Students are introduced to ‘statistical process control (SPC)’ and ‘process capability analysis (PCA)’ in terms of how they can help you be ‘in control’ of your manufacturing (or other) process and then know how to keep improving it. SPC is particularly useful at creating ‘early warnings’ or ‘red flags’ that a manufacturing process is starting to come out of control before defective parts or components are made. PCA is all about taking a process that is ‘in control’ and improving further, amongst other things, to help get more time from SPC warnings to address a potential problem.

Students are taught the immediate and long-term benefits of SPC and PCA and how these benefits generate individual and organizational value. Initiatives that try to introduce SPC and PCA tend to fail when those implementing them aren’t fundamentally motivated by these activities.

‘Processes’ are not machines or systems but interrelated activities that turn something (like raw material) into something else (like a finished product). Students will learn what a process is by looking at how bread is made in a commercial bakery.

Processes need to be visualized to help us understand what they are and what we need to do to make them ‘high quality.’ Students will learn how ‘process flow charts’ do this and how to measure processes to understand if and when they are ‘high quality.’

We need to be ‘in control’ of processes, from driving a car to providing healthcare to a patient in a hospital. When it comes to manufacturing, being ‘in control’ means the manufacturing process responds to our attempts to influence it by (for example) turning up temperatures, adjusting the speed, and changing the proportions of raw materials. Students will learn what a manufacturing process that is ‘in control’ is and how ‘control limits’ help us make that determination.

A ‘random process’ is any process that involves uncertainty. Even the highest-quality manufacturing processes are random in that there will always be some sort of variation in component and part characteristics. Students will learn about basic ‘random’ statistics such as ‘samples,’ ‘averages’ and the ‘spread’ of random variables.

Every random process has a ‘signature.’ This applies to manufacturing, where some variation in part dimensions will occur. Understanding this signature helps us understand how a manufacturing process that is ‘in control’ behaves. Students will learn how to use the Acuitas SPC-PCA Application,which is based on Microsoft ® Excel ® to uncover these signatures to help understand what the process is doing.

One of the most widely used ways to visualize a process’s randomness is with the ‘X-bar R chart’ that tracks things like the slight changes in part dimensions and characteristics to see if there is any warning of a process changing or coming ‘out of control.’ Students will learn the statistical underpinnings of the ‘X-bar R chart’ along with how they can help easily visualize a manufacturing process.

Students will learn how to create an ‘X-bar R chart’ using the Acuitas SPC-PCA Application to identify how ‘in control’ a process is. Part of this worked exercise involves truly understanding the concept of ‘X-bars,’ which are the average characteristic measurement of a small sample of manufactured parts or components.

Understanding that a process is not ‘in control’ is one thing. Understanding what needs to happen to address it is another. In this lesson, students will learn howbread can be made ‘chewy’ to set the scene for subsequent lessons where the baking process needs to be improved as part of worked exercises.

‘X-bar R charts’ can help visualize manufacturing process performance and help you see if something is changing (in a bad way). However, this can be subjective and perhaps challenging to standardize. Students will learn about well-known rules that can help identify the slightest change in a manufacturing process that could signify future problems and typical ‘corrective actions (CAs)’ that can rectify them.

Students will learn how ‘X-bar R charts’ are developed using the Acuitas SPC-PCA Application, where any statistical signatures that trigger the rules above are automatically detected.

Manufacturing processes fundamentally change all the time. Suppliers of raw materials may change, or a new manufacturing machine might be installed. This will always change the random distribution of key part and component characteristics. Students will learn how this changes the statistical signatures of a process that is ‘in control,’ and how to accommodate these changes into ongoing monitoring.

While history lessons on engineering practices are not often necessary, there is merit in examining the original, paper-based, ‘X-bar R charts’ where assumptions were made to create control limits and other features. This is important because many software applications designed to automate the creation of ‘X-bar R charts’ simply use these assumptions to create approximate control limits, even though they are easy to calculate. Students will learn how ‘X-bar R charts’ have been used to help manufacturing processes throughout the 20th century and how contemporary software packages are not as advanced as they should be (note that the Acuitas SPC-PCA Application doesn’t make any assumptions when creating control limits).

The ‘R’ in ‘X-bar R charts’ stands for ‘range,’ which is the difference between the largest and smallest value of a sample or subgroup of random variables. It is a basic measure of how ‘spread out’ these values are. A better way to measure spread is with the ‘standard deviation,’ which is based on all random variable values, not just the smallest and largest ones. ‘X-bar S charts’ use the ‘standard deviation’ and not ‘range’ to characterize the spread in manufacturing characteristics. Students will learn about ‘X-bar S charts’ and how they can better characterize the random nature of a manufacturing process.

Students will learn how ‘X-bar S charts’ are developed using the Acuitas SPC-PCA Application, where any statistical signatures that trigger the rules above are automatically detected.

‘X-bar R’ and ‘X-bar S charts’ involve routinely sampling a consistent number of parts and components. This could be (for example) five loaves of bread every hour or seven steel balls every shift change. There are some scenarios where this is either impossible or not ideal. Some components (like crankshafts for very large, maritime diesel engines) might be manufactured at a rate of one a week, with substantial breaks in production. Conversely, the height of each loaf of bread might be automatically measured by a laser meter. Students will learn how the ‘individual moving range (IMR) chart’ can visualize manufacturing processes where every item is measured, with a worked example using the Acuitas SPC-PCA Application.

There is a branch of SPC that doesn’t involve measuring component characteristics (like bread chewability or steel ball diameter). Instead, they are based on counting the number of defects in a batch or region of a component to see if this number starts to change. Students will learn not to use these charts as they inherently monitorlow-quality processes (as in, you need to expect a substantial number of defective parts for these charts to work).

New manufacturing processes take time to master. Students will learn about the ‘initial process study’ where ‘critical to quality (CTQ) characteristics need to be chosen and studied until they are under control before setting control limits for ongoing SPC.

Once a process is ‘in control,’ it must be made ‘high quality.’ A ‘capable’ process is one where the natural variation in component characteristics is largely within allowable bounds called ‘specification limits.’ Students will learn about how to measure and characterize ‘process capability.’

‘Control limits’ are used in SPC to help identify any early signatures that processes are starting to come ‘out of control.’ But just because a process is or isn’t ‘in control,’ doesn’t mean the products are ‘in’ or ‘out of specifications.’ Students will learn the difference between the two, how one relates to keeping a process in control, and the other relates to capability and quality.

Measuring quality can be challenging. Students will learn about key metrics called ‘capability indices’ that quantify the extent to which natural process variation of key component characteristics falls within specification limits (Cp, Cpk, Cr, and ‘target Z’).

Of the four ‘capability indices’ covered in the previous lesson, only one relates directly to the number of defects. Students will learn the relationship between Cpk and ‘parts per million (PPM)’ defects.

Students will learn how to conduct PCA using the Acuitas SPC-PCA Application, which allows process quality to be quantified andthe expected defect rate in terms of PPM.

A customer can receive products that result from many different, albeit slightly different, processes. For example, some components might have been manufactured with various raw materials. In contrast, others might have been manufactured with an upgraded manufacturing machine. Students will learn about the capability index Ppk, which covers all the products that customers receive.

The data needed to conduct SPC and PCA comes down to the number of measurements in each sample and how each is measured. Students will learn about measuring accuracy and sample size selection.

Students will confirm their learning using the Acuitas SPC-PCA Applicationto practice both SPC and PCA.

A summary of information learned throughout the course.