Locked lesson.
About this lesson
Exercise files
Download this lesson’s related exercise files.
Control Chart Design.docx60.7 KB Control Chart Design - Solution.docx
61.1 KB
Quick reference
Control Chart Design
The control chart is carefully designed to provide a visual representation of process stability.
When to use
When creating a control chart, the design methodology must be used so that the data is correct. And of course, when using control charts, the design of the control chart must be understood so that proper conclusions are drawn concerning process stability and process control.
Instructions
Control charts are carefully designed so that they reveal when a process is stable or unstable. They show the magnitude of the common cause variation which is always occurring in stable processes. They also indicate the introduction of special cause variation, which leads to process instability.
The figure below shows a control chart. The horizonal axis is time or the sequence of data points. This gives the chart a time-wise view of the process. The vertical axis is the magnitude of the data being charted. The mean or average value is plotted, which provides a sense of expected value. There are also two horizontal control limit lines plotted, which indicate the magnitude boundaries of the normal or common cause variation. These lines are called the upper control limit and the lower control limit. The plot of the data will reveal the introduction of special cause variation. This can be seen whenever data falls outside the control limits and in some cases when certain patterns develop in the data, even though it is inside the limits. These patterns will be discussed in other modules.
Hints & tips
- The method for calculating control limits varies depending upon the type of data and the chart. These will be discussed in detail in other modules.
- Statistical software can be programmed to automatically show when special causes exist. However, the rules are simply to apply so charts can be created with a simple software application like Excel and still be used to identify special cause variation.
- 00:04 Hi, I'm Ray Sheen.
- 00:06 It's time to dig into the nuts and bolts of statistical process control.
- 00:10 And I'll start that by looking at the aspects of the design of a control chart.
- 00:15 Control charting is at the heart of statistical process control
- 00:18 because the control chart is where the voice of the process is revealed.
- 00:23 Control charts are a tool or
- 00:24 technique that provides a statistical view of the process performance.
- 00:28 It is a time-based chart that plots the process results on each
- 00:32 successive iteration, and shows those results in relationship to some
- 00:36 statistical boundaries of the process.
- 00:39 So we see a time-based view of the process and
- 00:42 can quickly identify when something may have changed.
- 00:45 This means it's good for baseline control since we can see the process performance
- 00:50 with an established baseline then change the process and
- 00:53 see the impact on the baseline.
- 00:55 Control charts are used to manage the process in real-time.
- 00:59 In particular, control charts will pinpoint
- 01:01 when a process begins to display unstable or out of control behavior.
- 01:06 This helps identify the underlying causes of that change in behavior.
- 01:10 Because of the time dimension on control charts,
- 01:13 it's also easy to see when trends occur that change the process performance.
- 01:17 In addition, there are a set of rules in respect to the data that
- 01:21 it indicates the presence of special cause variation in the process.
- 01:25 In fact, when Shewhart developed the control chart technique, he was focused
- 01:29 on finding a way to easily recognize the presence of special cause variation.
- 01:34 So the corollary is that when there is no special cause variation,
- 01:37 the process is stable.
- 01:39 And the control chart provides a means to illustrate the stability of that process.
- 01:44 Let's look now at a control chart and the elements of its anatomy.
- 01:48 An obvious element of the control chart is that it is charting
- 01:51 the value of the process characteristics over time.
- 01:55 The horizontal axis can be actual clock time, like hours or days, or
- 01:59 it can be sequential units, unit one, two, three and so on.
- 02:03 In addition to the plot of the process data,
- 02:05 the control chart shows several statistical measures and variables.
- 02:09 There will be a line to the mean or center of the data.
- 02:12 There will also be a line called the upper control limit and
- 02:15 one called the lower control limit,
- 02:17 which is statistically derived boundaries of normal stable variation in the process.
- 02:22 We'll spend more time on other sections discussing how these are calculated.
- 02:26 There are a variety of rules associated with this plot that indicate
- 02:30 whether the only variation present is the normal random common cause variation or
- 02:34 if there is special cause variation present.
- 02:37 When using statistical software applications like Minitab,
- 02:40 these will be automatically identified.
- 02:42 When plotting the process results by hand,
- 02:45 or with a more basic software application like Excel, you will need to
- 02:48 regularly apply these rules to see if special causes have occurred.
- 02:53 The good news is that when there is no special cause present,
- 02:56 you have a stable predictable process that will consistently provide results
- 03:01 that are between the upper and lower control limits.
- 03:04 In fact, let me just go a bit deeper into a discussion of control limits on
- 03:07 a control chart.
- 03:09 A process is said to be stable and
- 03:11 in statistical control when there is only common cause variation.
- 03:16 And as you remember, common cause is random but within predictable boundaries.
- 03:21 A stable process has a predictable and stable mean, or average value.
- 03:25 It has a predictable and stable standard deviation or spread.
- 03:29 And while it is impossible to predict a precise magnitude or
- 03:32 value of the next data point,
- 03:34 that point is predictably between statistically determined boundaries.
- 03:39 A control chart and control limits are statistically designed so
- 03:43 as to discriminate between stable and unstable process performance.
- 03:47 They identify the presence or the introduction of special causes.
- 03:51 That is why a control chart is used to track and manage a process.
- 03:55 It tells the process operator or process manager whether a process is stable or
- 03:59 unstable.
- 04:00 And the process control limits let the system designers know the level
- 04:04 of variation they can expect from the process, and
- 04:06 that they must be prepared to accommodate in other parts of the system.
- 04:11 Control charts were designed to be a simple, visual technique for
- 04:15 determining if a process is stable or if there is special cause variation present.
- 04:20 When you understand them, they're easy to generate and use.
Lesson notes are only available for subscribers.
PMI, PMP, CAPM and PMBOK are registered marks of the Project Management Institute, Inc.