Kienle + Spiess is a German mechanical engineering company that was looking to strengthen its welding division by reducing its rework and improving its quality. To accomplish this, the organization utilized a dual approach where Six Sigma and Lean methods work in tandem.

Kienle + Spiess was able to practically eliminate shoddy splatter from the welding of its iron core product by combining tools from both Lean and Six Sigma methodologies. From Lean, the organization used tools like Big Picture Mapping and Value Stream Mapping, while it took tools like DMAIC and Design of Experiments from Six Sigma.

Kienle + Spiess Had a Problem

Kienle + Spiess is a mechanical engineering and manufacturing company based in Germany, with a second location in Hungary. The organization started in 1935 and specializes in providing its customers with a range of products, services, and technical solutions for the creation of energy-saving, efficient generators, and motors.

One of Kienle + Spiess’s products utilizes an iron core welded into a T-shape and is part of a larger transformer. A few years ago, it came to the attention of the company that way too many of the welds on these products were shoddy, leading to expensive reworking. In several cases, the welding was unsatisfactory enough that the part would be beyond reworking and dumped.

In order to maintain the value of its product in the eyes of customers and to keep the spiraling costs of reworking and replacement under control, the organization would need to address this welding issue.

The Company Opted to Alternate Between Lean and Six Sigma Toolkits While Addressing the Issue

When the team got together to address quality in its welding division, the first thing done was to create a big picture map and a value stream map. These are tools that are closely associated with the Lean methodology. Lean focuses on what is bringing value to the customer. Anything that falls outside of that is waste and eliminated from the process. Cutting down on waste in a process winds up not only providing more value to the customer, but also serves to improve efficiency, quality, and minimize costs.

Creating the Big Picture Map took about four days. This may seem like a long time, but it is worth noting that an organization might create one of these types of maps every three years. The reason for this is that this type of map will represent all of the projects in an organization over that long of a time period.

In looking at the map the team created, a rework loop stuck out like a sore thumb. It was obvious that there was a quality problem that was causing havoc in this particular value stream.

In looking at the welding of the product, it became clear that the issue was a technical problem. The Big Picture Map from the Lean toolkit had gotten the team this far, but once it became clear that the issue was a technical problem, it was time to switch toolkits. It was time to switch over to Six Sigma.

While Lean focuses on what provides value to the customer and eliminates the waste that falls outside of that, Six Sigma aims to reduce defects by identifying problems that have unknown causes, finding the root causes, and then solving how to best make improvements that can sustain.

In looking at the welds, it was clear that defects were occurring and that Six Sigma could help get to the bottom of them.

While the first go-to tools in Lean are Big Picture Mapping and Value Stream Mapping, the first line of defense in Six Sigma is DMAIC. DMAIC is an acronym for a problem-solving process that is fundamental to the Six Sigma method. This process has five sequential phases. These are Define, Measure, Analyze, Improve, and Control.

Here is what happens during each of these phases:

Define – During the first phase of the DMAIC process, it is defined exactly what the problem is that is going to be addressed. This is not always obvious, as multiple problems may exist, and it can take some work among a team to narrow down which issue is most important to tackle during a given time period.

Measure – In the second phase of the process, a team looks at how the current process performs and takes relevant measurements to establish a baseline from which to work. The data pulled that establishes this baseline provides a picture of the “as-is” state of the current process.

Analyze – In this third phase of DMAIC, the team carefully analyzes the data available in order to determine the root cause of the problem. This is a vital step to take before making any improvements. Improvements made before addressing the actual cause of the problem can wind up being a waste of resources.

Improve – This fourth phase of DMAIC is where the actual improvements happen.

Control – In this final phase of the process, the team ensures that measures are in place to maintain the gains that occurred during the Improve stage. The reason for this is that improvements are hardly worth much if the work isn’t done to prevent things from sliding backward again. The Control phase of DMAIC helps to ensure that the root cause that created the problem does not happen in the same way again.

For the team at Kienle + Spiess, the Define phase was simple enough; there was an obvious issue with the quality of the welding. For the Measure phase, the team looked at the amount of splatter that was present on the defective welds. These splatter markings provided a key indicator of the faultiness of the welds. It was decided that the splatter would serve as a measurement to work off of and improve upon. Counting the splatter markings on welds would provide a suitable baseline for where the current state of the process was and what needed to improve. The team then needed to analyze what was causing the markings.

