Statistical Process Control Flashcards
Precision vs Accuracy
both are very important
Accuracy: averages and means
Consistent precision: standard deviation and range
Sample Statistics vs Population Parameters
Statistics:
x bar: sample mean or average
s: sample standard deviation
Parameters:
mu: population mean or average
sigma: population standard deviation
The Normal Distribution
Property 1: normal distribution can be described completely by knowing only the mean and standard deviation
Property 2: area under sections of the curve can be used to estimate the cumulative probability of a certain event occurring
Voice of process = 3 standard deviations away from the mean
SPC
statistical process control, measures how we are doing
Basic assumptions (tenets) of process quality control:
1. every process has random variation in it
2. production processes are not usually found in a “state of control”
ONLY RANDOM VARIATION has “state of control” = stable and predictable
control chart components
upper control limit (UCL) = + 3 SD above the mean / center line / x bar
lower control limit: UCL = - 3 SD below the mean / center line / x bar
center line
Control charts are specific to the data type being observed - variable data (or “continuous”) uses “x-bar” and “R-charts” and attribute data (or “discrete”) uses “P-charts” (% defects)
two types of variation
Common cause (noise)
- present in every process
- randomly produced by the process itself (“the way we do business/work”)
- can be removed and/or lessened with a fundamental (large scale) change in the process
* a process is stable, predictable, and control when ONLY common (or random) cause variation exists in the process
Special cause (signals) NOT TYPICAL
- unpredictable / not random
- typically large in comparison to common cause variation
- caused by unique or assignable (“explain”) disturbances or a series of them
- can be removed/lessened by basic process controls and monitoring
* a process exhibiting special (or assignable) cause variation is said to be out-of-control and unstable (doesn’t mean it’s bad!)
Two types of data used for measurements
attribute data (qualitative) - discrete data are generally counted
- attendance, basketball/golf points, votes, demographics/gender, shipping
* uses P chart
variable data (quantitative, decimal point as a meaning) - continuous data are generally measured
- temperature, money, stock prices. grades, calories * everything can be measured
* uses x bar chart and r chart
Control chart rules
- helps identify special cause events present in our process
- we use phrase “out of control” when a rule has been broken
- Test #1: one point outside the UCL to LCL zone (3 sigma limit)
- Pattern Rule: any apparent pattern that repeats itself
- this means something unexpected has happened
match interpretations
- every trend has common cause variation
- common cause variation: process is stable and in-control
- special cause variation: process is unstable and out of control
- increasing trend: special cause variation, process is out of control and unstable
- pattern: special cause variation, signal, process is unstable and out of control
process variation
SPC is used to describe/analyze process outputs bc all processes has variation. “mean” = location/average, and “standard deviation” = range/spread
control charts classify variation as 1) common or naturally or randomly occurring causes) or (2) “special or assignable” (where investigation can reveal it’s cause)
interpreting control charts
*A process that appears be performing as expected is said to be “in-control” or
“stable”. However, because we are sampling we are never 100% sure. A type II error incorrectly concludes “in control”. A type I error is when we incorrectly conclude a process to be “out of control”.
*When patterns indicate other than natural or common cause variation, it is considered “out of control” or “un-stable”. In order to determine “why”, “special” or “assignable” causes (either negative or positive) are investigated.
Assignable
Variation that is caused by factors that can be clearly identified and possibly even managed is called ___________ variation.
upper and lower specification limits
The range of values in a measure associated with a process that is allowable given the intended use of the product or service.
Attributes
Quality characteristics that are classified as either conforming or not conforming to specification
Variables
Quality characteristics that are measured in actual weight, volume, inches, centimeters, or other measure
