Statistical Quality Control in Manufacturing

Statistical quality control methods utilize data-driven measurements to detect and eliminate defects or variations in the manufacturing process that lower the quality of products. This article summarizes the origins of quality control and statistical quality control; examines Six Sigma, a widely-recognized methodology for statistical quality control; notes the situations where the use of Six Sigma is undesirable; and provides a glossary of relevant terms.

Keywords Control Chart; Deviation; DMAIC; Lean Manufacturing; Quality; Quality Control; Six Sigma; Standard Deviation; Statistical Process Control; Statistical Quality Control; Tolerance

Overview

What is Quality Control?

Quality control refers to the oversight of the manufacturing process to ensure that products are manufactured as error-free as possible.

Origins of Quality Control

Quality control methods have been used in manufacturing for centuries. Formal quality control measures have existed at least as far back as the Middle Ages, when the craft guilds required apprentices to undergo long and rigorous training and also demonstrate proficiencies in their crafts before they would be considered master craftsmen (National Institute of Standards and Technology, 2006).

What is Statistical Quality Control?

Statistical quality control methods utilize data-driven measurements to detect and eliminate defects or variations in the manufacturing process that lower the quality of products.

(The term "statistical process control" is sometimes used interchangeably with the term "statistical quality control." However, according to the American Society for Quality, the two terms do differ somewhat; see the "Glossary" below.)

Origins of Statistical Quality Control

Statistical quality control methods were introduced in 1924 by Walter A. Shewhart, an engineer at Western Electric and parent company, Bell Telephone Laboratories. In 1924, Shewhart sent a memo that included a diagram of a modern control chart. (A control chart is a diagram that indicates "upper and lower control limits on which values of some statistical measure for a series of samples or subgroups are plotted. The chart frequently shows a central line to help detect a trend of plotted values toward either control limit" (American Society for Quality, 2005, Glossary).

Shewhart and his colleagues at Bell Laboratories continued to refine the theory and application of statistical quality control and in 1931, Van Nostrand published Shewhart's book, Economic control of quality of manufactured product (National Institute of Standards and Technology, 2006).

Six Sigma: A Contemporary Statistical Quality Control Methodology

Currently, Six Sigma is one of the most widely-recognized methodologies of statistical quality control. Six Sigma methodology "values defect prevention over defect detection. It drives customer satisfaction and bottom-line results by reducing variation and waste" (American Society for Quality,2995, Six Sigma Overview). Motorola is credited with creating Six Sigma in 1987. Soon afterwards, the concept was adapted by other leading manufacturers including Texas Instruments, IBM, General Electric, and Whirlpool (Dossenbach, 2004, p. 25).

Applications

This section examines Six Sigma methodology as it applies to statistical quality control in the manufacturing industry.

Six Sigma: Goal of the Six Sigma Methodology

The goal of Six Sigma methodology is to consistently manufacture products that have no defects by utilizing statistical quality control tools. According to the American Society for Quality (2005), quality manufacturing performance will yield no more than 3.4 defects per million opportunities.

Six Sigma: DMAIC: A Five Step Process

Six Sigma relies upon a five-step process:

• Define

• Measure

• Analyze

• Improve

• Control

The process is commonly known by the abbreviation "DMAIC" which denotes the first word in each step. Calabrese (2007, p. 31) further expands upon the five Six Sigma, DMAIC steps as follows:

• "Define: Identify the variable to be improved."
• "Measure: Capture data on the identified data."
• "Analyze: Brainstorm the root cause variables and their relationship with the variable that is to be improved."
• "Improve: Remove root causes and/or minimize variations around the mean of the target variable."
• "Control: Sustain the improvements in the process via control chart applications."

Six Sigma: DMAIC: The Significance & Use of Control Charts

Control charts (see step 5 in "Six Sigma: DMAIC: Five Steps to Perfection" above) are the most significant feature of the Six Sigma methodology. The control chart is the tool that is used to plot and graph a process over time in order to detect variations that deviate from the allowable standard deviation. The allowable standard deviation is also known as "tolerance."