The team looked at the inputs to the process. Inputs to the process can include methods, materials, human factors, machine factors, materials, and measuring systems. In this particular case, the team identified welding parameters and feed rates as machine inputs that could be important for the splatter issue. From input methods, it was determined that the mixture of gases could be relevant. Since these parts are pressed, the team decided that the level of oil pollution during the pressing should be an input tested, as it could be a factor in the bad welding. All in all, it was determined that nine variables were of interest in this particular problem.

Looking at the number of variables that were involved, the sampling needed would be at least 270 pieces to conduct a full sampling analysis. This meant potentially having to create that number of pieces that would likely have to be trashed. This did not seem realistic, so the team opted to do a designed experiment.

A Design of Experiment is the fastest way to statistically determine the methods of improving a process while minimizing the risk of wasting your effort. Instead of 270 tests for a full sampling, the team could find out enough information through analysis to make improvements in just 12 tests.

This was done by testing all nine variables at once during each of these 12 tests, utilizing two pieces per test. This made it so that only 24 pieces would potentially have to be scrapped, a much more cost-effective testing method.

In looking at the test results on a Y-Hat Marginal Means Plot, it was easy to spot that there were key variables that had a massive impact on the amount of splatter on the welds. One of these was the gas type. One gas type was clearly creating less splatter than another. By noting this and other key attributes, the team was able to determine the changes they could make to their settings to minimize the amount of splatter on the welds, which would indicate that the welding quality had been improved.

Noting the improvements that occurred from choosing the conditions that produced the least amount of splatter, the team then put measures in place for the Control phase of the DMAIC process. They created visual signals on their settings in order to safeguard the improvements that they had been able to make on the amount of splatter. Visual signals are a part of visual management, which is covered in Total Productive Maintenance. TPM is a tool that is associated with Lean. In this part of the team’s process, instead of alternating between the two methods, Lean and Six Sigma worked hand-in-hand.

The Outcome Was Astounding

By using the steps outlined above, the team was able to completely eliminate the splatter that was present on the product. This meant that the welds had improved significantly. In dialing in the settings that managed to eliminate the splatter and using those settings for all future welds, the team was able to control the improvements and consistently offer its customers a product of better quality, while avoiding the waste that came with having to scrap defective merchandise.

3 Best Practices When Utilizing Lean and Six Sigma Tools Together

The team at Kienle + Spiess found that utilizing tools from both Lean and Six Sigma methodologies worked best in tackling its particular problem. At some points during the process, alternating between each methodology’s tools made the most sense, while at other points, using tools from both methodologies at the same time proved beneficial. Here are some practices to keep in mind when using both methodologies together to handle an issue in your organization:

1. Understand the strengths of each methodology

The team understood the strengths of both Lean and Six Sigma and at what point along the process of addressing its problem the tools from each methodology could do the most benefit. Having an understanding of each method and what can be pulled from each can be a massive benefit in knowing how to approach a problem within your organization.

2. Know when to look at the other method

During the process of fixing the welding issue, the team was confronted with a large number of variables and cost-prohibitive testing that would have to be done in order to determine the root causes of the issue as well as understand the improvements that would need to be made. By switching toolkits, the team was able to utilize the Design of Experiment tool, which cut down on the number of tests it would need to do to a manageable amount. If you are working with either Lean or Six Sigma and you find yourself stuck or faced with an issue that seems insurmountable, looking to the other methodology can provide you with a path of less resistance.

3. Not every issue falls into simply a Lean or Six Sigma problem

The team at Kienle + Spiess found itself with what turned out to be a rather complicated problem once it began to take a closer look at it. There are some types of issues that are best handled by Lean, while others are ideally suited for Six Sigma. Not every single issue is going to fall into the category of one or the other. In your organization, there could be some issues that could benefit from alternating between the two methods or even working with both in tandem.

Lean Six Sigma

Utilizing tools from both Lean and Six Sigma methodologies, the team at Kienle + Spiess worked with what is known as Lean Six Sigma. This way of approaching a problem is worth looking at when there are complicated issues that do not fall squarely into the categories of issues that either methodology is meant to address alone.

About the Author