Stat 513 / IE 530 Midterm 1
Poisson Assumptions
-The probability of occurrences of an event is constant for all sub intervals -There can be no more than one occurrence in an interval -occurrences are independent
Common Cause Variation
-Variation inherent to a process -Can only be reduced by changing the process -Must be Sampled under homogeneous conditions (Rational Subgrouping)
Normal Distributions
1 standard dev from mean both ways - 68% data 2 " " - 95% data 3 " " - 99% data
Rules for out of control process
1.) 3 Sigma Limits 2.) Trends 3.) Runs 4.) Cycles 5.) 8 points on one side of side of the mean
Six Sigma Steps
1.) Define 2.) Measure 3.) Analyze 4.) Improve 5.) Control
Process Mapping Steps
1.) Identify process inputs and outputs 2.)Identify all steps in process 3.) List output variables at each step 4.) List input variables at each step and whether they are controlled or not 5.) Adde process specs for input variables
Gage to Precision to Tolerance Ratio
(P/T) = 6sigma(gage) / (USL - LSL) P/T < .1 is acceptable
Cp Pp
(USL - LSL) / 6 sigma
Rational Subgrouping
-Minimizes special cause variation w/i subgroups -Maximizes change special cause variation will be detected b/t subgroups -Special Cause variation appears in X-Bar chart between subgroups -Should not appear in R chart
Government Course Inspection
-Performed by US gov at suppliers facility -Very Expensive (taxpayers) -Can delay deliveries Used when -Required by contract
Process Audit
-Performed by customer personnel -Assesses one or more processes of suppliers -Used to assure capability and identify weaknesses in process
System Audit
-Performed by customer personnel -Assesses the entire quality system of the supplier -Used to assure capability and identify weaknesses in system
Customer Course Inspection
-Performed by customers personal at suppliers facility -Very Expensive Used when -required by contract -supplier owns large sophisticated test equipment -new supplier
Qualification Testing
-Performed by supplier or third part test lab -Redundant testing might be done by customer -Performed before product used during production AKA -First article testing -PPAP -ISR
Third Part Source Inspection
-Performed by the sub-contractors of the customer at suppliers facility Expensive, but less than customer personal Used When -required by contract -supplier owns large sophisticated test equipment -new supplier
Improving Capability
1.) Standardize Process 2.)Process is in control (can only reduce special cause var) 3.)Capability (process mean and variation dependent) 4.)Measurement System Analysis says estimated common cause variation is due to process/operator/equipment 5.)Process mean is a function of the process inputs 6.)These are obtained from process map 7.)Importance of process prioritized with cause and effect matrix 8.)Prioritized once can be used in a designed experiment to model process 9.)Model process mean 10.)Find the settings to make the best mean on target 11.) Settings can sometimes be used to reduce process variation 12.) Solution does not fundamentally change the process
Binomial
1.) n Independent Bernoulli trials 2.) Number of trials is fixed 3.) There are ONLY two outcomes (success and failure) (p and q) (p and (1-p)) E(x)=np, Var(x) = np(1-p)
Sustained changes in process mean
1.) sustained shift in process mean to a new value 2.)slow trending in the process mean up or down
Hyper Geometric
1.) the population or set to be sampled consists of N individuals, objects, or elements (finite population)
If we used two sigma limits in the X-bar chart instead of three sigma limits, how often would a point be outside of the limits if our process was in control
5%
Six Sigma and TQM
Differentiated by strategy Six Sigma requires commitment of top leadership continuously through years of implementation
No quality initiative can succeed without the support of
Everyone
Sample Receiving inspection
Expensive for customer Reduces risk of using nonconforming material
Skip lot receiving inspection
Expensive for customer, but less than sample receiving relies on inspection personnel to accurately select lots reduces risk of nonconforming material, but not as much as sample receiving
After using a Pareto diagram to decide which defect to work on, to find the cause one should then use
Fishbone Diagram, Cause and Effect Matrix
What is the usual reason for white space in a control chart
Fixed Differences, Overestimated theta
What causes white space
Fixed differences, overestimated sigma, limits too narrow, excessive roundoff
Cp and Cpk
