Post on 14-Nov-2014
description
QA-2 NKS 1
Quality Assurance - Process Control-tools & techniques
Srinivasan Nenmeli-K
QA-2 NKS 2
Q A through Process Control
◆ Process control is the key for QA◆ Identify the key processes & their parameters◆ Set limits to the parameters-higher and lower;
sometimes only one of these.◆ Suppose you make cakes---take baking step as a
key process◆ Set lower and higher limits for baking
temperature--say 50 degrees and 60 degrees centigrade--the "tolerance"
◆ What are the other key processes in baking?-list now!
QA-2 NKS 3
Tolerance -What it means?
◆ Based on experience or testing, you fix the tolerance.◆ As a rule,.closer the tolerance, costlier and difficult is
the process◆ Suppose you make shafts for automobiles rear axle
Suppose you set the length of shaft as: 120 mm with tolerance + or - 3mm.This may be suitable for a farm tractor.
◆ For passenger cars ,you may set 120 mm with tolerance + or - 1mm
◆ For smooth-riding,expensive car [Rolls-Royce ,say] you may set 120 mm with tolerance + or - 0.5 mm.
QA-2 NKS 4
Quality Control--Old Style
◆Quality control and inspectors go together!◆What did old style inspectors do for the
shafts mentioned?◆Suppose the length is 120 mm ,tolerance +
or - 1mm◆Upper specification limit [USL]is 121
mmLower specification limit [LSL] is 119 mm------Tolerance is: USL-LSL=2mm--
◆Inspectors checked a shaft with two gages of 121 and 119 mm-- 'gages' called 'Go-NoG0' gages
◆Acceepted the shafts between the limits
QA-2 NKS 5
QC- Old style
◆Inspectors passed if the length was within specs: between 121 mm and 119 mm.
◆ Others were rejected.
◆Do you want most of the shafts close to 120 mm?
◆Do you check for variation of length between 121 mm and 119 mm?
◆How many are close to the mean,120 mm?
QA-2 NKS 6
Variability of parameters◆ Statistics is a science of variations.◆ If all of us are of same height at certain age,there
is no need for statistical analysis.◆ How do we approach the variations?◆ Suppose your auto-plant makes 1000 shafts a
day.You take a sample of 100 shafts and measure their lengths.
◆ Most probably, you will find the average or mean value of the lengths.Suppose the mean is 120 mm .Are you satisfied?
◆ The old-style inspectors found the number falling within 121mm and `119 mm--let us say 68 shafts were within the limits and accepted.
QA-2 NKS 7
Mean and Sigma◆ The old style inspector would reject the remaining 32
shafts--outside the limits--32% rejection!◆ Suppose you find the "standard deviation" or sigma
from the 100 measurements of length. {Find the formula for sigma--most calculators and computer programs give this}
◆ sigma squared = sigma x sigma = variance ---> remember this
◆ Now our Engineer Tom finds that the sigma value turns out to be 0.98 mm, take it as 1mm--> sigma = 1mm
◆ Aha! you have got the clue now: ◆ Our tolerance: USL -LSL = 121-119 mm is close to
two sigma.!
QA-2 NKS 8
Tolerance and Sigma◆ Tom found that tolerance was mean + /- sigma.◆ Tom talked to his Stats Professor, Prof. Variant,
and he told that if you take those pieces with this tolerance ,you will get only 68.2% acceptable within tolerance.The rejection will be 32% [or nearly 2/3rd fall within tolerance , 1/3rd outside the limits---easy to remember]
◆ Prof Variant told that the lengths of the shafts followed "Normal Distribution" [the bell -shaped curve]with mean as 120 mm and sigma or Standard Deviation as 1mm.
◆ How to reduce rejections with the same tolerance?No use looking at the mean.The sigma should be reduced!
QA-2 NKS 9
Process Control-sigma◆ QA means "process control".◆ Process control means "reduce sigma to the desired level"◆ This is the great lesson the Japanese learnt from
Deming,Juran and others...control variability--reduce sigma.
◆ Tom improved the process of making shafts . Now Tom finds sigma = 0.5 mm for 100 shafts .
◆ Then tolerance limits are: 121 = 120 + 2x0.5 mm and 119 =120 - 2x0.5 mm.Within 2sigma above and below the mean.
◆ Prof Variant told now: Normal distibution tells that nearly 95 % of shafts will pass .Tom is sure that only 5% of shafts will be rejected.Great improvement! but not sufficient.
QA-2 NKS 10
Sigma control
◆ Reducing sigma or control of variance is the key to QA◆ From 1930's and particularly since World War II [say
1945] in most industries,many learnt this lesson..The Japanese applied to almost all manufacture-toys,electronics,optics,steel and the automobiles.Some did not learn and lost to competition their business.
◆ A standard rule was developed: Tolerance/sigma =6
◆ 'Maintain this -dummies '---told the Quality gurus.◆ Performance Index = tolerance/ 6 (sigma)◆ PI should be greater that 1 [can be 1.2 or more]◆ Suppose tolerance = 2mm sigma should be 2/6=0.33mm
or less!