The basic control chart procedure involves collecting and charting data for a specific time period and analyzing the data for out-of-control signals (deviations). Out-of-control signals may illustrate any of the following deviations (Tague, 2004, p. 155-158 (as cited by the American Society for Quality, 2005)):

• A single point outside the control limits;
• Points that deviate from the control limits;
• Data that indicates unusual data or process patterns.

Six Sigma: Training

The Six Sigma methodology has become so popular that training and certification in the five levels of expertise are offered by many organizations, including the American Society for Quality.

Five Levels of Expertise

There are five levels of expertise in Six Sigma; each is labeled with a martial arts term.

• White Belt
• Yellow Belt
• Green Belt
• Black Belt
• Master Black Belt

The following descriptions of the five levels are from the American Society for Quality (2005) and are arranged from lowest to highest level of knowledge and responsibility:

• "White Belt: Can work on local problem-solving teams that support overall projects but may not be part of a Six Sigma project team."
• "Yellow Belt: Participates as a project team member; reviews process improvements that support the project."
• "Green Belt: Assists with data collection and analysis for Black Belt projects; leads Green Belt projects or teams."
• "Black Belt: Leads problem-solving projects; trains and coaches project teams."
• "Master Black Belt: Trains and coaches Black Belts and Green Belts; functions more at the Six Sigma program level by developing key metrics and the strategic direction; acts as an organization's Six Sigma technologist and internal consultant."

Six Sigma: Case Study: Boeing's Satellite Development Center (SDC)

The Company

Boeing's Satellite Development Center (SDC) is the world's largest manufacturer of satellites. The satellites are manufactured for military, weather, space, and communications purposes.

The Situation

Precision in the manufacturing of the satellites and their parts is absolutely essential for the safety and accuracy of the finished satellites.

The Problem

One significant issue for SDC is to ensure that bolts and rivets are accurately torqued. In the past, this involved a lengthy, expensive verification process to observe and record the history of each torque and the calibration data: A quality control person would monitor the entire process and record the data, and since the accuracy of the torquing tools was paramount, each of the 3,500 torque tools had to be sent to an outside vendor for calibration.

The Solution

Vu D. Pham, leader of engineering assurance at SDC Pham's set out to improve the quality and accuracy of the torque tools. Pham, who holds a Six Sigma Black Belt designation, employed the Six Sigma methodology to analyze the people, systems, processes, methods, and torque tools. As a result of his findings, Pham commissioned Mountz to develop a software verification system that was more accurate and portable — it was mounted on each tool cart in a hand-held PDA-based system. The portability factor of the PDA allowed for torque testing and data recording at the point of use, which increased the accuracy of the information.

The Benefits ("Tightening up torque standards," 2007):

• SDC approached Six Sigma tolerance with the new PDA-based torque verification system.
• The new system is so convenient and reliable, that SDC no longer requires a quality control person to monitor and record each torque process because the individual operators now perform those functions at the point of use.
• The new system has eliminated the need to send torque tools to an outside vendor for calibration.
• SDC estimates that it reaped $200,000 in raw savings the first year, just by eliminating the need to send out the torque tools to an outside vendor for calibration.
• SDC continues to save money on the cost of failures, repairs, and personnel. (No estimate available.)

Six Sigma: A Useful Tool for the Manufacturing of Pharmaceutical Drugs

Although Six Sigma methodology is often associated with the manufacturing of electronics, aircraft, appliances, and machine parts, it can also be an effective tool in the manufacturing of pharmaceutical drugs. Control charts, in particular, can be created to control a critical attribute of drugs: tablet weight. Tablet weight is significant because its accuracy determines the uniformity, hardness, disintegration time, and breakability of the tablets. Six Sigma can be utilized to control tablet weight by measuring and identifying variables in the amount and ratios of ingredients and by identifying variables in tablet compression equipment. While the variables in the tablet compression equipment are usually not possible or are too expensive to modify, Six Sigma allows for detection of the variables. The identification and measurement of these two factors — amount and ratio of tablet ingredients and variables in compression equipment — then offer opportunities for design and process modifications that will yield the most accurate tablet weight with the available compression equipment (Calabrese, Foo, & Ramsay, 2007).

Discourse

If Six Sigma is so beneficial in controlling quality, then why don't all manufacturers adopt it? When does Six Sigma become burdensome or undesirable?

