From data to efficiency

How do you optimize the efficiency of a machine in practice? That question is worth a lot of money, so that’s why we’ve dug into the stack of projects and written a story about one of our clients.

For the sake of confidentiality, the story is anonymized, but it is real enough. We have used the exact data from the production and the simulations of investments that were used.

A starting point is a machine that produces different series sizes. Conversion in production costs money, so the project was to analyze production data and identify efficiencies.

Data was missing, which is why Opticloud was set up. There was an expectation that production could be optimized, but no basis for decision-making.

First, a data logger was mounted on the machine, and then data collection and analysis started. Along the way, service checks were carried out on the machine in order to improve the OEE.

With two service checks, the OEE was lifted from 21% to up to 41%.

Service checks were conducted in October and February. After both service checks, the OEE was lifted to another level.

It turned out that the machine’s productivity was closely linked to the series size in production. The machine can produce small and large items, and the many shifts were a core problem.

OEE could be lifted by reducing the number of series types and item numbers in production.

From April, the machine will start producing significantly larger series sizes. It makes a big difference in earnings that OEE rises from 30% to 41% in five months. The increase is from “under 150 topics per series” to “over 500 topics”.

How was the increase created?

An investment analysis was carried out on the basis of data. The analysis showed that it made sense to buy one more machine to produce small elements.

The small machine has no set-up but produces slower than the large one. The little one wins, on the other hand, when it comes to small series. After the investment, the factory has two efficient machines instead of one that is idle and in set-up mode too often.

The advantage of analyzing production data is that it is possible to view OEE for each series size. It is not enough to look at the machine as a whole. One machine can have a high OEE for some types of items, while others drag the average down. It is a common problem to lack concrete data at the level of detail. This can lead to wrong conclusions.

The red bar shows how many items a series size represents the total production. On the left are the main series, and on the far right there is the series “0000-0049”, where the OEE is low. This makes perfect sense because when you produce a very small series there are still conversion costs. They’re driving down productivity.

At the bottom of the blue bars, you can see the OEE for each series size. OEE is over 50% for pillars 1, 2, and 6. These are the three series sizes where we produce 500 items or more before we convert the machine.

The conclusion is easy:

When we produce large series sizes we have high OEE.

Therefore, the extra and the smaller machine will be a good business, but can it be calculated exactly what the effect will be?

Yes, you can.

Opticloud collects fairly detailed data about each machine in production. For example, all downtime is recorded in categories, so there is data on downtime due to maintenance, cleaning, logistics, conversion, human error, materials, environment, and several other categories.

Before investing in a new machine, OEE was 21%. After purchasing the small machine, the figure rose to 41%.

A little extra data

The number of shifts needed to produce 6,119 units fell from 3.0 shifts to 2.3 shifts. A shift in production amounts to 8 working hours. This is the equivalent of taking 18.5 hours to produce what used to take 24 hours. A saving of 5.5 hours a day.

There are two ways to calculate what the payback time for a new machine is – and the truth is probably in between the two methods.

One method we call fixed cost. Here we are counting on how quickly the new machine is written off with the current production. The total investment in the project including the new machine was DKK 3.5 million and the saving per year was DKK 750.000. This is a positive return on investment after 4.6 years.

The second method is earnings maximization. Instead of producing in less time, we use the time we get in surplus to produce more. Given that the extra goods can be sold, it’s a great idea.

In this case, the value of the additional production was DKK 2.96 million per year. If all the additional production is sold, it takes 1.2 years to achieve a positive return on investment. After that, it’s a profitable business.

The art is to find the easiest way to optimize production. The method changes from time to time, but each project starts with data collection. Stay focused on the small details.

Want to know more about how data can be used to optimize your production? We know a lot about that. Call or write to us. We always have time to help you.