Quality Engineering学习笔记。
课件引用于香港理工大学ISE369课程
Content
- Chapter 1: Eight Dimensions of Quality
- Chapter 2: Quality System Planning in an Organization (QC,QA,CQI,QM,QMS; QC Tools)
- Chapter 3: The Seven Tools of Quality
- Chapter 3: QMS & Certification; Quality Audit; ISO 9000
- Chapter 4: Statistical Quality Control
- Chapter 5: Process Variation and Process Capability
- Chapter 6: Statistical Tolerancing
- Chapter 7: Acceptance Sampling
- Chapter 8: Reliability fundamentals and life Testing
- Chapter 9: Fault Tree Analysis (FTA)
- Chapter 10: Failure Modes, Effects and Criticality Analysis (FMECA)
Chapter 1: Eight Dimensions of Quality
1.Performance(will the product do the intended job?)
2.Reliability(how often does the product fail?)
3.Durability(how long does the product last?)
4.Serviceability(how easy is it to repair the product?)
5.Aesthetics(what does the product look like?)
6.Features(what does the product do?)
7.Perceived Quality(what is the reputation of the company or its products?)
8.Conformance to Standards(is the product made exactly as the designer intended?)
Chapter 2: Quality System Planning in an Organization (QC,QA,CQI,QM,QMS; QC Tools)
(i) Quality Control (QC)
- Quality Control is the operational techniques and activities that are used to fulfill requirements for quality.
- It aims both at monitoring a process and at eliminating causes of unsatisfactory performance at relevant stages of the quality loop, in order to result in economic effectiveness.
(ii) Quality Assurance
- Quality Assurance (QA) is “all those planned and systematic actions necessary to provide adequate confidence that a product or service will satisfy given requirements for quality”.
(iii) Continual Quality Improvement (CQI)
- Establish, govern and sustain by a quality system; Management programme such as Six Sigma, TQM; Permanent function in the company (usually under quality department or customer service department)
(iv) Quality Management (QM)
- Planning, establishing, staffing, controlling QC/QA/CQI/QS/other quality functions
Quality Management System (QMS)
Systematic steps: P-D-C-A Cycle
Plan
- establish the objectives and processes necessary to deliver results in accordance with customer requirements and the organization’s policies;
Do
- implement the processes;
Check
- monitor and measure processes and productagainst policies, objectives and requirements for the product and report the results;
Act
- take actions to continually improve process performance;
ISO 9001: 2008 Quality management systems –requirements
Chapter 3: The Seven Tools of Quality
Flow Chart
Whether the (non-value-adding) incoming inspection could be eliminated by using supplier certification.
Run Chart
The run chart shows the history of variation over a period. Run charts are useful in providing an indication of a possible shift in the characteristic being plotted.
Process Control Chart
The control chart is a tool which provide a means of determining what type of variation is present in a process and whether the process is performing predictably.
These lines are the upper control limit (UCL)
, lower control limit (LCL)
, and centreline (CL)
.
Check Sheet
Pareto diagram
Pareto diagram provides the same type of insight into the most and least frequent occurrences.
The Pareto Diagram sorts the data categories from highest to lowestand also shows the cumulative frequencies. This information is helpful in focusing attention on the highest-priority category.
Cause and Effect Diagram
It is also called an Ishikawa diagram (or fish bone diagram)that is useful in many analysis, as it illustrates the relationship between cause and effect in a rational manner.
There are three types of branches in a fish bone diagram:
- Primary branch–represents the effect and is typically labeled on the right side of the diagram.
- Major branches–Each of them corresponds to a major cause that directly relates to the effect.
- Minor branches–They correspond to more detailed causal factors.
the R&D engineer-> the methods
the quality engineer-> the materials
the production supervisor-> the workers (man)
the process engineer-> the machines
Scatter diagram
Scatter diagram is graphical tool that is used to describe the influence or correlationthat one variable has on another.
It usually displays points on a graph each of which is representing the observed value of one variable, and corresponding value of another variable.
