Title: Predictive Maintenance for Continously Variable Tractor Transmissions

Project Partners: AGCO GmbH, Professorship of Agrimechatronics TU Munich

In modern agriculture, tractors are indispensable machines for efficiently performing a wide range of tasks.
Their availability, reliability, and cost-effectiveness are decisive factors that not only influence farmers' efficiency, but also have a lasting impact on the responsible use of resources.
In this project, the transmission is considered a central topic due to its key role in the performance and durability of the tractor.
As part of this project, modeling approaches will be developed and compared to enable predictions for individual transmission units.
Specifically, the focus will be on three areas :

• (i) the remaining useful life (RUL) of the transmission oil filter,
• (ii) the RUL of the transmission oil, and
• (iii) the RUL of the entire transmission.

To create the prediction models, advanced sensors are used, which are installed both in real tractors and on the transmission test bench.
These sensors provide valuable data that is processed using innovative modeling approaches such as machine learning and digital twins to enable accurate predictions about the condition of the transmissions and their components.
The result is a sound basis for predictive maintenance that optimizes operations, reduces downtime, and at the same time contributes to more sustainable use of resources in agriculture.

Acknowledgments