QA-2 NKS 11
Taguchi and Quality Cost
◆ Taguchi was a great quality guru.◆ No use inspecting using tolerance only!;◆ Any deviation from the mean entails cost-cost to
customers & cost to society◆ To make it quantitative,Taguchi gave a simple
cost equation:◆ If the mean is xo and the actual value is x, the
cost is : Cost = k( x - xo)^2 where k is a constant.
◆ The cost increases as the square of the deviation from mean --- a parabolic curve or quadratic equation
QA-2 NKS 12
How to reduce sigma ?
➢QA involves reducing sigma for critical parameters
➢Reduction of sigma applies to both manufacturing and service industry--examples follow
➢Suppose you run a transportation company.The delivery time from New York to Chicago may vary with mean 32 hours ,with sigma +/-2 hours.Your effort should be to reduce the sigma from 2 to 1 hour.
➢ In process control, first you reduce sigma ,then you may shift the mean to lower or higher value---first you control the process ,that is, reduce variations.
➢You can now reduce the mean for shipping from NY to Chicago to 31 hours from 32 hours.
QA-2 NKS 13
Find sources for variation
◆ What are the sources of variation? What factors increase sigma?
◆ In manufacturing: sigma squared = sum of sigma squareds for material, machine,labor,tools and instruments.
◆ Find which sigma-squareds are high.If you are using poor quality materials and worn-out tools,their sigmas will be large.
◆ This is called variance analysisor Anova to separate out the causes of variation or sigma.
QA-2 NKS 14
Simple QA tools
◆ Use check-lists as often as you can--for simplest of tasks...checklist is a powerful QA tool.!
◆ Develop internal company-wide standards--they should be written down,discussed and reviewed periodically
◆ 'Auditing'--What is this? It is "checking against specifications" --Encourage auditing and review audit reports;train internal auditors;
◆ Do "quality audit" at least once a quarter[3 months]
◆ Use the magnifient "Seven Q C tools"
QA-2 NKS 15
Process Flow Chart
◆ Draw a process flow chart--Can you simplify the chart?
◆ Making cake: measure ingredients-->mix-->make dough-->put in trays-->bake in oven
◆ For each process step,identify the critical parameters
◆ Can you avoid some steps?Can you reduce waiting times between steps?
◆ Involve your staff in making and analyzing process flowcharts
◆ Can you combine steps?--example: your receptionist combines reception and phone-switch-board
QA-2 NKS 16
Measurements and process control
◆ " What you can measure,You can control"◆ Measure critical parameters of processes--not
just productivity◆ Measurements relating to:1
materials,2machines,3methods & tool,4men & women
◆ 5 Mother Nature--environment [humidity for instance in some industries]
◆ Metrics is the word used for measurements and their analysis
QA-2 NKS 17
Process Ownership
◆Give process ownership to your staff/managers
◆Do they follow QC/QA concepts?
◆Are they willing to follow Quality Audits?
◆Do they write standards for their processes?
◆Do they perform defect measurement and analysis?
◆Do they measure critical parameters?
QA-2 NKS 18
Q A for service industry
◆ Several service industries apply QA --health care,information technology,hospitality,telecom and education
◆ Measure mean and sigma for service parameters and reduce the sigma value.The most important service parameters are often 'time-related factors'.We saw the example of transportation earlier.
◆ Response time to fix customer complaints/problems is an important parameter to control---find its mean & sigma.Can you reduce this?
◆ Learn about 'lean processes', "Toyota system"--reduce cycle time & waiting times in various process steps, including 'decision-making times','authorization times'
QA-2 NKS 19
Practical means to reduce sigma
● See my other presentation : "QA- 5 tips"---Learn about the Deming cycle or Plan-Do-Check-Act cycle
● Training of employees at all levels is the potent method to reduce sigma.Without sufficient training ,they cannot control processes.
● Develop the culture of 'Continuous Improvement'--Quality is not a quick-fix program.
● Control the sigma of incoming raw material or products--this is very essential for product industries---Control the process sigma of equipment and tools used.
● In service industry, improve documentation and logging of records.The documents serve as 'raw materials' to control the sigma.
● Above all, apply the 'magnificent seven' QC tool
QA-2 NKS 20
Summary
◆ See the connection between quality and tolerance and sigma values of process parameters
◆ Practice "Continuous Improvement"◆ Identify key processes and their critical
parameters◆ Use PDCA cycle and the seven QC tools to
reduce sigma of these parameters◆ Emphasize training for all employees
QA-2 NKS 21
Suggested Reading
□ Dale Besterfield et al --- Quality control
□ Dale Besterfield et al --- Total Quality Management
□ Peter Pande -- The Six Sigma way
QA-2 NKS 22
Contact
Contact me for further presentations and info:
email ID: nksrinivasan@hotmail.com