Six Sigma is not the cure for every manufacturing quality issue or process. It works best when uniformity of product and process characteristics is the goal.

Six Sigma may still be effective if it is applied to selected aspects of a manufacturing operation. However, if the following four factors are present, Six Sigma may be impractical for the entire manufacturing process:

• Innovative Intention
• Highly-customized Product
• Lack of Resources
• Ambivalent Management

Factor One: Innovative Intention

The first factor that may deem Six Sigma impractical is an innovative intention. If a company wants to try out new methods or processes, or create an entirely new prototype, it will expect and desire variations in specifications and quality. (see "Notes from 3M" below for an example of the tension between innovation and Six Sigma.)

Factor Two: Highly-customized Products

The second factor that may deem Six Sigma impractical is the manufacturing of highly-customized products. If a company manufactures each item to different customer specifications, an overall Six Sigma approach may not work.

Factor Three: Lack of Resources

The third factor that may deem Six Sigma impractical is a lack of resources. A company needs to devote time, funding, and manpower to train employees in Six Sigma methods.

Factor Four: Ambivalent Top Management

The last factor that may deem Six Sigma impractical is ambivalent top management. If top management is not convinced that Six Sigma is valuable enough, implementation will be spotty. Six Sigma depends upon the commitment of top managers for support and effective implementation.

Notes from the Woodworking Industry

Dossenbach (2004) stresses that Six Sigma is not necessarily the best quality control methodology for the woodworking industry. He agrees that of course it is desirable to measure performance and eliminate defects, but disagrees that a full-blown Six Sigma approach is necessary for the woodworking industry. Instead, Dossenbach recommends focusing on customer-centered quality control by following the continuous improvement approach of lean manufacturing.

Dossenbach (2004, p. 26) offers the following continuous improvement basics as more appropriate quality control measures for the woodworking industry:

• Keep "customer requirements and expectations in focus at all times."
• "Set goals and objectives."
• "Educate and train suppliers and employees."
• "Continuously measure the performance quality of your company and suppliers."
• "Identify causes of unacceptable performance and take corrective action to prevent recurrence."
• "Take a proactive or preventive approach to managing quality."
• "Get total involvement and cooperation from suppliers and employees."
• Begin a never-ending journey to perfection.

Dossenbach's basics cited above certainly sound comprehensive and effective. What distinguishes these quality control measures from Six Sigma methodology? Dossenbach's basics are just that — basics — they provide a general outline of quality control measures. They don't provide a precise blueprint for actually controlling quality. On the other hand, Six Sigma relies upon statistical quality control measures, most notably control charts to identify and measure variables and defects.

Notes from 3M

3M is the company famous for inventing masking tape, Thinsulate, and most notably, Post-it notes. In 2000, When James McNerney left General Electric (GE) to become CEO of 3M, he quickly implemented cost-cutting strategies and GE's quality control methodology: Six Sigma. His strategies were successful in reducing defects and increasing efficiencies and rejuvenating the company's stock, perhaps at the expense of creativity. By the time McNerney left 3M after 4 1/2 years to become chairman of the board, president and chief executive officer of The Boeing Company, 3M's introduction of ground-breaking products had dwindled to a trickle. Now, the new CEO George Buckley is cutting back on Six Sigma and instituting strategies to help 3M regain its creative edge. Buckley is credited with returning an atmosphere of creativity to 3M, chiefly by relaxing the Six Sigma controls (especially in the research labs); increasing the research and development budget; and encouraging risk-taking (Hindo, 2007).

Conclusion

Statistical quality control methods evolved from quality control measures. Six Sigma, one of the most widely-recognized methodologies for statistical quality control has been adopted by many manufacturers. Because Six Sigma aims to detect and eliminate defects, it offers distinct quality and financial benefits for manufacturers. Because of its wide adoption by large companies, many opportunities for training and certification in Six Sigma are available. However, Six Sigma is not the best choice for all companies and processes.

Terms & Concepts

Control Chart: "A chart with upper and lower control limits on which values of some statistical measure for a series of samples or subgroups are plotted. The chart frequently shows a central line to help detect a trend of plotted values toward either control limit" (American Society for Quality, 2005, Glossary).

Deviation: "In numerical data sets, the difference or distance of an individual observation or data value from the center point (often the mean) of the set distribution" (American Society for Quality, 2005, Glossary).

