The 5 steps of Lean to optimize production processes with IIoT

In this article, we share a strategy for manufacturing process optimization with Industrial Internet of Things (IIoT). We do this based on the 5 steps of Lean with the DMAIC cycle.

Start small with process optimization

Machine data is crucial to gain insight into and improve production processes. Much has already been published about this in professional literature. If you start to improve production processes based on data, start small. Choose a production line or a part of it. Realize that more is needed than simply the technical aspect of collecting and presenting the data. Employees also need to gain experience to understand the data in order to implement improvements based on it.

The Lean DMAIC method to improve production processes

The DMAIC cycle (Define – Measure – Analyze – Improve – Control) is a proven Lean methodology to achieve improvements in production processes. Let’s take a look at how these five steps are used in the optimization process.

#1: Define – which production line to improve?

In the Define step we determine in which production line we want to realize improvements. This can be done in two ways. Firstly, based on which production line causes the greatest problems. Or secondly, based on the simplicity of the production line process. Which production line data is the easiest to interpret?

After this decision has been made, there are three steps in the Define phase:

A. Mapping the production process
This is a simple diagram showing the steps of the production process. Often this already exists. Does it need to be made? Then involve the people involved in the production process in drawing it accurately.

B. Determining the optimization goal
Once the production process has been mapped out, it is time to set goals. What do you want to improve? Some examples of improvements could be:

• Reduce the number of short stops by 20%
• Increase Overall Equipment Effectiveness (OEE) by 5%
• Reduce line energy consumption by 10%

The next step is to determine which machine data is needed to gain insight into the problem. For example, if the goal is to increase the OEE of a production line by 5%, then we need to know what the current OEE percentage is.

We map out which sensors or data points of the production line define the performance, which data points define availability and which data points define quality. It is important that we avoid manual collection of data.

C. Connecting production line to the innius platform
The third step is to connect the production line to innius. Only the relevant data points from the production line are sent to the innius platform. It is important to keep the dataset small, aimed at the defined goal. This is to keep it clear and to prevent an overload of data.

#2: Measure – real-time production line monitoring

In this Measure phase, we will monitor the production line in real-time with innius based on the connected data points. As soon as the production line is connected, there is immediate insight into the state of the production line. The first measurement provides a baseline. From here we get started to achieve the set goal. For example: if the goal is to increase the OEE by 5%, the OEE percentage of the first measurement reflects the current state of the production line. After innius has received data from the production line for 14 days, we make the step to the Analyze phase.

#3: Analyze – where to improve the production line?

In the Analyze phase, we will analyze the collected data more deeply. Two weeks of data provides a lot of insight. We look at how the production line behaves to determine where improvements are needed to achieve the set goal.

Pareto analysis
Back to our example where the goal is to increase OEE by 5%. There we analyze the data where we see deviations in the performance of the machine. A simple line graph, which compares the maximum performance against the actual performance, provides direct insight into where there is a potential bottleneck in the process. A pareto analysis (80% of the problems can be traced back to a few problems ~20%) provides a simple insight into which ‘stops’ have occurred and how long they have lasted. From this we can conclude which stops we need to solve first to increase the availability of the product line. We also analyze how the quality produced was based on the data. This analysis provides input for the next step.

Identify problems
Based on the analysis of the data from the previous step, we identify the problems, and determine which problem we will solve. Normally, this is the problem that takes the most value from the stated goal. In our example, the low availability of the production line could be the problem of the low OEE. Solving this contributes the most to the goal of increasing OEE by 5%. We then focus on the problem that emerges from the pareto analysis. In other words, the problem that causes the longest downtime. Now that we know this, we can move on to the Improve phase.

#4: Improve – make improvement proposals

In the Improve phase, improvement proposals are generated and planned. Together with those involved in the production process, it is discussed which improvements are possible to achieve the set goal. In our example, the team (including the maintenance team) discusses what can be improved to avoid the stop that shuts down the production line for the longest. If several improvement proposals have been made, the best option is determined to implement the improvement. Subsequently, a schedule is drawn up to implement the selected improvement on the production line.

#5: Control – did the adjustment have the desired result?

In this step, measurements are used to check whether the implemented improvement has had the desired result. The chosen solution may require an adjustment in the production process. All production employees must of course be aware of this and be trained where necessary.

In our example, it may be that the production line has been idle the longest because the supply of raw materials is not on time and conducted correctly. It is not a technical problem but an organizational problem. Those involved in this part of the process must be informed and trained in the adapted production process.

When this is done, comes the step of measuring the outcome of the improvement made. In our example, we regularly check whether the OEE percentage is increasing. This is also done with the pareto analysis report, to see if the downtime is decreasing.

Conclusion – Learning from Lean

We have now taken the first step to improve the productivity of a production line with IIoT data. The five steps of the Lean DMAIC cycle can easily be repeated on the same production line. Until there is satisfaction with the productivity of the line. Now that experience has been gained with data management, the next production line can be linked to innius. The DMAIC cycle can also be completed for that production line.

Finally, we would like to mention several pitfalls that must be avoided in this process. These are:

• Wanting to tackle a too big project; Lean emphasizes a step by step approach.
• Wanting to collect too much data; limit yourself to relevant data.
• Working in an isolated manner; involve the whole team in the process.