Defining Quality: Towards a Better Understanding of “Statistical Quality Control”Ece MutluBlockedUnblockFollowFollowingJun 29Quality of products and services plays an important role in decision making processes of different customer segments.
Maintaining quality at the desired level, though may be challenging, is imperative in achieving high level of customer satisfaction, as well as, maximizing revenue and market share, rendering elimination of waste product for companies and prolongation of product loyalty for the customers.
The second law of thermodynamics states that all systems move towards a state of greater entropy (or disorderliness).
This phenomenon is observed also in manufacturing processes: operators may make a mistake, machines may break down, etc.
, when equipments are not controlled or regulated.
Therefore, quality of a process is strongly affected by its stability and repeatability.
Although it is impossible to prevent the inherent tendency towards disorderliness of processes, various techniques may reduce the process variability sufficiently.
This is why quality control and improvement is an important concern to industries.
Reduction in variability of quality characteristics is achieved via online statistical process monitoring (SPM) techniques.
SPM may be defined as the engineering tools and statistical techniques applied to measure and observe the quality characteristics of a process for monitoring and quality improvement.
Process monitoring may be achieved via seven quality tools, which are histogram plots, pareto charts, cause and effect diagrams, defect-concentration diagrams, scatter diagrams, sampling inspections and check sheets, which yield limited discrete knowledge about the process rather than distinguishing the variation over a period (left figure (a)).
Control charts, on the other hand, offer continuous information about the process, and provide an opportunity to apply systematic approaches to decrease the variability in quality characteristics (right figure (b)).
Representation of process variation via a) histogram b) control charts.
The concept of quality control was introduced by Dr.
Walter Andrew Shewhart, a physicist, during his service in Bell Telephone Laboratories in 1920s.
He needed to decrease the frequency of failures and repairs of transmission systems, so he employed control charts.
In the following years, quality control drew the attention of researchers and practitioners.
Deming, who is another widely known proponent of statistical quality control (SQC), brought a new perspective to the application of control charts by emphasizing the responsibilities of the top management.
He developed his complete philosophy of management and popularized the Deming PDCA cycle based on the idea to “plan the necessities’’, “do what is needed’’, “check whether it works or not’’, and “act to prevent abnormalities’’.
Juran defined the quality as “fitness for use’’ to emphasize that achievement of desired quality cannot happen accidentally, it should be planned and controlled.
He advocated a triology for quality management as planning, improvement and control.
Usage of pareto charts to prioritize variables was his suggestion, as well.
Ishikawa’s ideology, on the other hand, focused on the importance of total quality control of an organization, rather than just focusing on products and/or services.
He urged managers to resist becoming content with merely improving a product’s quality, insisting that quality improvement can always go one step further.
He emphasized the importance of “internal customers’’; customer would continue receiving service even after receiving the product.
This service would extend across the company in all levels of management, and even beyond the company to everyday lives of those involved.
Importance of quality control, however, was not widely recognized particularly by the U.
After World War II, due to necessity of improving the manufacturing efficiency, application areas of SQC were highly extended in chemical industries.
Numerious studies demonstrated the positive relationship between quality control and manufacturing efficiency.
In 1980s, an ingenious perception of Taguchi showed the importance of design of experiments in SQC.
In the following years, further research was conducted in statistical techniques more efficiently.
Aim of online application of control charts is not to decrease inherent variability in a process regardless of design and maintenance, but to detect unusual changes.
Shewhart, Deming and Juran view the unusual changes as a result of local faults/special causes and sporadic problems, while they see inherent variation as a consequence of chance causes, system faults/common causes and chronic problems.
Common causes vs.
Special causesProcess is defined as set of causes and conditions, which come together in different combinations to transform inputs (chemical compounds, information, etc.
) into outputs (chemical product, service, behaviour, etc.
If one samples products in equal time intervals, these measurements are subjected to variation which may stem from common and/or special causes.
When variation is only caused by common causes, the process is stable and statistically in-control, since variability of common causes generates a fluctuation within an acceptable and predictable bandwidth.
Common causes do not have a prominent effect on equilibrium point, because the underlying probability distributions of mean and standard deviation estimates of a process result remain stable over time.
Special causes, on the other hand, are the causes that stem from assignable sources and indicators of problems, which need to be solved.
Special causes may be due to extraordinary conditions, incidents or lack of attention, or a new supplier for raw material, etc.
A process, in which outputs are affected by both common and special causes, is called an unstable, or statistically out-of-control process.
The magnitude of the variation and change in the quality characteristics are unpredictable.
Consequently, special causes should be identified and removed for maintaining a stable process.
Deming illustrated the necessity of a stable process as follows.
A stable process has an identity, and its performance is foreseeable; therefore, it allows rational planning.
Maximum productivity and minimum costs are obtained in the existence of stability in a process.
Effect of changes for a stable process can be identified in a more reliable way.
In an unstable process, it is difficult to distinguish special causes from common causes.
If the inherent variability of a process is not identified separately, it would be more challenging to know when a change results in improvement.
As importance of quality management has been accepted over years, application areas of quality control charts have been widened in manufacturing processes.
Regardless of the characteristics of the process, customer need must be recognized to dene and manage quality.
While subjectivity and polysemy of quality makes the concept dicult to comprehend, quality should be attained and kept at a satisfactory level as much as possible, regardless of its definition.