Six Sigma primarily focuses on preventing defects. Unfortunately, approaching this problem for many organizations can be incredibly daunting. Take this time to explore the different methods for quality improvement that exist to make defect prevention significantly easier.
To get that process started, this article will discuss the ins and outs of control charts, an excellent visualization tool for quality improvement.
Control charts help operators and management get on the same page in terms of process improvement. Consisting of data points as well as both lower and upper control limits, the chart users can plot data as it becomes available and pinpoint when a process may be having issues. Not only does it work as a prediction tool to prevent out of control processes, but it also serves as a visualization tool for future improvement.
There are two primary types of variation that organizations must be able to distinguish from each other to be able to identify areas requiring improvement successfully. The two different types of variation are described as follows:
- Common Cause Variation – This type of variation is nothing to worry about. It’s often referred to as “noise.” In fact, a controlled process that sees some variation is expected. It comes with normal operation, the design of the process, maintenance activities, and even slight differences in the materials being used.
- Special Cause Variation – This is indicative of something unlikely caused by random common cause variation. Rather than “noise” special cause variation is often referred to as a “signal” that must be investigated. When a data point falls outside of the normal control limits, which denotes the acceptability of some variation, it can mean something is wrong and needs to be examined for improvements.
It also must be noted that common cause variation is always random. There should never be a pattern with data points, even if all the data points are within the upper and lower control limit.
What are Process Control Charts?
Known as process behavior charts, control charts, and statistical process control charts, this Six Sigma process visualization tool utilizes data to plot outputs in a time-ordered sequence.
The key idea to remember with control charts is that there will always be variation within a process. Problems begin to appear if every little variation that occurs insights a change from operators or management. Those little “adjustments” can be detrimental to a manufacturing sequence, causing even more drastic changes to the objectives than previous instances. This statistical theory was proven with Deming’s funnel experiment.
Prevent unnecessary tampering by thoroughly going over data. By taking the time to methodically make decisions via visualizing a process through control charts, the organization will be able to make more informed decisions. This also allows the users to set up a plan for re-analyzing the process after any changes were made to eventually confirm if the new process was a success or a failure.
The Three Elements of a Statistical Process Control Chart
Made up of three primary elements, control charts can show how a process changes over time. Those three elements are as follows:
- A time series graph with data points that have been collected at specific intervals.
- A horizontal median line to be able to visualize trends and any variations within the process.
- Two more horizontal lines above and below the median line represent the calculated upper and lower control limits. These lines show when the level of variation becomes out of control.
Then the user must calculate both the UCL and LCL to be able to monitor when the system is producing too much variation.
The Upper and Lower Control Limits
Both the upper and lower control limits, also known as UCL and LCL, are essential for control charts. Control limits are based on gathered data and used to monitor expected variation. Calculating them isn’t the easiest task, but it is doable. Otherwise, many people also use pre-made LCL and UCL calculators to help them along.
Let’s go over the math.
The formula to find the upper control limit is as follows:
UCL = x - (-L * σ)
The formula to find the lower control limit is as follows:
LCL = x - (L * σ)
X – This is the control mean. It is the average of all the numbers you’ve gathered.
σ – This is the sigma symbol. It represents the standard deviation from the mean. Essentially, it is a measurement of variability in a dataset. This is probably the most difficult component to solve as each type of control chart has its own unique formula for calculating sigma/the standard deviation.
L – This represents the control limit you wish to examine. It is the number of sigma lines from the center control line, or the mean.
The Different Types of Control Charts
There are two categories of control charts that modern factories can choose from:
- Variable Charts – These types of charts are used to visualize the variation in measurement such as height, weight, length, and concentration.
- Attribute Charts – These types of charts are used to display information regarding defects (a non-conformity) and defectives (so many defects that it is unsuitable for use).
Both are used to categorize following seven unique control charts.
Under the variable chart category there are three different charts:
- The Xbar-R Chart – This is the most common type of Statistical Process Control chart. Operators can use the Xbar and range chart to monitor a variable’s data. This is particularly useful when the sample size is small on constant, as they must be collected at regular intervals.
- The IX-MR Chart – This chart is helpful in situations where gathering data on the regular is too expensive for the company to pursue.
- The Xbar-S Chart – This chart is for facilities that wish to show how much variation happens within the average value. Not only will this help facilities understand variation on product better, but it is used in a wide array of business types such as manufacturing and engineering.
Under the attribute chart category there are four different charts to choose from:
- P Chart – The proportions chart is used to assess the proportions of a non-conforming/defective item in a process. It utilizes binomial distribution to measure this.
- NP Chart – This chart is used to monitor the number of non-conforming/defective items in the measurement process.
- C Chart – This is the control chart that counts the number of defects in constant size units.
- U Chart – This chart measures the defects per unit chart. It counts the data where the sample is greater than one.
Keep in mind, the chart to pick depends on the data obtained and how you would prefer it to be presented. The math for calculating the standard deviation also differs with each.
Know the Four Process States
No manufacturing process ever stays the same. In fact, all processes always lean towards a state of chaos if maintenance and improvements are not provided regularly. The point of charting data points within a process control chart is to be able to spot problems before a disaster strikes, such as multiple defects or a machine breakdown.
Let’s go over the four states of data monitoring within process control charts:
- The Ideal State – The operator or floor manager will find that 100% of the specified goals and objectives are being met. This state is characterized by predictability.
- The Threshold State – While this state isn’t 100% accurate, it only exhibits occasional non-conformity in the manufacturing process.
- On the Brink of Chaos – The process is not predictable, but it has yet to produce defects. This state is often missed in observation since there are no defects to be found.
- Out of Control – This stage is often where something wrong is noticed. Defects are beginning to appear, and the process is completely unpredictable.
The goal is to not wait until the process is completely out of control. By plotting out points on a control chart, erratic behavior can be spotted before critical levels of defects and lack of regularity strike.
Remedying these problems before they get worse often makes working safer, and it saves the company money.
Making and Implementing a Process Control Chart
Now that the basics have been covered, it’s time for you to start building your own visualization strategy with control charts. Get started by following these six simple steps:
- Decide on a duration of time to measure data. This is typically noted on the X-axis of the control chart.
- Collect the necessary data and plot it on the control chart.
- Add a control line by finding the mean from the collected data points.
- Calculate both the UCL and LCL, then add these lines in your chart.
- Note any instances where the data falls outside of the established control limits. If this occurs, investigate the cause, and adjust your process to reduce these abnormalities.
- Continue to track the process.
How do Control Charts Improve Quality?
There are all kinds of benefits that come with utilizing process control charts. Those who use continuous improvement in terms of quality primarily use this system to take advantage of the following benefits:
- To effectively communicate the performance of a process whether it be in manufacturing, those who use stock trading algorithms, or any other business focused on Six Sigma and TQM.
- To establish a baseline for potential future improvements.
- To visualize performance metrics over a large span of time.
- To analyze how a change has altered the process performance.
- To predict future performance.
- The ability to decide what is noise and what are signals for change.
Overall, control charts enable organizations to get on the same page about a process in a simple and easy to understand way. Not only do these simple charts enable businesses to improve their output, but it can also alter how the workplace looks towards safety efforts. This is because unpredictable processes can be dangerous ones.