# Statistical Process Control - DOC by Pc9tHgL

VIEWS: 12 PAGES: 3

• pg 1
```									                    Statistical Process Control
Statistical process control is a technique for ensuring any process used to
delivery a service or produce goods has the fidelity to meet standards. All
processes are subject to variability and in the 1920’s Dr. Walter Shewhart
developed a system using statistics to track this variability. Shewhart further
refined his thinking to distinguish between types of variations found in all
processes i.e. common and special causes of variation. These variations are
now commonly called natural and assignable causes. Investigation into this
phenomenon led to Shewhart’s development of a statistical tool to separate the
two types of variation—the control chart.

Statistical process control (SPC) is a funnel quality and productivity. It uses very
basic statistics to control processes. A process is any collection of activities that
effects a sequential change towards an objective. Statistical process control is
used to measure performance of a process. A process is said to be in statistical
control when the only source of variation is common or natural causes. The
process is brought into statistical control by detecting and eliminating special
causes of variation. Performance is predictable and the ability of the process to
meet customer needs and expectations can be evaluated when a process is in
control. A process that is in control will consistently produce parts or service
within its own natural tolerance limits. This is done by eliminating all of the
special causes of variation that exist. The first objective of SPC is to get the
process in control, which means the identification and elimination of special
causes of variation.

Natural variations affect almost every production process and are to be
expected. This is variation that has several common causes. Common or
natural causes are such issues as heat, vibration, humidity, that either
individually or collectively causes variation. Common causes are very difficult to
identify and correct. Natural variations, behave like a constant system of chance
causes. In effect this variability is inherent in the process. Although individual
values measured in a process are different, as a group they form a pattern that
can be described as a distribution. When distributions are normal (output
measurements) remain within specified limits, the process is said to be in control
and natural variations are tolerated.

Assignable or special variations are those effects on a process that are not
built in. Causes of the assignable variations are unpredictable and can be
directly identified, corrected (particularly when the time they occur is identified) to
bring the process into control. Assignable variation can be traced to a specific
reason. Factors such as machine wear, misadjusted equipment, fatigued or
untrained workers or new batches of raw material are all potential sources of
assignable variation. The objective of management is to identify and eliminate
assignable variations so that processes will remain in control.

WEC 10/05 MET 345
Shewhart’s contribution to statistical process control took form in a tool he called
a control chart. Control charts are the best tool to bring a process into control.
These charts are simple statistical charts for detecting special causes of variation
in the process at the time they exist. In addition these charts will measure the
natural tolerance of the process due to normal variation or common causes.
Control charts are the main tool to distinguish between random, natural variability
and nonrandom variability. The basis for building a control chart is the concept of
sampling and distribution which describes random (natural) variability. Sample
measurements are made and plotted on the chart. Examinations of the
characteristics of this data help to distinguish between natural variations and
variations due to assignable causes. A process in control will produce
measurable data (and presumably parts) within established control limits.

After a process is in control (stable) and producing consistently within its natural
tolerance, it can then (and only then) be compared to the engineering tolerance
limits to see if it is capable of meeting those limits. A capable process is directly
related to the ability of the process to produce parts consistently within the
drawing tolerance limits. Capability cannot be studied until the process is in
control, because it will not be consistent enough to trust the results of the study.
It is possible to have a stable process (in control) that does not meet the
tolerance limits specified on a blueprint for producing an acceptable part.

Consider a drilling operation that can consistently produce round holes, always
within ±.010” location on raw material and review of the data from sampling
measurements confirm that no assignable causes are detected. The process is
in control. A look at the blueprint specifications may show that the location
tolerance for the holes is only .005 inches. The process is in control but not
capable. Capability by definition means the control limits that define the process
must be inside the part tolerance limits. In this example the process is in control
but in capable of producing parts to specifications.

Three primary conditions are needed for a capable process;
1. The process must be in a state of statistical control
2. The control limits must be well inside the tolerance limits.

The goal is a process capability ratio (PRC) of
.075 max.

WEC 10/05 MET 345
Where:

3.    Prior to calculating the PRC, test the distribution of individuals for
normality and centrality, (which means that X should be close to the
nominal dimension).

Advantages of a process in control:

1. There is more uniformity (less variation) between units, which decreases the
chances of producing a defective product
2.                     Fewer samples are necessary to judge product quality,
since they are more uniform
3. Inspection costs are reduced.
4. A more accurate definition of capability and sound business decisions can be
 Selection of specification limits
 Knowledge of the yield of the process
 Selection of the appropriate process

5. Percentage of product produced within specification limits can more
accurately be calculated (% yield).
6. Consistently less scrap, rework and other wastes of productive time and
money.

WEC 10/05 MET 345

```
To top