Skip to content

Statistical Process Control & Management

Obviously Delphi started its focus in the financial markets. But, it was fortunate to have been able to engage in some fairly innovative forays into very different spheres of work. One of the areas that was particularly enjoyable for Delphi involved the solving of problems with very physical characteristics – process control or process management.

In the manufacture of certain products – say machined parts in the aerospace industry – the manufacturing process involves the use of physical tools (like bits or other cutting instruments) that might become dull or loose their initial cutting properties over time and use. As the process progresses, the product being manufactured can become less-and-less consistent and harder and harder to stay within the manufacturing tolerances or requirements.

Consequently, when the manufacturing characteristics of complex systems need to be monitored carefully for detailed fault analysis, then all parts’ batch data need to be collected and stored with permanent tracking of each part in each assembly – then a real, solid process management program is very useful. Delphi has made specific contributions in this field of endeavor.

Western Electric Rules

As an example, consider the Western Electric rules – or WECO rules to some… These rules are most easily understood as relating to control charts.

Imagine that for each ‘run’ of a manufactured part, let’s say that 1,000 items are produced. Let’s further imagine that these are subdivided into ‘batches’ of 100 items, and these batches are further divided into sample ‘sets’ of 10 items – and all of these are sequentially produced parts from a single milling machine, for example. We can imagine that each item has one or more ‘tolerances’ that must conform to a specification. Let’s say the thickness measure of the milled part must be 1.5 inches, plus or minus 3 thousandths of an inch. Any part that exceeds the tolerance limit is automatically rejected.

From each sample set of 10 items, we can conceive of a measurement or control process that requires that we randomly choose one of these 10 items and record in a control ledger that item’s precise measurement; 1.500 inches is the target, but each sampled item will be in the range of 1.497 to 1.503 inches.

The Western Electric rules for control consist of rules like:

  • Rule 1: Mean beyond 3 sigma
  • Rule 2: Two of three consecutive points beyond 2 sigma (same direction)
  • Rule 3: Four of five consecutive points beyond 1 sigma (same direction)
  • Rule 4: Nine consecutive points on same side of center
  • Rule 5: Six consecutive points increasing/decreasing

This type of evaluation – the monitoring of statistics relating to measures on a process – allows the process to be controlled. It is by continually assessing this statistical process that problems can be detected and adjustments made.

Leave a Reply

Your email address will not be published. Required fields are marked *