e²LEAD

Enabling EU Leadership in automotive ECU for future Autonomous Driving

Supercomputing systems face tremendous challenges regarding the packaging technologies for communication and signal processing units, as well as heat removal and power delivery. This is particularly true for automotive supercomputer platforms designed for autonomous driving at Level 5, where high information interchange rates across multiple physical channels must lead to real-time driving decisions in the most trustworthy and reliable manner. e²LEAD focuses on advanced packaging to enable the development of supercomputing platforms 'Made in Europe' for the automotive industry, working across three research domains: Connectivity for high-performance data and thermal interfacing, Technology for automotive chiplet and smart power system (SPS) packaging and integration, and Reliability/Safety for new test strategies and digital twin-based design methods.

Project Facts

Project Lead
BOSCH

Countries

Partners

Costs

>9 Mio. €

Research Area

Transport and Smart Mobility

Duration

36 Months

Latest News

07/2024 e²LEAD - Enabling EU Leadership in automotive ECU for future Autonomous Driving – alignment meeting.

Partners of the e²LEAD project work collaboratively on the advanced packaging to enable supercomputing platforms for highly automated driving. On July 1st and 2nd we meet in Berlin at Fraunhofer Forum to discuss the progress of the current project, address any issues or roadblocks, and make decisions on next steps. The team members provided updates on their tasks and shared their insights on potential challenges and opportunities. During the meeting NANOTEST made a demo presentation of the Thermal Test Chip that we are using in the project. The meeting concluded with a clear plan of action and assigned responsibilities for the upcoming tasks. Our goal for this year is to prepare demonstrator. This project is funded by Germany and the Netherlands.

Thanks to all partners for the contribution!

04/2024 Mike Feuchter received Outstanding Paper Award during EuroSimE Conference 2024 in Catania

Reduced-Order Model for Solder Balls – Potential of projection-based approaches for representing viscoplastic behavior

The complexity of automotive and industrial electronic control units does not allow the use of regular finite element analyses for the complete design optimization with respect to thermomechanical reliability. Therefore, a novel approach is proposed that massively reduces the computational effort without compromising the result accuracy. It is based on a modular system of reduced-order models. In this paper, projection-based methods for capturing the complex material behaviour of solder balls are introduced as key enabler for this approach. It is shown that the discrete empirical interpolation method can reduce the number of degrees of freedom massively to speed-up the simulations with a high accuracy.