Histogram
Histogram gives a simple graphical view of accumulated
data including its dispersion and central tendency. It also provides the easiest way to evaluate the distribution of data.
Chapter 3: QMS & Certification; Quality Audit; ISO 9000
Chapter 4: Statistical Quality Control
X bar and R Charts
- average of x and r, and the total average
- from variable table for n=? A2 D3 D4
- UCL LCL formula
- chart
- in control/ out of control?
to understand whether it’s in control, if one or more points is more than UCL or less than LCL, then that’s not in control limits. / also increasing or decreasing trend, but still need to know as eventually it will decrease. Out of control-> investigate, set up quality improvement team.
Charts for non-conforming units: p chart and c chart
If there is C/P chart, you may consider both of them
P chart -> you cannot have any 0; if it’s possible, better to use p chart
C chart ->
P chart (for defective units %)
- find np and p, and p average
- UCL LCL formula
- chart
- in control/ out of control?
p = n p ( n u m b e r o f n o n c o n f o r m i n g ) / n ( n u m b e r i n s p e c t e d ) p = np (number of nonconforming) / n (number inspected) p=np(numberofnonconforming)/n(numberinspected)
in control, but do you accept 2% defect? It’s normal, but if you don’t accept, you need to improve.
• The number of points on each side of the centerline should be balanced, with the majority of the points near the centerline.
• There should be no patterns in the data, such as trends, runs, cycles, or sudden shifts in level.
• One difference between the interpretation of a variables control chart and a p chart is the desirability in the latter of having points that approach the lower control limits.
C chart (for defective units -> no need total inspects)
- find number of sample , number of defects
- average defects/sample
- UCL LCL formula
- chart
- in control/ out of control?
In control, number of non-conformity is around 5.4, it’s fine -> carry on/
Point going out/ trend, then need to stop the process
• The data points on a c chart should flow smoothly back and forth across the centerline and be balanced on either side of the centerline. The majority of the points should be near the centerline.
• Once again it is desirable to have the points approach the lower control limits or zero, showing a reduction in the count of nonconformities.
Chapter 5: Process Variation and Process Capability
The larger the Cp, the better the process capability.
- A process controlling the length of a shaft is in control with its process mean equal to 100 units, and estimated standard deviation (s) equal to 1.05. The process specifications on its length are 95±10 units. The process has a normal distribution.
(a) Estimate the potential capability (Cp).
(b) Estimate the actual capability(Cpk)
© By comparing Cp and Cpk values calculated above, what can you comment on the process
Cp=Cpk -> the process is centered at the midpoint of the specifications.
cpk<0 which means the process lies outside specification
Cp= 0.767, you need to consider whether you can accept 20000 defective PPM when buying from ABC company. If in hurry, perhaps yes.
Chapter 6: Statistical Tolerancing
The essence of technical specification consists of two elements:
- nominal size
- tolerance
Control of quality is largely a matter of controlling variation.
Chapter 7: Acceptance Sampling
Producer Risk (alpha): Acceptable quality level (AQL)
Consumer Risk (beta): Limiting quality (LQ)
- 1000 samples (N=1000), take n= 240 samples, and if c <=2 (less than or equal to 2 defects), then it will be accepted
- 1000 samples (N=1000), take n= 170 samples, and if c <=1 (less than or equal to 1 defects), then it will be accepted
- 1000 samples (N=1000), take n= 100 samples, and if c <=0 (less than or equal to 0 defects), then it will be accepted
Otherwise rejected the whole lots (1000)
(i) All is okay because at 2% defectives, there is more than 10% acceptance.
(ii) how many lots will be rejected?
1 % -> 0.56 0.48 0.35
0.5 % -> 0.92 0.8 0.6
0.2 % -> 0.99 0.96 0.8
(iii) whatever the perceived defects, the 1000,240,2 should be chosen although there is more workload, the risk is lower
Chapter 8: Reliability fundamentals and life Testing
Mean-time-to-failure (MTTF)
Mean time to failure (MTTF) is the length of time a device or other product is expected to last in operation. MTTF is one of many ways to evaluate the reliability of pieces of hardware or other technology.
