How Data Analytics Tools Support the Five DMAIC Phases of Six Sigma

When it comes to continuous quality improvement and removing defects from a process, Six Sigma continues to be the gold standard in manufacturing and process management. This structured, data-driven methodology for discovering problems relies on rigorous analysis of production and process data. For many companies, engaging in a Six Sigma process can be time consuming or even a bit daunting.

One of the core pieces of a successful Six Sigma approach is conducting an effective root cause analysis based on data analytics. This sort of statistical approach to problem solving benefits from data analytics tools that make the process simpler.

^{_{This article is posted on our Science Snippets Blog.}}

The Umetrics Suite of Data Analytics tools can help optimize the process in three of the phases of DMAIC using the Six Sigma method for quality improvement.

In fact, Six Sigma is heavily reliant on data analytics. It gets its name from the idea of managing a process to stay within six standard deviations of the upper and lower limits of the mean (sigma is the statistical notion for a standard deviation).

One of the methods used for Six Sigma is called DMAIC, which stands for Define > Measure > Analyze > Improve > Control. Let’s take a look at how these five steps can be simplified using data analytics tools.

Using DMAIC to Correct Process Problems

DMAIC is the main Six Sigma methodology designed to uncover deviations and solve problems. It’s characterized by a strong customer focus (a focus on delivering the features or specifications the customer wants), and uses data and statistics to draw conclusions.

The DMAIC methodology is divided into five different phases. Each one has three steps.

1. Define. The first phase involves defining the problem and mapping it out.

The steps include:

Understand the problem from the customer’s perspective.
Describe the problem, set the goal (what you want to achieve) and determine the resources needed (for example, create a team charter).
Describe the process and verify with the stakeholders that you’re focusing on the right things. (Define a process map and verify).

The Define phase of DMAIC involves mapping out the problem from the customer’s perspective.

2. Measure. The second phase of DMAIC involves defining the project’s main metrics. How will it be measured? These steps include:

Describe the problem with numbers. What are your units of measurement?
Define the performance standards (determine the limits for Y). What is the critical parameter for the customer? Is this a discrete measurement?
Evaluate your measurement system. Ensure that the measurement data can be trusted. Can your measurement system validate whether or not your goal is met? (For example, if you want to improve yield from 90% to 95%, you can’t have a measurement system with a margin of error around 3%-units). It will be hard to see the difference between 90% and 95%).

3. Analyze. In the third phase of DMAIC, you look at data from your current process to see how stable or capable it is. Your operation or process must be able to produce results (products) that meet the specified requirements, so this is where you evaluate the standard deviation compared to the specification limits (tolerance limits). The steps are:

Establish process capability. Determine if your process is producing what is needed.
Define the project goals in numbers. (Define performance objectives, such as reducing amount of defects by 15% or reduce variability of process by 30%).
Identify variation sources. Look at the historical data from the process to determine what factors are influencing it. Is there anything in the historical data that can suggest what is causing the issue? (SIMCA, from the Umetrics Suite of Data Analytics Software, is a good tool for this phase).

Charts provided by SIMCA can help with the Analyze phase of DMAIC. The illustration relates to a loading plot showing which process parameters were influential in the past. Process parameters far away from the origin of the plot have been most influential in the past.

4. Improve. In the previous phase, you identified which of the factors have an impact on Y. In this phase you’ll investigate how changes in X will impact the result. (MODDE, the Umetrics Suite tool especially for Design of Experiments (DOE), is useful in this phase). The steps in this phase are:

Screen potential causes. (Sort out the important X variables.) Here you design experiments to test which of the possible X identified in the previous step influence Y.
Discover variable relationships. (Describe the relationship between X and Y). You go forward with a few big influencers and do response surface modeling.
Establish operating tolerances. (Which boundaries need X to hold Y within specifications?) Find out what operating conditions will affect the outcome. Establish operating tolerances by using tools like design space estimation, robust optimization and set point validation. Find out more about robust optimization in this webinar.

MODDE is useful during the Improve phase of DMAIC to manage Design of Experiments (DOE). It can also be used for robustness testing during the last phase. The main output of the robust optimization is a plot called design space plot. The design space (green) is a region in which there is a low risk of failing to comply with response specifications. Inside the design space, the robust optimum or robust setpoint is situated.

5. Control. The focus of this phase is to ensure that the performance objective you identified in the Improve phase is well implemented and maintained. (SIMCA-online from Umetrics Suite is useful for this phase).

Define and validate measurement system in the actual application. Ensure measuring system for important Xs is working.
Determine process capability. Here you will see if you met your goal. If your goal was to improve yield, you’ll verify that the yield now meets the new target you set.
Implement process control. Confirm that your changes can be sustained.

SIMCA-online provides the data needed to verify results and implement process Control during the last phase of DMAIC.

Tools to Help with Six Sigma Analysis

The Umetrics Suite of Data Analytics tools are uniquely designed to help you manage analytics and uncover the details in your data. As we’ve seen, the Umetrics Suite offers tools that can enhance the effectiveness of your the Six Sigma DMAIC projects:

For Analyze, SIMCA is very good at looking at the historical data.
For the Improve phase, MODDE helps you with screening, response surface modeling, and design space estimation.
And for the Control phase, SIMCA-online helps to monitor the process and make sure it behaves consistently.

In summary, the proven strength of DMAIC to solve your manufacturing issues can be enhanced by utilizing the tools and technics provided by the UMETRICS Suite.

Want to Know More?

Get a Free Trial