/k Introduction to SPC 2WS02 Industrial Statistics A. Di Bucchianico

/k What is Statistical Process Control (SPC )? The set of operational and managerial activities that a company uses to ensure that the quality characteristics of a product are at the required level?? A series of control charts??

/k Statistical methods acceptance sampling (AS) statistical process control (SPC) design of experiments (DOE)

/k Dimensions of Quality Performance Reliability Durability Serviceability Aesthetics Features Perceived Quality Conformance to standards

/k Definitions of Quality Fitness for use –design –conformance Inversely proportional to variability

/k Why study history of quality control? understand and foresee trends and future developments companies may still be working with concepts from the past

/k History of Quality control craftsmen: quality is what producer and consumer agree about start of mass-production –parts should be exchangeable –all the same –no variability

/k History: in the beginning Quality really became an issue during the Industrial revolution (19th century). Inspection-oriented quality was adopted in the emerging factories (usually through inspection department) F.W. Taylor started research on quality and management in the 1890’s (for more information, read the introduction to the Taylor collection at the Stevens Institute of Technology)introduction

/k output input processes specifications scrap customer History part I: inspection inspection just before shipping of product related definition of quality: conformance to specifications

/k Problems with total inspection mistakes have already been made –“one cannot inspect quality into a product” does not yield insight in production process very expensive inspection is not water-tight

/k How many times does the letter f occur ? Finished files are the result of years of scientific study combined with the experience of many years.

/k How many times does the letter f occur ? Finished files are the result of years of scientific study combined with the experience of many years.

/k Example of acceptance sampling simplest form: –lot size N –sample size n –acceptance number c (lot is accepted if number of defectives in sample is  c) OC (Operating characteristic)-curve gives P(acceptance of lot | percentage defectives)

/k input processes output scrap customer specifications sampling plans History part II: Acceptance sampling

/k History: W.A. ShewhartW.A. Shewhart Shewhart developed at Bell Labs in the 1920’s a process oriented control approach. Instead of inspection, production process is monitored Sampling plans were used to achieve acceptable quality levels (AQL) Techniques employed in quality control were statistical, basically sampling methods and control charts

/k History part III: Monitoring process output outputinput OCAP control charts limits single process

/k SPC: Statistical process control new important tool: control chart measurements of production process during production prevention instead of detection afterwards monitoring variance behaviour of production corresponding definition of quality: –variation of process fits within tolerances

/k History: World War II Increase of production during World War II caused a large expansion of quality control activities in both the U.S.A. and Britain. Statistical techniques were so effective that they were classified as military secrets during World War II! After World War II, the U.S.A. only had major production facility left. Lack of competition and enormous post-war demand caused disinterest in quality control.

/k History: Japan Before World War II, Japanese products were cheap but of low quality. After World War II, Japan had to rebuild its factories (including the management structure). American consultants (in particular Deming and Juran) were brought in to teach quality.Deming Juran A national centre for quality control was established (JUSE) and many national activities were organized.JUSE Japanese companies have extensive training programme for “everyone” in the company.

/k History: W.E. DemingW.E. Deming Deming applied Shewhart’s principles of quality control at various places From 1950 on, he often visited Japan as lecturer and consultant (the Japanese honoured him by naming the highest Japanese quality award after him) In spite of this, his fame in the US did not really start until Deming puts much emphasis on management rather than production floor.

/k History: J.M. JuranJ.M. Juran Early Japanese efforts in quality control overemphasised statistical methods under the influence of Deming. Juran’s 1954 visit to Japan initiated a change of focus on technology to a focus on management (especially involvement of top- management). Juran also emphasised the importance of education and the cost of quality.

/k History part IV: Problem solving tools 7 problem solving tools (‘Magnificent Seven’) fishbone/Ishikawa diagram Pareto chart run chart defect concentration diagram Box-and Whisker plot...

/k History: K. IshikawaK. Ishikawa Ishikawa was very important to the quality revolution in Japan because of his many activities (consulting, member of various committees,...). Ishikawa stressed that quality control must try to meet the requirements of customers. Technical contribution: the cause-and-effect (or fishbone) diagram.

/k Process thinking all deviations from product specifications have causes in production process outcome production process is caused by many process factors Methods/ adjustments Peopl e Environme nt Tools Machin e Materials deviation

/k Design of Experiments preferably design phase (cf. EVOP) off-line team detect causes of variation optimise process

/k History: G.E.P. BoxG.E.P. Box Introduced DOE in industrial environment since early 1950’s Developed response surface methods at Imperial Chemical Industries Moved to US in 1960 and initiated Center for Quality and Productivity Improvement in Wisconsin ASQC history: George Box award:

/k History: TaguchiTaguchi Systematic experimentation is necessary for product development that leads to high quality products Taguchi developed a system of statistical experimentation that meet the needs of the engineer working in industry and is easier to understand than the classical statistical methods Taguchi introduced the concept of “loss function” (i.e., how “ much ” do I conform to specifications?) and “robust design”.

/k Conformance to specifications why is product just outside specifications worse than product just inside specifications? classical: yes/no LSLUSL LSLUSL modern: Taguchi loss function

/k Conformance to specifications II specifications of characteristic 1 specifications of characteristic 2

/k History: US rediscovers quality control In the 1970’s and 1980’s Japan established itself as world leader in quality and became an economic superpower Several American industries could not compete with Japanese industries, because of their neglect of quality control Quality control starts to prosper again in the 1980’s in the US under the guidance of Deming, Juran and new names like Feigenbaum and Crosby. Examples of successful American companies include Ford (QS9000) and Motorola (Six Sigma).QS9000Six Sigma ASQASQ: American Society for Quality

/k History: Europe Interest in quality control in Europe did not start until the 1990’s. Much emphasis in Europe on ISO 9000 certification.ISO 9000 European quality control initiatives like EFQM focus mainly on management issues.EFQM In 2000, ENBIS (European Network for Business and Industrial Statistics) was founded. Initiative by Bisgaard, then successor of Box in Wisconsin.ENBIS

/k History part V: Monitoring process input Control of process factors Prevention of disturbances Tools: SPC, TPM, Poka Yoke,... corresponding definition of quality: –variation of process fits within tolerances PROCES S output specifications input

/k History: new developments An important new development is the Six Sigma approach developed by Motorola (one of the first winner of the Baldrige Award)Baldrige Award Six Sigma approach uses known statistical and management concepts to come to an overall approach to quality: –use statistical tools (DOE, regression analysis) to understand relations between input and output parameters of processes –use control charts on inputs instead of outputs –direct link of efforts with reduction of costs

/k History part VI: Quality by design Actions in design phase: –prevent quality problems in production phase –prevent quality problems when customer uses product Typical tools: Design of Experiments, Robust Design Corresponding definition of quality: –quality is conformance to customer requirements and expectations

/k The main line of history to be remembered is total inspection acceptance sampling control charts on output control charts on input technical solutions/ TPM /... quality by design Summary of historical development inspection prevention