Process Behavior and Control
The terms in control
and out of control
are typically used when referring to a stable or unstable process. A process is in control (stable) when the average and standard deviations are known and predictable. A process is out of control (unstable) when either the average or standard deviation is changing or unpredictable.
- In control: Stable, predictable, consistent, unchanging
- Out of control: Unstable, unpredictable, inconsistent, changing
An in-control process has many benefits:
- Scrap and rework estimates can be made prior to production.
- Machine settings can be adjusted to optimize throughput.
- Engineers can incorporate statistical tolerance into their drawings, increasing component tolerances without compromising assembly performance.
- Product designs can be statistically modeled to accurately predict fit and performance yields prior to prototype assembly.
- Machine utilization can be optimized (e.g., high-precision machines and resources will not be wasted on manufacturing low-precision dimensions).
- Process-improvement resources will be better spent.
Remember, being in control does not mean that the process is within specification. A process can be extremely stable while consistently producing bad product.
Out of Control
A process is usually judged to be out of control based on five commonly used control chart rules. These rules signal a change in either the process average or the variation.
- Points are beyond control limits.
- Eight or more consecutive points are either above or below the centerline.
- Four out of five consecutive points are in or beyond the 2-sigma zone (referred to as zone B in the graphic).
- Six points or more point in a row are steadily increasing or decreasing.
- Two out of three consecutive points are in the 3-sigma region (referred to as zone A in the graphic).
Even an out-of-control process can reveal useful information. By using SPC to measure out-of-control processes, you can do the following:
- Detect both unwanted and desirable process changes.
- Prove whether a process change resulted in an improvement.
- Determine when to make a process change.
- Verify measurement system improvements.
Control charts, sometimes called process behavior charts, are tools to determine whether a process is stable or unstable.