Chapter 3

Statistical process control

"a powerful collection of problem-solving tools useful in achieving process stability and improving capability through the reduction of variability."

How to change a Pareto chart to situations with serious consequences

(1) Using a weighting scheme to modify the frequency counts (2) Accompany the frequency Pareto Chart analysis with a cost or exposure pareto chart

Estimating Device

(ex. Control chart) from the chart you may estimate certain process parameters such as mean, std dev, fraction nonconforming, or fallout - Used to determine *capability* of a process to produce acceptable products

Check Sheet

- For collecting historical or current operating data - summarize all historical defect data in a time oriented summary for looking into trends or meaningful patterns

Non stationary variation

-data that "wanders about" without any sense of a fixed mean -typical in chemical and process industries -Stabilized using engineering process control such as feedback control.

Autocorrelated process data

-moves in long "runs" on either side of the mean

7 major tools of SPC ("The Magnificent Seven)

1. Histogram or stem-and-leaf plot 2. Check sheet 3. Pareto chart 4. Cause-and-effect diagram 5. Defect concentration diagram 6. Scatter diagram 7. Control chart

Limitations of control chart and what to do

A control chart will only *detect* assignable causes. *action* must be taken to eliminate these causes. -find the *root cause* of the problem to attack it

Out of control Action Plan (OCAP)

A flow chart or text-based description of the sequence of activities that must take place following the occurrence of an *activating event* - Consists of *checkpoints* (potential assignable causes) and *terminators* (actions taken to resolve the out-of-control condition) -*living document* and will be modified over time

Cause-and-Effect (or Ishikawa Diagram)

A formal tool frequently useful in un layering potential causes constructed by teams

Pareto Chart

A frequency distribution (or histogram) of attribute data arranged by category - can be used in nonmanufacturing applications

Control Chart

A graphical display of a quality characteristic that has been measured or computed from a sample versus the sample number or time.

Defect Concentration Diagram

A picture of the unit showing all relevant views to analyze whether the location of the defects on the unit conveys any useful information about the potential causes of the defects.

Out-of-control process

A process that is operating in the presence of assignable causes.

Chance causes of variation

A process that is operating with only chance causes of variation. A "stable system of chance causes." only has natural variability.

Scatter diagram

A useful plot for identifying a potential relationship between two variables. Data is collected (yi, xi) - can be positively correlated, correlation is not causation

Factorial Experiment

An experiment design in which all possible combinations of these factors levels would run

Fractional Factorial Design

An experimental design in which only a portion of all possible combination of factor levels is run.

Process

An organized sequence of activities that produces an output (product or service) that adds value to the organization

Non-parametric statistical procedures

Don't have underlying assumption of normality and can be used as alternatives to procedure t-tests.

Upper control limit (UCL) and the lower control limit (LCL)

If the process is in control, nearly all of the sample points will fall between them

current-state value stream map

It shows what is happening in the process as it is now defined.

Shewhart Control Charts

L is the distance of the control limits from the center line expressed in standard deviation units. UCL=Mu +Lsigma LCL=Mu - Lsigma

Activating event

Out-of-control signs in a control chart

Attributes control charts

Quality characteristics that can be judged as either conforming or nonconforming on the basis of whether it has certain attributes.

Control Chart: Center Line

Represents the average value of the quality characteristic corresponding to the in-control state

sigma in the control chart

Standard deviation of the statistic plotted on the chart

Tolerance Diagram

The extremes of the samples are connected with a vertical line

Type of variability: Stationary Behavior

The process data vary around the fixed mean in a stable or predictable manner. - In-control process

Variable control chart

The quality characteristic can be measured and expressed as a number on some continuous scale. -quality characteristic described with variability

Design of the control chart

The selection of sample size, control limits, and frequency of sampling.

Use of a control chart

To *improve* a process because (1) Most processes do not operate in a state of statistical control (2) Identifying assignable causes and if they can be eliminated from the process

future state value stream map

constructed to show what a redefined process should look like.

Generalized linear models

example logistic regression can be used with binomial data and poisson regression can be used with many kinds of count data

Process-capability studies

have considerable impact on many management decision problems that occur over the product cycle,

transforming data

how to deal with non-normality.

Benefits of Scatter diagram

identifying potential relationships, designed experiments, and must be used to verify causality

Assignable causes of variation

improperly adjusted or controlled machines, operator errors, or defective raw material.

Assumptions made with Shewhart control charts

incontrol process data that is stationary and correlated

Value stream mapping

like a flow chart, but it usually incorporates other information about the activities that are occurring at each step of the process and information that is required or generated

Operation Process Chart

symbolic flow chart

Uncorrelated data

the observations give the appearance of having been drawn from at random form a stable population (perhaps a normal distribution) -*"white noise"*: past data does not help in predicting future values

Stacked pareto chart

the two categories are stacked to compare (ex. supplier A vs. B components)