrams to show the relationship between outcomes (characteristics) and causes (factors) ； 2) Countermeasure-seeking type: Focuses on how to prevent problems and how to achieve goals; it uses a causal diagram to illustrate the relationship between desired effects and countermeasures. 2. Implementation steps: ① Form a cause-and-effect diagram analysis team; 3 to 6 members is ideal, with representatives from various departments preferred ； ②Identify the problem areas ； ③ Draw the main skeleton, intermediate bones, and minor bones of the backbone, and identify the root causes (generally, these are found by examining all six aspects of 5M1E: Man, Machine, Material, Method, Measure, and Environment) ； ④The participants had lively discussions, conducted analyses based on significant factors, identified intermediate or minor causes, and plotted them on a cause-and-effect diagram ； ⑤ The causal diagram team needs to reach a consensus on identifying the items that are most likely to be the root cause of the problem, and mark them with red pen or special symbols ； ⑥ Record the necessary items. 3. Key application points and precautions: ① To determine the cause, it is necessary to draw on the knowledge and experience of all team members and encourage collective brainstorming to avoid omissions ； ②The more detailed the analysis of the causes, the better; the greater the detail, the easier it is to identify the key causes or solutions to the problem ； ③There should be as many cause-and-effect diagrams as there are quality characteristics ； ④ If no actions can be taken based on the identified reasons, it means the problem has not been resolved. To achieve effective improvements, the reasons must be broken down further until actions can be taken ； ⑤Objectively evaluate the importance of each factor based on data ； ⑥Focus on solving the problem, and list the various elements using the 5W2H method. When creating a cause-and-effect diagram, prioritize understanding \"why such causes and effects occur\"; once analysis is complete and solutions need to be proposed, focus on \"how to resolve the issue\"” ； Why——why do it? (Object) What—what to do? (Purpose) Where——Where to do it? (Location) When – When should it be done? (Sequence) Who——Who will do it? (Person) How—what method to use? (Methods) How much—how much does it cost? (Fees) ⑦ The cause-and-effect diagram should be based on the problems that occur on site ； ⑧After drawing the causal diagram, a consensus should be reached to determine the cause, which should then be marked with red pen or a special symbol ； ⑨ Causal diagrams must be continuously improved when used. 5 Scatter plot: Data showing the changes associated with a causal relationship are plotted on the X-Y coordinate axes, in order to determine whether there is a correlation between the two variables and to assess the extent of that correlation. This type of graph is called a “scatter plot,” or also a “correlation diagram.” 1. Classification 1) Positive correlation: When variable X increases, the other variable Y also increases ； 2) Negative correlation: As variable X increases, the other variable Y decreases ； 3) Unrelated: When variable X (or Y) changes, the other variable does not change ； 4) Curve correlation: As variable X begins to increase, Y also increases, but after reaching a certain value, Y decreases as X continues to increase. ; 2. Implementation steps 1) Identify the two variables to be investigated, and collect the most up-to-date relevant data, with at least 30 data sets ； 2) Find the maximum and minimum values of the two variables, and plot the two variables on the X-axis and Y-axis respectively ； 3) Mark the coordinates of the corresponding two variables on the coordinate system in the form of points ； 4) Include items such as map title, creator, and creation date ； 5) Interpret the correlation and degree of correlation in the scatter plot. 3. Key application points and precautions: 1) There should be at least 30 pairs of corresponding variables; ideally, 50 to 100 pairs. Too few data points can lead to incorrect conclusions ； 2) Usually, the horizontal axis is used to represent the cause or independent variable, while the vertical axis represents the effect or dependent variable ； 3) Since the acquisition of data is often affected by changes in 5M1E, the correlation of the data is impacted; in such cases, it is necessary to stratify the conditions under which the data is obtained, otherwise the scatter plot cannot accurately reflect the relationship between the two variables ； 4) When an abnormality occurs, the cause should be identified immediately, rather than deleting the abnormality ； 5) When the correlation in a scatter plot does not match technical experience, it is necessary to further investigate whether there are any reasons causing this illusion. 6 Histograms A histogram is a tool used for analyzing the characteristic values of a certain product or process. By applying the principles of the normal distribution, more than 50 data points are grouped together, and the frequency of occurrence for each group is calculated; these values are then represented graphically using a histogram, with the horizontal axis showing the relevant values. 1. Implementation steps 1) Collect data of the same type ； 2) Calculate the range (full span) R = Xmax - Xmin ； 3) Set the number of groups K: K=1. 3.23logN. Total number of data points: 50–100, 100–250, over 250. Corresponding ranges for the total number: 6–10, 7–12, 10–20. 4) Determine the minimum unit of measurement; when the number of decimal places is n, the minimum unit is 10-n ； 5) Calculate the class interval h; h = range R / number of classes K ； 6) Determine the upper and lower limits for each group. The lower limit for the first group = X_min – 10/n/27. The lower limit for the second group (which is also the upper limit of the first group) = Lower limit of the first group + interval width h ； 7) Calculate the central value for each group; Group central value = (Lower limit of the group + Upper limit of the group) / 2 ； 8) Create a frequency table ； 9) Draw a histogram based on the frequency table. 2. Common shapes of histograms and their identification: 1) Normal type: It follows a normal distribution, adheres to statistical laws, and the process is normal ； 2) Missing tooth pattern: Not normally distributed, does not follow statistical laws ； 3) Skewed type: Not normally distributed, does not follow statistical laws ； 4) Island type: Not normally distributed, does not follow statistical laws ； 5) Plateau type: Not normally distributed, does not follow statistical laws ； 6) Bimodal: Not a normal distribution, does not follow statistical laws ； 7) Irregular type: It is not normally distributed and does not follow statistical laws. 7 Control Charts 1. The meaning of control chart methods: There are many factors that affect product quality, including both static and dynamic factors. Is there a method that can monitor the product manufacturing process in real time, identify potential quality issues promptly, and thus improve the production process and reduce the amount of defective products? Control chart methods are such quality control approaches that focus on prevention; they utilize quality characteristic values collected on-site to create control charts, and by analyzing these charts, it is possible to assess the