If you have ever been involved in manufacturing or quality-related conversations, you may have heard of "Statistical Process Control" or "SPC". And if you Google SPC you will find a bunch of ‘textbooky’ definitions which are likely going to make you run away and never think of it again.
But you shouldn’t.
SPC is a fancy way of saying that you should …
… listen to your machines (or processes) – so you can hear them cry if they are hurting.
This might sound a little too emotional and abstract - but it's true. SPC starts with creating a ‘language’ that allows your machines and processes to talk to you, and for you to understand what they are saying.
When it comes to SPC, some people will tell you it's all about making sure that your products are ‘within specifications’ or ‘meet tolerances.’ This is not the case. If you are making components that don’t meet requirements, then those components (for lack of a better analogy) are already dead. We want to do something well before then.
Which is why SPC is all about being able to tell if your machines or processes are hurting so we can do something about it before they ‘die.’
So what is the language that SPC uses? Statistics. A statistic is nothing more than a number that represents the characteristics of a bunch of things. And the idea is that we can pull some statistics from key variables of our machines or processes to essentially track how everything is going.
Think of a doctor doing the rounds in a hospital. They check a patient’s charts to see how the patient is recovering or dealing with a sickness or injury. The patient’s temperature is an example of a statistic. As is their blood pressure. In fact, SPC involves charts that look a lot like the charts you see at the feet of hospital beds.
So this SPC language that is based on statistics … what does it tell us? Change.
We are looking for any changes in the behaviours of our processes and machines because these usually come before catastrophic failures. So if we identify changes in our process or machine, we can prescribe some sort of remedy (just like doctors do). This is analogous to providing a patient with antibiotics to fight an infection as early as possible instead of waiting for the infection to cause gangrene and then amputating a leg.
So what does this language look like? Well, SPC involves a lot of different charts. Here is what we call an ‘X - R chart.’ Someone who is trained to ‘speak’ the language of SPC would notice that there is something going on in this process or machine.
This chart is based on a machine that manufactures the lids used for medicine bottles. The metric that is being measured is the ‘removal torque’ which represents how much effort it takes to unscrew the lid. This chart captures the initial ‘startup’ period of the machine before it settles into a more constant behaviour, creating lids with a much more consistent removal torque. We are OK with ‘change’ … as long as we know why. But charts like this can tell us if our machine is starting to deviate from its current state. And that is where we step in – well before all our lids are ‘dead.’
Are you interested in learning more? Do you have any stories to tell in this regard? We would love to hear from you.
コメント