Hi, guys. Welcome back. I'm Katherine McKeever, and this is your lean six sigma green belt. So today we're going to go over statistical process control. So before I jump into our objectives, I wanna per you know, my usual caveat out there
so much like very of constraints. Statistical process control can be done independently off your domestic or your PDC air. Your process improvement efforts.
It is not the realm of just lean and six Sigma. It is its own independent discipline that is very complimentary to our lean six Sigma. Objectives and goals specifically are six Sigma as we're looking a decreasing variation. So with that, we are learning this in our control phase
because it's really good point in her Damanik project to use it. However you because you listen to our introduction and you know that you're going to be a super savvy business person
can use this on any process independent of whether or not you have ever done a process improvement activity for it. So if you just want to keep your finger on the pulse of your, um, processes, statistical process, control is a great way to do it.
So with that switching back until lean six sigma and, um AIC. You have now finished your pilots because you did such a great job developing your solutions on we've got highlights and we're ready to go. And we are. We're doing
where we have done full implementation. So everybody who uses this process
is following our was future state now new and improved current state on. Do you have officially shifted into the control phase off your DM a EC project. So you're done with improved, and now you're in control. So today we're going to go over an introduction to statistical process control.
You're gonna understand the objective of us PC.
You're gonna understand the assumptions. There are some key assumptions for this to work that I want you to be comfortable with because you're probably going to think about him as you go through using this on. And then we're gonna go over really, really high level components of the control chart literally were going over the world's simplest control chart,
and then we'll do more of it later on.
Statistical Process Control, or SPC, is an applied statistics technique that is used for monitoring and controlling processes. What that means is, if you remember early on, when we said I was telling you about descriptive statistics and inferential statistics and they said, Someday we're going to get to the place where statistics we're gonna help you make your decisions
today is that day. You're going to be able
to use statistics to make some determinations about your process and both capability and stability. So there are two important things that come from Spc. It's understanding your process capability, which hopefully this is ringing a bell and sound a little bit like CPK and PPK
from our voice of the process, lecture
and stability, which doesn't ring a bell yet other than I think I've sprinkled it in like fairy dust throughout the class. But we're going to go into it in depth in the next couple of lectures, and then we're gonna go. So the last thing about Spc. Is it uses control charts to monitor process.
So if you if you think through this course, everything here can be done either mathematically
or visually, and we like visual because it's more intuitive. So think about our spc or excuse me, think about R, C, P K and R S P. C so we can calculate CPK. And we can say these are good like we're better than one or were worse than one. Or we can view it
using a control chart under Spc. So you have some options,
but we like it. It's very intuitive as we go through it. I think that you will learn to love it as much as I do. Hopefully. All right, So the first thing off I promised you is we have to have some assumptions. These must be in place for you to use statistical process control, and some of them are esoteric and some of them are not.
The 1st 1 is. All processes have some form of variation,
and this is true All processes, regardless of how refined and how maney improvement and how much streamlining you've done on this, your process will have some form of variations. So think about your fully automated robotic processes. So, um, I live in Colorado. We have a lot of
canning and bottling factories for a lot of people's favorite beverages.
Beer. Let's just call it. There's a lot of beer here, and I've gone through quite a few of their manufacturing planes and then one of the things that it's always noticed mazes. You have this entirely robotic process. One of the brewery's here actually only has, like, three employees on their factory floor as Q A and then the operators for the
the machines. But so you have this entirely robotic process.
You still have variation because you will still have down time. You will still have, you know, a clog or a broken bottle or something like that. So even though it is entirely automated and it is like the pinnacle of greatness, you still have some variation,
even if it's DME diet and been insolent. There are two types of variation. There is common cause, and there was special cause. We're gonna talk a lot about variation of the next few lectures, but I'll give you kind of a spoiler. Now. Common cause is normal. This is that we it always kind of happens. We know it's there. It's okay. It's your ambient level of variation
this is non normal. This is what you want to do something about. It's a special case. The next thing up is that you're process. Stability has to be able to be tracked. We talked a little bit about the python distribution. Ah, few lectures ago.
If you have a process where your stability cannot be tracked,
you cannot use statistical process control. So, textbook example poison distribution, hurricane distribution. If you can't say when the next time you're gonna have in occurrences, you are not going to be using statistical process control on then. The last assumption is a faith in you guys
that you, as practitioners, will be able to tell if your variation is special or common, cause
you're going to do that by using shoe hurts. Rules were have a whole module on this coming up in a few lectures, but this is empowering us. The practitioner. You're gonna be the one who raises your hand and says, Yep, this is OK. This is common cause or no, this is not special, cause
or no, this is special cars, and we got to do something about it. So with that, the last assumption is that you was the practitioner will be in a place where you could make those determinations.
this is a control chart, guys, this is statistical process control in like the world's simplest version. A couple of things that you're gonna look at. We have our data points. So a control chart is different than, um a any of the other charts that we've looked at because your ex axes is time.
They are also called Times Studies. Your Y axes are going to be the measures
of either your sample or your um, well, I guess you'll always be a sample. It's just whether or not your sample sizes one or more will be in the measure of your sampling for your process outcomes. So in the middle blue dots, these air going to be your actual measurements, and they're based on time.
So from left to right, we're going from older to newer.
With that, you're always gonna have an upper control limit and the lower control limits. So these were going to be your customer specifications. You can actually determine control limits two ways. One customer requirements, which is my preference
into your upper control limit, can also be three standard deviations from the mean. So if your customer doesn't give you a requirement
which all customers have requirements, so don't believe them. Um,
you can do three standard deviations above, above or below the means. One of the examples that I'm showing you later on has a little bit of variation on it. I'd like you to use that because it's more more specific. The last thing up is, is the started line in the middle of this is gonna be your average, so this is going to be where the middle of your process performs.
This is important because it tells us
how close or far away from our Upper Control LTD. Or lower control limit is. That's the difference in the K or than normalising factors in our process. Capability on sore CPK and PPK. If your average line is not centered,
you are using C, P or PP formulas
to determine your process capability. But so this is a control chart. In its simplest form, we have whole modules on how to get into this in more depth. But what this tells us is we use our descriptive statistics and eso. Both are
measures, measures of central tendency and our measures of dispersion, and it now gives us an idea of what to dio.
So with that today we introduced statistical process control you guys know that this is an applied statistics tool that's gonna help us make decisions about our processes. You understand the objectives and the assumptions. Of course, the objective is to monitor and control our process. The assumptions are
that all processes have variation, etcetera, etcetera,
and you can recognize the major components of a control chart. If you see a chart with four lines three straight and one jig it you're probably looking at a control chart or a run chart. With that, we're actually gonna talk quite a bit more about process, stability and our next module. So I will see you guys there.