Spc training[1]
Download as PPT, PDF0 likes157 views
The document discusses statistical process control and data collection. It covers measuring performance, collecting and representing data, process control using statistical methods, sources of variation, and calculating process capability indices. Process capability analysis determines if a process is capable of meeting specifications and identifies when processes are in or out of control. Maintaining statistical control is important for process stability and minimizing non-conforming output.
Spc training[1]
- 1. Minda Huf Limited STATISTICAL PROCESS CONTROL
- 2. Minda Huf Limited MEASURING PERFORMANCE PROCESS EVERY ACTIVITY IS A A SEQUENCE OF OPERATIONS & ANALYSING DATA COLLECTING,REPRESENTING CONTROL PROCESS STATISTICAL SPC & DATA COLLECTION STATISTICAL PROCESS CONTROL
- 3. Minda Huf Limited NOT ACCEPTABLE,GO GAUGE & NO GAUGE OK NOT OK ,ACCEPTABLE e.g IT IS A QUALITATIVE OR NOT ACCEPTABLE & AS OK NOT OK,ACCEPTABLE (ANY DATA IS EXPRESSED e.g DIM.,TEMP.,ANY VALUE) UNITS OF MEASUREMENT AS QUANTITATIVELY IN SPECIFIC BE MEASURED & EXPRESSED R/O , LINEAR DIMENSIONS MOT,PCD R/O,FACE (ANY DATA THAT CAN ATTRIBUTE VARIABLE INDUSTRIAL AREA INDUSTRIAL DATA CATEGORIES
- 4. Minda Huf Limited DATA ; DATA IS A COLLECTION OF NUMERICAL FACTS & FIGURES WE NEED FOR:- UNDERSTANDING THE CURRENT STATUS & PERFORMANCE EVALUATION ANALYSIS OF PROBLEM CONTROL OF PROCESS MAKING IMPROVEMENTS FINDING VARIATION & TRENDS JUDGEMENT OF ACCEPTANCE OR REJECTION
- 5. Minda Huf Limited IMPORTANCE OF DATA FACTS ARE BETTER THAN OPINION KNOW WHAT FACTS ARE EXPRESS FACTS THROUGH DATA CLARIFY THE PURPOSE OF COLLECTING DATA COLLECT DATA IN A SUITABLE MANNER FIND THE RIGHT WAY TO RECORD DATA ALL RELEVANT INFORMATION WHICH COULD BE OF VALUE SHOULD BE RECORDED PROVIDE QUICK FEEDBACK ON ANALYSIS OF DATA PREVENT FALSIFICATION OF DATA ENSURE THE SAMPLE IS REPRESENTATIVE
- 6. Minda Huf Limited A PROCESS CONTROL SYSTEM IT IS A FEEDBACK SYSTEM .SPC IS ONE TYPE OF FEEDBACK SYSTEM , FOUR ELEMENTS ARE:- 1.THE PROCESS:- MEANS COMBINATION OF SUPPLIERS , SUBCONTRATORS, PEOPLE , EQUIPMENT , INPUT MATERIAL , METHODS & ENVIRONMENT THAT WORK TOGETHER TO PRODUCE OUTPUT & CUSTOMER WHO USES THAT OUTPUT. 2.INFORMATION ABOUT PERFORMANCE:- UNDERSTANDING OF THE PROCESS & ITS INTERNAL VARIABILITY GIVES INFORMATION ABOUT PROCESS 3.ACTION ON PROCESS:- ACTION TAKEN TO PREVENT IS ECONOMICAL. THIS MAINTAINS THE STABILITY & THE VARIATION OF PROCESS OUTPUT. 4.ACTION OF THE OUTPUT:-LEAST ECONOMICAL WHEN RESTRICTED TO DETECTING AND CORRECTING OUT OF SPECS PRODUCTS W/O ADDRESSING PROCESS PROBLEM. IT SHOULD BE INTERIM MEASURE ONLY.
- 7. Minda Huf Limited VARIATION A NATURAL PHENOMENON INHERENT TO EVERY PROCESS 1. NO TWO THINGS ARE EXACTLY ALIKE AND WILL ALWAYS VARY. 2. A) PANDIT RAVISHANKAR CAN NOT PLAY ‘RAG BHAIRAVI’WITH HIS SITAR, TWICE IN EXACTLY THE SAME DAY. B) WE DO NOT KNOW WHICH WAY PANDIT RAVI SHANKAR’S NEXT SITAR RECITAL ON BHAIRAVI WILL DIFFER FROM THE LAST ONE. C) THERE IS SOMETHING IN HIS BHAIRAVI’S THAT MAKE THEM RECCOGNIZELY DIFFERENT FROM OTHERS.
