Scania 4: Predicting Failures

This project is the fourth Tech For Future (TFF) project in a row that is being carried out at and with Scania with the aim of increasing production uptime. This research project is a collaboration between Saxion, Windesheim University of Applied Sciences and TFF. Saxion focuses on collecting and analyzing the data and Windesheim contributes its knowledge of mechanics and failure causes. In the previous TFF project (Scania 3), ‘Scania Production Line: from corrective to predictive maintenance’, a first step has been made towards collecting detailed information with sensors from the carrier during the tour of the production line.

With this data, the different detailed parameters of the carrier are recorded during the production runs and then analysed to gain insight into how the step from corrective (after the occurrence of a malfunction) maintenance to predictive maintenance (predictive) can be made. The start and results of this prototype are the motivation for Scania to place such data logging on all carriers.

This 4th Scania project focuses not only on reducing production stoppages due to carrier breakdowns, but also on answering the question “how can Scania predict failures for the production line in order to facilitate predictive maintenance?”.
The research directions are twofold and reinforce each other. First, it is investigated how (detail carrier) data can be used to predict failure and thus to reduce downtime (downtime). Secondly, the carriers are mechanically analyzed to detect failure causes. Combining both studies should lead to Scania's primary goal of increasing current uptime from 96% to 99%. As a tool, it is used to predict failures and to carry out preventive maintenance for this.

To this end, sensor technology is used to collect detailed data from the carriers. Data from, among other things, machines, log files, maintenance, environment, product, and ERP is also used to retrospectively analyze failures. In addition, if enough historical data has been collected, machine learning will be used to generate models that can predict failures. By linking the mechanical analyzes and failure to the data obtained, it is expected that the uptime of the production line can be improved through a systematic approach.

Program objectives

Where the previous project addressed the research question “how can the synergy of data analysis and the model-based approach of failure analyzes be applied for predictive maintenance?” had, the aim of this project is implementation-oriented to use the synergy between data analysis and model-based failure analysis for predictive maintenance. 

Partners

Scania Production Zwolle B.V., University of Applied Sciences, research group AmI (penvoerder), University of Applied Sciences Windesheim.

Duration

1-9-2021 tot 1-2-2023

More Information

For more information, please contact Jan Veltman.

Financing

This project was made possible by Tech For Future: http://techforfuture.nl/