"I truly value the professional and interdisciplinary collaboration with Motius, especially their flat hierarchies. Over the years we have developed a very amicable work mode."
Our Final Result
Our Final Result
On average, the identification time of machine parts was reduced by 30%. Furthermore, the reliable classification of expensive precision machine parts allows their reuse, reducing the costs for spare parts.
Proven 30% Faster Identification of Engine Components
- Targeted process duration reduction of 50% can be achieved with further code and interface development
- We improved the algorithm, partially achieving 95% top-1 accuracy, and scaled the software.
- Less new spare parts are needed, due to reliable component classification
- MTU patent application (WO2022/237930A1)
A solid business case and the successful implementation in production has led to a long-term collaboration between MTU and Motius. Currently we are exploring the adaptation of the use case in other MTU business units and developing further features.
Looking for an Contact Us
During maintenance, the entire aircraft engine is disassembled, and all essential components are recorded. To shorten this time-consuming process and free up valuable time for skilled workers, MTU aimed to (partially) automate the process.
Motius developed a computer vision solution combined with a box containing multiple cameras to ensure optimal lighting conditions. A dashboard presents the user with a sample image of the recognized machine part.
All About MTU Aero Engines
Germany’s Leading Engine Manufacturer
MTU Aero Engines is an expert in the development, manufacturing, and maintenance of civil and military engines across all thrust and performance categories, as well as stationary industrial gas turbines. Their innovative propulsion systems, high-tech solutions, and comprehensive services enhance aviation efficiency, safety, and sustainability. In the field of civil maintenance, MTU Maintenance ranks among the top 3 global providers for civil aviation propulsion systems and industrial gas turbines.
Tackling the Challenge in a Four-Step Process
POC & Prototype
To see if the concept is feasible, we started with a simple, wooden PoC, with only 50 components, 4 cameras, and a scale. The performance was around 70%, which showed that the concept works, but needed more improvement. We came up with a more elaborate prototype that had better hardware and implemented faster software for quicker detection results.
For the MVP we increased the size of the detection box, so it can be used with bigger components as well. The design stayed similar, so it is compatible with the older, smaller prototype.
To ensure user acceptance of the released MVP, we conducted UX interviews and integrated the feedback. This resulted in faster software, better cameras, and a new distance sensor that can differentiate in the nanometer range. For better usability we centralized the data storage and minimized the cold-start issue by using the same ML model on each device.
Rollout and Optimization
As the MVP has been improved and significantly shortened maintenance time, the product was rolled out to all MTU locations. For better scalability, we further optimized up-time and reliability.