- 8. Minda Huf Limited VARIATION EXPECTED (THIS WE MUST LIVE WITH) UNEXPECTED (THIS WE DO NOT HAVE TO LIVE WITH) TOTAL PROCESS VARIATION + = WHEN THE UNEXPECTED VARIATION IS ELIMINATED , WE HAVE THE TRUE PROCESS CAPABILITY.
- 9. Minda Huf Limited CONCEPT OF PRECISION & ACCURACY NOT ACCURATE A P A P ACCURATE A P A P
- 10. Minda Huf Limited COMMON VS. ASSIGNABLE CAUSES COMMON CAUSES (CHANCE CAUSES) (RANDOM CAUSES) ASSIGNABLE CAUSES (SPECIAL CAUSES) - CONSISTENT OF MANY INDIVIDUAL CAUSES. - ANY ONE CAUSES RESULTS IN ONLY A MINUTE AMOUNT OF VARIATION EXAMPLE : - SLIGHT VARIATION IN RAW MATERIAL - LACK OF HUMAN PERFECTION IN READING INSTRUMENTS & SETTING CONTROL - CANNOT BE ECONOMICALLY ELIMINATED - PROCESS FOLLOWS A PREDICTABLE (STATISTICAL) PATTERN - CONSISTENT OF JUST ONE OR TWO INDIVIDUAL CASES. - ANY CAUSE CAN RESULT IN LARGE VARIATION EXAMPLE : - BATCH OF DEFECTIVE MATERIAL - UNTRAINED OPERATOR - FAULTY SET UP - EASY TO DETECT & GENERALLY ECONOMICAL TO ELIMINATE - NO SPECIFIC PATTERN GOD ONLY KNOWS HUMAN ONLY KNOWS
- 11. Minda Huf Limited WHAT IS PROCESS ? - IT IS A SEQUENCE THROUGH WHICH INPUTS ARE CONVERTED OR TRANSFORMED INTO DESIRED OUTPUT. - IT IS A COMBINATION OF PEOPLE, M/Cs, MATERIALS, METHODS, ENVIRONMENTS, PLANNING AND ASSOCIATE SYSTEM. WHAT IS PROCESS CONTROL ? A PROCESS IS CONTROLLED WHEN IT IS OPERATING UNDER CHANCE (COMMON) CAUSES ONLY (FREE FROM ASSIGNABLE CAUSES) AND IS CENTRED AROUND THE TARGET. ADVANTAGES OF CONTROLLED PROCESS 1. PROCESS OPERATES AT ITS BEST WHEN UNDER COMMON CAUSES ALONE ( PREDICTABLE CAUSES). 2. IT EXHIBITS MINIMUM VARIATION. 3. LOWER LEVEL OF NON -CONFORMANCE. 4. SAMPLING INSPECTION POSSIBLE. STATISTICAL PROCESS CONTROL IS STATISTICAL METHOD OF SPOTLIGHT ABNORMAL PROCESS VARIATION LEADING TO THE SEARCH AND REMOVAL OF ASSIGNABLE OR SPECIAL CAUSESOF PROCESS VARIATION.
