Achieve Process Control by using SPC!
Process control is about understanding your process, only then can you control it. To understand the process, you must understand variation.
Six Sigma use SPC (Statistical Process Control) to control processes. SPC is a tool to distinguish between normal and special caused variation.
Why should you distinguish between normal and special cause variation? Because special and normal cause variation requires different approach when identifying causes of variation.
If scrap at a manufacturing company arise from special cause variation, you can identify a specific reason that caused it. If scrap is caused by normal variation, it is more challenging to find the cause. It is not possible to identify a specific reason; the scrap is due to variation caused by the process natural behaviour. Several factors can contribute to the normal variation.
Let me illustrate with an example. A machine has too much normal variation and produce scrap. All machines have normal variation as a result of its design. By adjusting a machine that has undesirably normal variation you can risk making it worse. Adjusting a machine will only help if the problem is accuracy. If the problem is precision (too much normal variation) it will not help to adjust it. You may choose to invest in a machine that is more precise. Another option is to include the expected quality cost in the cost calculation of the product in order to take the right price for the product. I made a video with applications for statistical process control.
When it comes to the quality of the final product, it can be subject to variation from several factors that varies during the production process. If you want to improve product quality, you need to identify the critical factors of your process. Then you can reduce and control the variation of these factors. I was part of the team who implemented online statistical process control - SPC at a production line at Axis-Shield (now Alere). Online SPC was available for all control parameters.
Before implementing SPC, consider the purpose. Here are some examples of applications:
- Do you want to determine the normal variation of a process in order to determine its robustness? Are the normal variation well within your customer requirements? Otherwise, you should work to make the process more robust, or include the cost of scrap in your calculations.
- Do you want to determine the normal variation in order to react to special causes?
- Do you want to connect the process behaviour to maintenance tasks? Wear on equipment will gradually increase normal variation. Alarms related to the standard deviation can alert you when parts need to be replaced.
- Do you want to assess the effect of a change you have done? If you made an improvement which will reduce the normal variation or change the average of a process, the effect can be determined using SPC. SPC chart before and after the change will show if the change had an effect.