2020: Behind the scenes – Zeppelin oil lab

We are very successful in selling and servicing Caterpillar construction machines and are proud Caterpillar dealer since 1954. So far, so clear. But did you know that Zeppelin also owns an oil laboratory, to conduct construction machine analysis? And that the Zeppelin oil lab is very often part of innovation processes to improve Caterpillar construction machines? In this blog post we want to take you on a behind the scenes tour through the Zeppelin oil laboratory in Garching, near Munich in Germany. What exactly happens in the oil lab on a daily basis? And what have data analytics to do with that? Read more below to become an expert in the field of oil laboratories combined with data analytics. And have fun scrolling!

Black gold – not only in value, but also data!

Oil samples from all around Europe are being sent to the oil lab after they have been taken from the customer’s machine. The samples then get a registration number and are being prepared for further processing and analysis for their contents with the help of various machines. The data gained from the analysis enables the technicians to evaluate the condition of the respective part of the machine.

During easy going days, there are “only” approximately 300 to 500 oil samples per day rushing through the oil laboratory. But when it’s busy, that number can rise up to 800 or even 1,000 samples per day! So the question is: How can a small team handle that big amount of samples in an easy and effective way, keeping high quality standards? The solution was: The creation of an algorithm that was developed by the Zeppelin Data Analytics & Fleet team in close cooperation with colleagues from the oil lab!

An algorithm that helps with oil analysis

Prior to sending out the final oil results to customers, the data is transferred to a database. Based on input there, technicians can evaluate the results and put it in perspective and comparison to other results. Zeppelin technicians have been evaluating these samples according to a traffic light principle for the past ten years. And exactly at this point, the newly created algorithm is of great help! With the collected data from the past ten years (which represents about 1,000,000 data sets) and some initial information, the development of the algorithm started in 2018. Basic rules and guidelines were defined to provide a framework for the algorithm to classify the samples according to the traffic light principle. That way, the technicians can see with one blink of an eye, in which category the sample belongs without taking too much time. The algorithm is currently used in the form of a second traffic light in the analysis dashboard to give a second opinion for the evaluation of oil samples. This way the algorithm now supports not only new colleagues and colleagues who are less used to interpreting the samples, but contributes to a consistently high quality in interpretation also with an increasing amount of samples.