- 12. Minda Huf Limited CONTROLLED PROCESS V/S CAPABLE PROCESS (SPECIAL CAUSE V/S CHANCE CAUSE) - PROCESS OPERATING UNDER CHANCE & SPECIAL CAUSES - PROCESS OPERATING UNDER CHANCE CAUSES ONLY - PROCESS IN CONTROL BUT NOT CAPABLE - PROCESS IN CONTROL & CAPABLE
- 13. Minda Huf Limited MEASURES OF DISPERSION RANGE ( R ) - DIFFERENCE BETWEEN LARGEST & SMALLEST. STANDARD DEVIATION (S) NORMAL DISTRIBUTION - -IT IS A CONTINUOUS DISTRIBUTION - BELL SHAPED, SYMMETRICAL CURVE - AREA COVERAGE WITHIN ONE SIGMA - 68. 26% WITHIN TWO SIGMA- 95. 44% WITHIN THREE SIGMA- 99.73% WITHIN FOUR SIGMA - 99. 994%
- 14. Minda Huf Limited RATIONAL SUB GROUPING - ALL THE UNITS IN THE SAMPLE SHOULD HAVE BEEN PRODUCED UNDER HOMOGENEOUS SET OF CONDITION SO THAT VARIATION WITHIN THE UNITS IN THE SAMPLE MAY BE IRREDUCIBLE MINIMUM. - THE SUB GROUP SHOULD BE CHOSEN SO THAT OPPURTINITES FOR VARIATION AMONG THE UNITS WITHIN A SUBGROUPARE SMALL. IF THE VARIATION WITHIN A SUBGROUP REPRESENTS PIECE TO PIECE VARIABILITY OVER A VERY SHORT PERIOD OF TIME, THEN ANY UNUSUAL VARIATION BETWEEN SUBGROUPS WOULD REFLECT CHANGE IN PROCESS THAT SHOULD BE INVESTIGATED FOR EXISTENCE OFASSIGNABLE CAUSES AND TO TAKE APPROPRIATE ACTION. - SUB GROUP SHOULD TYPICALLY CONSIST OF 4 TO 5 CONSECUTIVELY PRODUCED PIECES. THE INTENTION IS IS THAT PIECES WITHIN EACH SUBGROUP WOULD ALL BE PRODUCED UNDER VERY SIMILAR PRODUCTION CONDITIONS OVER A VERYU SHORT TIME INTERVAL WITH NO OTHER SYSTEMATIC RELATIONSHIP TO EACH OTHER; HENCE VARIATION WITHIN EACH SUB GROUP WOULD PRIMARILY REFLECT COMMON CAUSES. - DURING INITIAL STUDY SUBGROUPS ARE OFTEN TAKEN CONSECUTIVELY / IN SHORT INTERVAL , TO DETECT WHETHER THE PROCESS CAN SHIFT TO SHOW OTHER INSTABILITY, TIME GROUP BETWEEN SUB GROUP CAN BE INCREASED.
- 15. Minda Huf Limited CONTROL LIMIT V/S SPECIFICATION LIMIT CONTROL LIMIT SPECIFICATION LIMIT - INHERENT TO THE PROCESS - CALCULATED FROM DATA GATHERED DURING THE PROCESS - USE TO JUDGE WHETHER A PROCESS IS IN “ STATISTICAL PROCESS CONTROL VOICE OF THE PROCESS - EXTERNAL TO THE PROCESS - GIVEN BY CUSTOMER (MAY BE INTERVAL CUSTOMER) ON DRAWING, OPERATION SHEET OR OTHER SPECIFICATION - TO JUDGE ACCEPTABILITY OF INDIVIDUAL PRODUCT. VOICE OF THE CUSTOMER
- 16. Minda Huf Limited WHAT IS PROCESS CAPABILITY PROCESS CAPABILITY IS SIMPLE THE VARIATION EXHIBITED BY A PROCESS UNDER COMMON INFLUENCE ONLY. IT IS THE VARIATION THAT WOULD BE SEEN IF ALL ELIMINABLE (SPECIAL CAUSES) SOURCES OF VARIATION WERE ELIMINATED. IT IS ALSO CALLED THE NATURAL TOLERANCE OF THE PROCESS. IT REFLECTS THE INHERENT VARIABILITY OF THE PROCESS AND TELLS THAT CAN BE EXPECTED FROM THE PROCESS IN FUTURE.
- 17. Minda Huf Limited WHAT IS PROCESS CAPABILITY STUDY A SCINTEFIC SYSTEMATIC PROCEDURE FOR DETERMINING THE CAPABILITY OF A PROCESS , AND IF NECESSARY , CHANGING/ MODIFYING THE PROCESS TO OBTAIN A BETTER CAPABILITY. A PROCESS CAPABILITY STUDY SHOULD :- 1. IMPLY SOLUTION TO THE PROBLEM. 2. FIND & ELIMINATE THE SPECIAL CAUSES THAT UPSET THE PROCESS. ROUTINE DATA COLLECTION & Cp ,Cpk REPORTS ARE MEANINGLESS , IF THE PROCESS IS NOT UNDER ‘STATISTICAL CONTROL’.
