Statistical process control

from Wikipedia, the free encyclopedia

The statistical process control (including statistical process control or statistical process control , English statistical process control , SPC is called) is usually understood as a procedure for the optimization of production and service processes the basis of statistical methods.

history

SPC was developed by Walter A. Shewhart . The scientific principles were comprehensively derived and described by him in 1931 in the book Economic Control of Quality of Manufactured Product . This work was triggered by the intention of the management of the Hawthorne Plant of the Western Electric Company in Chicago to manufacture products that were as uniform and therefore reliable as possible. Attempting to do this using common sense has failed. Shewhart was subsequently approached by Bell Telephone Laboratories in New York for assistance.

Shewhart started from the assumption that the quality of the end product essentially depends on the combination of the variation in the parameters of the individual parts. He found two fundamentally different mechanisms to be the cause of this spread:

  1. Scatter due to general causes (random deviations from the mean value resulting from a stochastic process, noise (physics) ) and
  2. Variation due to special causes (material defects, machine defects, construction defects, etc.)

Shewhart's second important finding was that in trying to minimize this variance, two mistakes can be made:

  • Error 1: Assigning a deviation to a special cause even though it was caused by a general cause.
  • Error 2: Assigning a deviation to a general cause even though it was caused by a particular cause.

Either one or the other error can be avoided completely, but never both at the same time. So a way had to be found to minimize the cost of error prevention. Extensive statistical studies and theory led Shewhart eventually to the development of control charts (dt. Quality control charts ) to implement as the optimal tool to the findings into daily practice.

SPC found its first large-scale industrial application in World War II , where it was used in the manufacture of armaments.

William Edwards Deming later realized that these insights and tools could be applied to all types of processes (business processes, administrative processes, etc.) with the same positive results. This teaching fell on fertile soil, especially in Japan, where it was further developed within the Toyota production system, among other things .

Today, statistical process control is seen as part of a quality management system and accompanies the core process of production or service as a service process . All statistical methods used to monitor and optimize the core process are summarized under the term statistical process control . These methods go beyond the various control card techniques and also include e.g. B. the methods of statistical test planning, the FMEA or the Six Sigma method collection . SPC values ​​are incorporated into customer-supplier relationships as process capability indices.

Action

After the process to be examined has been clearly defined, a process expert must determine which measured variables are important. These must then be recorded as planned during ongoing production. The evaluation is then carried out using quality control cards (e.g. -, - or - card).

There are now software packages that try to establish statistical process control as a central component of the computer-aided quality assurance CAQ . The measurements taken for this purpose are partly automatically by a M achines d ata s rfassung (MDE) carried out and further processed statistically accordingly.

Use

SPC is used to maintain a predefined level of quality as cost-effectively as possible . On the other hand, it is unsuitable for increasing the quality of products . A quality level that goes beyond the required level would result in additional costs to which only an insignificant additional benefit would be assigned. Typically, the required amount of defect-free parts is at a value of X, which can be achieved with relatively little control effort and thus low costs. An increase in quality to 100% defect-free parts would, however, greatly increase the control effort, namely by significantly more than the difference of 100% -X, since the total costs increase exponentially with the desired quality level (including more tests, better test equipment and Production machines, more suitable production methods, etc.). It is referred to as uneconomical "over- provisioning ". Thus SPC serves the economic minimum principle (Engl. Minimal principle ).

Other quality management methods, such as FMEA, are necessary to increase the desired quality level in a targeted manner .

software

SPC is usually used with software support. Three types of software are used here. Firstly, generic calculation software such as spreadsheets or statistics packages, secondly, comprehensive CAQ total solutions such as those from the following manufacturers (in alphabetical order): AHP, Babtec, Böhme & Weihs Systemtechnik GmbH & Co.KG, camLine GmbH, CAT GmbH, CAQ AG, Fink & Partner GmbH, Gewatec GRIPS, Guardus, IBS, iqs Software GmbH, Minitab, Pickert & Partner, Predisys, QDA Solutions, Q-DAS, QUIPSY, SAP, SCIIL AG, Solvtec GmbH or Syncos. Beginners mostly use spreadsheets. Special SPC and CAQ products can, however, reduce the workload and enable central evaluations: PLC software is usually preconfigured with manufacturer-specific rules and can read measurement data directly from measuring tools and machines, while a CAQ system has other quality management modules such as FMEA, Contains supplier evaluation or process control plans over several levels and articles.

literature

  • Alfred Schulze: Statistical Process Control (SPC). In: Tilo Pfeifer, Robert Schmitt (Hrsg.): Masing Handbook Quality Management , 6th revised edition. Carl Hanser Fachbuchverlag, Munich / Vienna 2014, ISBN 978-3-446-43431-8 , chapter 30.
  • Edgar Dietrich, Alfred Schulze: Statistical procedures for machine and process qualification . 6th, completely revised edition. Carl Hanser Verlag, Munich / Vienna 2009, ISBN 978-3-446-41525-6 .
  • Josef Heinhold, Karl-Walter Gaede: Engineer statistics . 4. verb. u. essential exp. Edition. Oldenbourg-Verlag, Munich, Vienna 1986, ISBN 3-486-31744-X .
  • Donald J. Wheeler, David S. Chambers: Understanding statistical process control . 2nd Edition. SPC Press, Knoxville, Tenn. 1992, ISBN 0-945320-13-2 .
  • Horst Rinne, Hans-Joachim Mittag: Statistical methods of quality assurance . 3. revised Edition. Carl Hanser Verlag, Munich 1995, ISBN 3-446-18006-0 .
  • Walter Andrew Shewhart: Economic control of quality of manufactured product . ASQ Quality Press, Milwaukee 1980, ISBN 0-87389-076-0 .
  • Walter Andrew Shewhart: Statistical Method from the Viewpoint of Quality Control . Ed .: William Edwards Deming. Dover Publications Inc., New York 1986, ISBN 0-486-65232-7 .
  • William Edwards Deming : Out of the Crisis . 2nd Edition. MIT Press , Cambridge 2000, ISBN 978-0-262-54115-2 .
  • Günter Faes: SPC - Statistical Process Control: A practical introduction to statistical process control and its use . 2nd Edition. Books on Demand , Norderstedt 2009, ISBN 978-3-8370-5156-8 .
  • Horst Quentin: Statistical process control: SPC . In: Pocket Power . tape 55 . Carl Hanser Verlag, Munich 2008, ISBN 978-3-446-41637-6 .

Web links