M T T F = t h e t a = ( ∑ i = 1 n t i ) / n MTTF = theta = (\sum_{i=1}^{n} t_{i}) / n MTTF=theta=(i=1∑nti)/n
Estimate the mean-time-to-failure and the failure rate per 100 hours if a sample of five units is tested to failure and the time to failure of each is t1=111 hours, t2=147 hours, t3 = 152 hours, t4 = 176 hours, t5 = 195 hours.
Failure number Truncated Test
Non-replacement test
t h e t a = ( ∑ i = 1 r t i + ( n − r ) ∗ ( t r ) ) / r theta = (\sum_{i=1}^{r} t_{i} + (n-r) * (t_r)) / r theta=(i=1∑rti+(n−r)∗(tr))/r
Replacement test
t h e t a = ( n ∗ ( t r ) ) / r theta = (n * (t_r)) /r theta=(n∗(tr))/r
Time Truncated Test
In such cases, as the TTF of the last failure, in general, will not happen exactly at the time of truncation ‘t’, according to common convention, it is assumed that just one more item was about to fail and, therefore, the divisor for the total item test- hours is taken as (r + 1) when the actual number of observed failures is ‘r’.
Non-replacement test
t h e t a = ( ∑ i = 1 r t i + ( n − r ) ∗ ( t ) ) / ( r + 1 ) theta = (\sum_{i=1}^{r} t_{i} + (n-r) * (t)) / (r+1) theta=(i=1∑rti+(n−r)∗(t))/(r+1)
While testing a product sample of 12 items, it was decided to truncate the test at 5000 hours. The observed failures were at 1450, 3492, and 4013 hours.
( 1450 + 3492 + 4013 + 9 ∗ 5000 ) / ( 3 + 1 ) = 13488.75 H o u r s (1450 + 3492 + 4013 + 9*5000) / (3+1) = 13488.75 Hours (1450+3492+4013+9∗5000)/(3+1)=13488.75Hours
Replacement test
t h e t a = ( n ∗ t ) / ( r + 1 ) theta = (n*t) / (r+1) theta=(n∗t)/(r+1)
In a time-truncated, replacement type of test, 7 items were tested for 1000 hours during which 4 failures were observed to occur.
7 ∗ 1000 / ( 4 + 1 ) = 1400 H o u r s 7 * 1000 / (4+1) = 1400 Hours 7∗1000/(4+1)=1400Hours
Weibull Distribution
The Weibull distribution is a general purpose reliability distribution used to model material strength, times-to-failure of electronic and mechanical components, equipment or systems.
Weibull found that in general, distributions of data on product life will fit an expression of the following form:
Situation 1:
R ( t ) = e x p ( − ( ( t − t 0 ) / n ) β ) β − S h a p e F a c t o r t 0 − S t a r t i n g P o i n t n − C h a r a c t e r i s t i c L i f e R(t) = exp (-((t-t_0)/n)^{\beta}) \\ {\beta} - Shape Factor \\ t_0 -Starting Point \\ n - Characteristic Life R(t)=exp(−((t−t0)/n)β)β−ShapeFactort0−StartingPointn−CharacteristicLife
Situation 2:
β = 1 a n d t 0 = 0 R ( t ) = e x p ( − λ t ) {\beta} = 1 \\ and \\ t_0 = 0 \\ R(t) = exp(-{\lambda}t) \\ β=1andt0=0R(t)=exp(−λt)
An exponential model with
λ = 1 / n \\ {\lambda} = 1/n λ=1/n
Chapter 9: Fault Tree Analysis (FTA)
Chapter 10: Failure Modes, Effects and Criticality Analysis (FMECA)
Overall
b. is better because Quality control -> probability
For very complex materials, users cannot identify all of the components.
Part-level FMECA
Functional-level FMECA
Severity Index
Consequences are often placed into one of various categories.
Probability Index
How likely is the event to occur under the circumstances described and given the required precursor events?
Criticality Analysis (calculated index)
Rank failures by risk (severity and probability)