Short Term Spans days-weeks 30-50 data points >1 capable <1 not capable =1 marginally capable
Name three ways in which we can use a Pareto chart to prioritize which problem to address first
Shows which defects show up the most, what the customer prioritizes, which are critical to the process
You would use an XmR chart if the subgroups were
Size of 1
Shifts in process mean and process knowledge
detected in x-bar chart these are usually not detected by the 4 rules of thumb
Poisson
lambda is mean occurrence gives probability of a given number of events occurring in a fixed interval of time E(x) = lambda = V(x)
Cpk Ppk
min(USL - Xbar, Xbar - LSL) / 3 sigma
How many observations are used in computing ranges for moving a range chart
one per time period
Traceability
organizations are contractually required to maintain a measurement system that is traceable, like ISO
Hierarchy of Standards
owing to the technology required to maintain the primary reference standards, a hierarchy of standards
Cpk
short term capability
Give one reason for using an XmR chart instead of an X-bar and R-chart
subgroup is too small
Estimate of common cause
sum of true product variation, variation due to different operations (reproducibility) and variance of measurement equipment error (repeatability) sigma(m)^2 = sigma(p)^2 + sigma(o)^2 + sigma(e)^2 Product variation too large, invest in process. Operator variation too large, invest in training Error variation too large, invest in measurement eq.
Cpk >= Ppk
this is because variability increases over the long term
Process Flow Charts
Graphical Display of Process, identifies value added and non value added process steps, inputs/outputs, data collection points
How do you make sure that within a subgroup variation reflects Common Cause
Homogeneous Conditions
The main components of Quality Cost
Prevention, Appraisal, Failure
Three Components of Quality Cost
Prevention, Appraisal, Failure
Name two important uses for Flowcharting a process
Understanding a process, identifying where/what you need to control
Pp and Ppk
Long Term weeks-months 100-200 data points <1 not meeting specs >1 meeting specs =1 marginally meeting specs
Ppk
Long term capability
Six Sigma
Measure of variability, name is based of the laws of six sigma (Data distribution) Six sigma defect rules 3.24 defects per million
Sampling Plan - Rational Subgrouping
choose subgroups in small region of space, time
White Space
comes from subgroups across fixed differences
Deming said that the improvement of processes should be
critical
What are the four parts of the continuous improvement cycle
Plan, Do, Check, Act
Signs that a process is out of control
8 points on one side, cyclical, out of limits, seasonal,
Distribution Rules of Thumb
Binomial - number of successes/failures in a sample size n Poisson - number of blemishes in a continuous area of a certain size Normal- continuous characteristics of an item and average the values Geometric- monitor for stability and wait for a process to become unstable
Qualitative Data
C & U, Poissan, p and np binomial, XmR Binomial
What attributes charts are used for discrete data
C, u ; np, p ; XmR
Why Define Processes?
Individuals can understand tasks at hand, documentation and standardization, common understanding of procedures
Certificate of Compliance
Legal documents signed by an officer or supplier, part fully comply with contract requirements
If the data is skewed to the right, which is larger the mean or the median
Median
What is a standardized Process
One where everyone follows the same agreed upon procedures
In doing a root cause analysis, name three people who must be involved
Operator
Standardized Processes are not necessarily
Optimal, Current Best Methods
If you detect the presence of an out of control point that probably means that the process is
Out of control, special cause
Moving Average Charts
UWMA (uniformly weighted moving average chart) -for sustained shift in the process mean to a new value) EWMA (exponentially weighted moving average chart) -for a slowly trending process mean
For quantitive data use
X, R, XmR
What attributes charts are used for cont. data
X-bar, R charts, XmR
Gage Capability
accuracy, repeatability, reproducibility, and stability combined into a single unit Poor reproducibility would indicate a need for the operator training in use of the gage or a req for better readability. Poor repeatability would suggest need for gage maintenance, gage redesign or selection of another measure device
Modern methods of training should be instituted for
all employees to institute control
Best way to decrease variation
better measurement equipment
