Predictive Maintenance after Niels Bohr:

It’s difficult to make predictions, especially about the future.

In this project we developed and implemented the control software of a blood analysis transport system. The blood samples are transported by small electric vehicles (cars) on a specially designed track (Carrera track) to the analysis machines. Every day, 20,000 blood samples from 1,000 cars are transported. Cars and tracks send more than 5 million status and position reports each day, resulting in 2 GB of log files that need to be processed and transformed into meaningful, usable information.

The batteries of the cars are charged at charging stations via a mechanical contact. The quality of this contact can be reduced by external factors.

Using the information from the log files, we were able to train a model that would allow us to optimize the timing of maintenance at the charging stations. The system combines a variety of big data technologies, but more about that later.

Considering the amount of data, we have chosen technologies like Hadoop, Spark, Python and R for processing, analysis and visualization.
In order to scale dynamically, the analysis was performed on a cluster in the Amazon AWS Cloud.

The predictive model identifies suspicious situations and reduces failures, raising the overall performance of the system to an unthinkable level.

Would Steadforce be able to implement this process in other projects?

The underlying process can easily be implemented in other industrial applications. The core competence of our Advanced Analytics team is to offer our customers a full service: starting with the raw data, continuing with the complete data analysis, and finishing with the productive software system.