- 18. Minda Huf Limited CALCULATING THE PROCESS CAPABILITY INDEX Cp = PROCESS CAPABILITY Cpk = PROCESS CAPABILITY INDEX Cp = SPECIFICATION WIDTH PROCESS WIDTH = USL - LSL 6 s
- 19. Minda Huf Limited C p k = USL - X s3 ( WHEN ONLY USL EXISTS ) C p k = X - LSL s3 ( WHEN ONLY LSL EXISTS ) X & ARE USUALLY UNKNOWN & ARE ESTIMATED USING THE PROCESS DATA . s CALCULATING THE PROCESS CAPABILITY INDEX Cpk 1 = USL – X s3 Cpk 2 = X - LSL s3
- 20. Minda Huf Limited USES OF PROCESS CAPABILITIES CHOOSING FROM AMONG COMPETING PROCESSES THAT WHICH IS MOST APPROPRIATE FOR THE TOLERANCES TO BE MET . PLANNING THE INTERRELATIONSHIP OF SEQUENTIAL PROCESSES. FOR EXAMPLE , ONE PROCESS MAY DISTORT THE PRECISION ACHIEVED BY A PREDECESSOR PROCESS, AS IN HARDENING OF GEAR TEETH . QUANTIFYING THE RESPECTIVE PROCESS CAPABILITIES OFTEN POINTS THE WAY TO A SOLUTION . PROVIDING A QUANTIFIED BASIS FOR ESTABLISHING A SCHEDULE OF PERIODIC PROCESS CONTROL CHECKS AND ADJUSTEMENT. ASSIGNMENT MACHINES TO CLASSES OF WORK FOR WHICH THEY ARE BEST SUITED. SERVING AS A BASIS FOR SPECIFYING THE QUALITY PERFORMANCE REQUIREMENTS AT MACHINE’S STAGE.
- 21. Minda Huf Limited USLLSL INCAPABLE PROCESSCAPABLE PROCESS USLLSL 66 PROCESS CAPABILITY INDEX Vs PERFORMANCE Cp CONFORMING OUTPUT(%) NONCONFORMING OUTPUT(%) .5 86.66 13.5 .6 92.80 7.2 .7 96.40 3.6 .8 98.40 1.6 .9 99.30 0.7 .10 99.70 0.3 1.1 99.90 0.1 1.2 99.97 0.03 1.3 99.99 0.01 1.33 99.994 0.006
- 22. Minda Huf Limited PROCESS CAPABILITY CAPABLE BUT NOT PERFORMING PROCESS USLTXLSL PROCESS CAPABILITY INDEX Vs PERFORMANCE Cpk CONFORMING OUTPUT(%) NON CONFORMING OUTPUT(%) 0.5 93.3 6.7 0.6 96.4 3.6 0.7 98.2 1.8 0.8 99.2 0.8 0.9 99.65 0.35 1.0 99.86 0.14
- 23. Minda Huf Limited NORMAL DISTRIBUTION
- 24. Minda Huf Limited Cp - a measure of variation (Upper spec. limit= 40;Lower spec. limit =20; Process width Defined as ± 3 sigma limits) Cp = 20/30 = 0.67 -2 U.S. in 1970s 20 +2 A 40 40 1980s U.S. in +3 B 20 Cp = 20/20 = 1.0 -3 Cp = SPEC.WIDTH (S) PROCESS WIDTH (P)
- 25. Minda Huf Limited Cp - a measure of variation (Upper spec. limit= 40;Lower spec. limit =20; Process width Defined as ± 3 sigma limits) Cp = SPEC.WIDTH (S) PROCESS WIDTH (P)
- 26. Minda Huf Limited OUT OF CONTROL SITUATIONS When the actual variation exceeds the control limits, or when there is a pattern or a trend, or when all points are too close to the central line, or all points are too close to control limits, or there is run of 7 or more points. All such cases are abnormal (probability of such thing happening is very remote) and are signals of assignable causes. (A) Unusual Patterns 1. Trend 2. Points on same side 3. Stratification (all points lying within 1 *) 4. Mixture (All points close to control limits) 5. Cyclic Pattern (Time related assignable cause) s
- 27. Minda Huf Limited CONTROL CHARTS FOR VARIABLES A) Plot of X B) Plot of R
- 28. Minda Huf Limited CONTROL CHARTS FOR ATTRIBUTES A) np Charts (Number of defective - charts) B) c Charts (Number of defects - charts) A) np Charts It follows binomial distribution UCL, LCL = np + 3 np (1-p) B) c Charts It follows poisson distribution UCL, LCL = C + 3 C