Manufacturers love to hear the word “in control” when talking about processes. From a Statistical Process Control (SPC) point of view, an in-control process means that a process is stable or predictable. After putting in the work to get a process in control, how do you make sure it stays in control?
Every process has variation. While some sources of variation may be known and considered minor; others, if deemed critical, must be detected and removed in order to maintain a stable process.
Dr. Shewhart/Deming identified the following two sources of process variation:
- Common Cause – Variation that is inherent as part of the process. Examples of Common Cause: natural wear and tear, changes in humidity, old machines, etc.
- Special Cause – Variation that is outside of the normal process. Example of Special Cause: operator error, broken part, power outage, etc.
SPC provides statistical methods to observe the performance of a process in order to predict, identify, and remove sources of variation. Below are several methodologies that can be applied to help maintain a stable process:
- Real-Time Data – Data collection in real-time provides early detection. Immediate corrective action can be taken to minimize making bad products.
- Control Charts – Control charts provide process performance relative to specified control limits and, therefore, can differentiate between common cause and special cause of variation.
- SPC Control Rules – When a process triggers a control rule, it is detecting an “out of control” or non-random condition. Depending where the data lies in the control chart, further investigation will be warranted.
- Corrective Actions – Methods for eliminating a source of variation may include proper training, well-defined process standards, and developing a robust process through process refinement.
Variation is present in all things. The challenge is to identify what is and is not natural variation and then create an action plan to eliminate the variation. The approach listed above will help to maintain a historically-established level of variation.
It’s not time to uncork a bottle of champagne just yet, first, a little reminder. Don’t confuse control limits with specification limits, which represent the desired final product. Just because a process is in-control, it does not always mean that the process is “good”. In other words, it is possible to have a process that is in a state of statistical control but producing bad, or out-of-specification, parts.
What if you’ve got a process that is in control and producing products that are well within specification limits? Does continuous improvement mean that process should be improved at all costs? This is where the economics of SPC come into play. If you are running a process in control with a high capability, it probably isn’t worth the time and cost necessary to improve that process. Moreover, you may even consider reducing your sampling frequency and focus your efforts on another process that is struggling.
An in-control process simply implies that a process has performed to a degree of stability in the past and that stability is expected to continue going forward. If the process is producing good products, this means you can expect the process to continue to do so. The same is true if the process is producing bad products. Having a process that will allow for a predictable outcome of saleable goods and services is the first step. Making that process perform within desired specification is the ultimate goal. So, identify process variation, use Manufacturing Intelligence to make strategic decision, and put yourself “in control.”
For more information on how to get processes in control you can contact an InfinityQS expert.