quality of the product manufacturing process. Control charts can provide a lot of useful information and are one of the important methods in quality management. A control chart, also known as a management chart, is a quality management chart that includes control limits. One of the purposes of using control charts is to analyze and determine whether any abnormalities have occurred in the production process by observing the distribution of product quality characteristics on the control chart. Once an abnormality is detected, necessary actions must be taken promptly to eliminate it and restore the production process to a stable state. Control charts can also be used to bring the production process 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 from probability theory. A control chart is a graphical representation of the quality of a production process; it includes a center line and upper and lower control limits, as well as data points showing the statistical values of samples taken over time. The center line is the average value of the statistic being controlled, with the upper and lower control limits lying a multiple of standard deviations away from the center line. Most manufacturing applications use three-standard-deviation control limits; other control limits can also be used if there is sufficient evidence. Common control charts are divided into two main categories: measurement-type and attribute-type, which are suitable for different production processes respectively ； Each category can be further divided into specific control charts; for example, measurement control charts can be divided into mean-range control charts, single-value moving range control charts, and so on. 2. Drawing of control charts The basic format of a control chart is shown in the figure. To create a control chart, the following steps are generally required: ① Select samples at specified sampling intervals and sample sizes ； ②Measure the quality characteristic values of the sample and calculate their statistical values ； ③Plot points on the control chart ； ④Determine whether there is parallelism in the production process. When control charts provide managers with a great deal of useful information on the production process, the following points should be taken into consideration: ① Select management points appropriately based on the quality conditions of the process steps. Management points generally refer to key locations, critical dimensions, aspects of the manufacturing process that have special requirements, and key factors that affect subsequent operations; for example, areas where quality is unstable or where a high number of defective products are produced can be identified as management points ； ②Based on the quality issues at the management points, select the appropriate type of control chart: ③ When using control charts for process management, it is necessary to first determine reasonable control limits. ④ If the points on the control chart show abnormal conditions, the cause must be identified immediately, and measures taken before production can resume; this is the primary prerequisite for control charts to be effective ； ⑤The control line is not the same as the tolerance line; the tolerance line is used to determine whether a product is qualified, while the control line is used to determine whether there have been changes in the quality of a manufacturing process ； ⑥When an abnormality occurs in the control chart, it is necessary to identify the responsible party and address the issue promptly or report it. When creating control charts, the control limits are not calculated every time; so how are these limits determined initially? If the current production conditions are similar to those in the past, it is possible to rely on previous empirical data, that is, to use the control limits that were effective during periods of stable production. The following describes a method for determining control limits, namely the field sampling method. The steps are as follows: ① Randomly select more than 50 samples, measure the data for these samples, calculate the control limits, and create a control chart ； ②Check whether the control chart is in a controlled state, that is, stable; if all points are within the control limits and there are no abnormal patterns in the arrangement of the points, then it is possible to proceed to the next step ； ③If there is an abnormal condition, or although it is within the control limits but the arrangement is abnormal, it is necessary to identify the cause of the abnormality and take appropriate measures to bring it back under control. Only then should new data be collected to calculate the control limits, allowing progress to the next step ； ④Create a cubic plot using the data obtained above, and compare this plot with the standard limits (upper and lower tolerance limits) to determine whether the conditions are ideal or fairly ideal. If the requirements are not met, measures must be taken to reduce the mean deviation or standard deviation. After taking these measures, repeat the aforementioned steps to collect new data and establish control limits until the standards are satisfied. 3. How to use control charts to identify abnormal phenomena: Control charts are used to determine the status of a production process, with analysis and judgment being based on the positions of the sample points formed from sample data as well as their trends over time. The out-of-control state is mainly manifested in the following two situations: ① The sample points exceed the control limits ； ②The sample points are within the control limits, but their arrangement is abnormal. When data points exceed the management limits, it is generally considered that there is an anomaly in the production process; at this point, the cause must be investigated and corrective actions taken. Abnormal arrangements mainly refer to the following situations: ③ When seven or more consecutive points are all above or below the center line, it is necessary to check whether there have been any changes in the production conditions. ④When two of the three consecutive points enter the area near the control limits (that is, the area starting from the center line and extending more than two-thirds of the way to the control limits), it is necessary to check whether the variability in production is excessive. ⑤The successive appearance of upward or downward trends indicates that the process characteristics are changing upward or downward. ⑥The arrangement of the points changes periodically; at this time, the processing time can be handled hierarchically, and new control charts can be created to identify the cause of the problem. The ability of control charts to detect anomalies varies depending on the number of data points in each group when the data is grouped, the method used for collecting samples, and the way in which layers are defined. One should not be satisfied with merely using one type of control chart; instead, various data collection and utilization methods should be employed to create different types of charts, in order to achieve better results. It is worth noting that if abnormal phenomena beyond the limits of management are detected but no efforts are made to investigate their causes or take corrective actions, then despite the control chart’s usefulness, it is nothing more than a piece of paper.