DMAIC: "A data driven quality strategy for improving processes and an integral part of a Six Sigma quality initiative. DMAIC is an acronym for define, measure, analyze, improve and control" (American Society for Quality, 2005, Glossary).

Lean Manufacturing: An ongoing, systematic effort to eliminate the sources of waste in a production process (Mark, 2007).

Quality: "The characteristics of a product or service that bear on its ability to satisfy stated or implied needs; a product or service free of deficiencies" (American Society for Quality, 2005, Glossary).

Quality Control: "The operational techniques and activities used to fulfill requirements for quality" (American Society for Quality, 2005, Glossary).

Six Sigma: "A method that provides organizations with tools to improve the capability of their business processes. This increase in performance and decrease in process variation lead to defect reduction and improvement in profits, employee morale and quality of products or services. Six Sigma quality is a term generally used to indicate that a process is well controlled (±6 s from the centerline in a control chart). Six Sigma quality performance means no more than 3.4 defects per million opportunities" (American Society for Quality, 2005, Glossary).

Standard Deviation: "A computed measure of variability indicating the spread of the data set around the mean" (American Society for Quality, 2005, Glossary).

Statistical Process Control (SPC): "The application of statistical techniques to control a process; often used interchangeably with the term 'statistical quality control'" (American Society for Quality, 2005, Glossary).

Statistical Quality Control (SQC): The application of statistical techniques to control quality. Although SQC is often used interchangeably with the term "statistical process control," SQC includes acceptance sampling and statistical process control does not include acceptance sampling (American Society for Quality, 2005).

Tolerance: "The allowable deviation from a standard, especially the range of variation permitted in maintaining a specified dimension in machining a piece" ( Merriam-Webster's collegiate dictionary, 2000).

Torque, Torqued: A turning or twisting force (Merriam-Webster's collegiate dictionary, 2000).

Bibliography

American Society for Quality (ASQ). (2005). Six Sigma overview. Retrieved September 12, 2007 from ASQ Website. http://www.asq.org/learn-about-quality/six-sigma/overview/overview.html

American Society for Quality (ASQ). (2005). Glossary. Retrieved August 31, 2007 from ASQ Web site. http://www.asq.org/glossary

Bandyopadhyay, J., Jenicke, L.O. (2007). Six Sigma approach to quality assurance in global supply chains: a study of United States automakers. International Journal of Management, 24, 101-107. Retrieved August 27, 2007, from EBSCO Online Database Business Source Complete. http://search.ebscohost.com/login.aspx?direct=true&db=bth&AN=25139351&site=ehost-live

Behbahani, M., Saghaee, A., & Noorossana, R. (2012). A case-based reasoning system development for statistical process control: Case representation and retrieval. Computers & Industrial Engineering, 63, 1107-1117. Retrieved November 15, 2013, from EBSCO Online Database Business Source Complete. http://search.ebscohost.com/login.aspx?direct=true&db=bth&AN=82840027&site=ehost-live

Borawski, P. (2006). The state of quality: 1947 and 2006. Journal for Quality & Participation, 29, 19-24. Retrieved August 29, 2007, from EBSCO Online Database Business Source Complete. http://search.ebscohost.com/login.aspx?direct=true&db=buh&AN=23965963&site=ehost-live

Calabrese, R., Foo, L., & Ramsay, O. (2007). Reducing variance. Drug Discovery & Development, 10, 31-33. Retrieved August 29, 2007, from EBSCO Online Database Business Source Complete. http://search.ebscohost.com/login.aspx?direct=true&db=buh&AN=26073247&site=ehost-live

Dossenbach, T. (2004). Six Sigma: Is it for you? Wood & Wood Products, 109, 25-26. Retrieved August 27, 2007, from EBSCO Online Database Business Source Complete. http://search.ebscohost.com/login.aspx?direct=true&db=bth&AN=15223569&site=ehost-live

Fugee, T., Zhou, Z., & Jiang, W. (2007). Applying manufacturing batch techniques to fraud detection with incomplete customer information. IIEE Transactions, 39, 671-680. Retrieved August 27, 2007, from EBSCO Online Database Business Source Complete. http://search.ebscohost.com/login.aspx?direct=true&db=bth&AN=24471429&site=ehost-live

Hindo, B. (2007). At 3M, a struggle between efficiency and creativity. Business Week, (4038), 8-14. Retrieved September 17, 2007, from EBSCO Online Database Business Source Complete. http://search.ebscohost.com/login.aspx?direct=true&db=bth&AN=25263157&site=ehost-live

Kay, S. (2007). Save money by understanding variance and tolerancing. Medical Device Technology, 18, 40-43. Retrieved August 29, 2007, from EBSCO Online Database Business Source Complete. http://search.ebscohost.com/login.aspx?direct=true&db=buh&AN=25280903&site=ehost-live

Knoth, S., & Steinmetz, S. (2013). EWMA p charts under sampling by variables. International Journal of Production Research, 51, 3795-3807. Retrieved November 15, 2013, from EBSCO Online Database Business Source Complete. http://search.ebscohost.com/login.aspx?direct=true&db=bth&AN=89100637&site=ehost-live

Mark, A. (2007). Organic is best. Modern Machine Shop, 79, 10-10. Retrieved July 12, 2007, from EBSCO Online Database Business Source Complete. http://search.ebscohost.com/login.aspx?direct=true&db=bth&AN=24510244&site=ehost-live

Merriam-Webster's collegiate dictionary (10th ed.). (2000) Springfield, MA: Merriam- Webster.

Oakham, M. (2007). Taking control. Engineer, 293(7716), 35-36. Retrieved July 23, 2007, from EBSCO Online Database Business Source Complete. http://search.ebscohost.com/login.aspx?direct=true&db=bth&AN=24103962&site=ehost-live

Parkinson, J. (2007). Worshipping Six Sigma. CIO Insight, 83, 72-72. Retrieved August 27, 2007, from EBSCO Online Database Business Source Complete. http://search.ebscohost.com/login.aspx?direct=true&db=bth&AN=26047321&site=ehost-live

Tightening up torque standards. (2007). Quality, 46, 16-18. Retrieved September 18, 2007, from EBSCO Online Database Business Source Complete. http://search.ebscohost.com/login.aspx?direct=true&db=bth&AN=26259147&site=ehost-live

United States Department of Commerce National Institute of Standards and Technology (NIST). (2006). How did statistical quality control begin? NIST/SEMATECHe-Handbook of Statistical Methods. Retrieved September 13, 2007, from NIST Website. http://www.itl.nist.gov/div898/handbook/pmc/section1/pmc11.htm

Zantek, P., Shan, L., & Yong, C. (2007). Detecting multiple special causes from multivariate data with applications to fault detection in manufacturing. IIE Transactions, 39, 771-782. Retrieved July 19, 2007, from EBSCO Online Database Business Source Complete. http://search.ebscohost.com/login.aspx?direct=true&db=bth&AN=25228069&site=ehost-live

Suggested Reading

Djurdjanovic, D., & Ni, J. (2007). Online stochastic control of dimensional quality in multistation manufacturing systems. Proceedings of the Institution of Mechanical Engineers — Part B — Engineering Manufacture, 221, 865-880. Retrieved July 23, 2007, from EBSCO Online Database Business Source Complete. http://search.ebscohost.com/login.aspx?direct=true&db=bth&AN=25640429&site=ehost-live

Dwyer, J. (2007). 'P' before 'Q.' Works Management, 60, 36-39. Retrieved August 29, 2007, from EBSCO Online Database Business Source Complete. http://search.ebscohost.com/login.aspx?direct=true&db=bth&AN=24991679&site=ehost-live

Mairani, J. (2007). No matter the plant size, quality management systems measure up. Plant Engineering, 61, 25-26. Retrieved August 29, 2007, from EBSCO Online Database Business Source Complete. http://search.ebscohost.com/login.aspx?direct=true&db=buh&AN=24806760&site=ehost-live

Schofield, J. (2007). When did Six Sigma stop being a statistical measure? Journal of the Quality Assurance Institute, 21, 25-26. Retrieved August 27, 2007, from EBSCO Online Database Business Source Complete. http://search.ebscohost.com/login.aspx?direct=true&db=bth&AN=25246490&site=ehost-live