BUP46 - OpticalRNN

The production of a competitive battery that stores enough energy for an acceptable range is crucial for the further development of high-performance electric vehicles. A promising solution is lithium-ion batteries, whose performance and cost largely depend on electrode production. To enhance the competitiveness of automakers, the electrode production process must be further optimized.
Ellipsometry is a standard tool used to characterize the thickness and properties of thin films and could be applied to monitor electrode production. However, the slow measurement process prevents its direct use during production (in-situ application).

Aim

The goal of this project is to develop a neural network that can meaningfully segment measurement signals, which were until now analyzed only as a whole. The network uses individual ellipsometric  dataas input and provides thicknesses and refractive indices as output. This should enable the identification of necessary measurement signals and allow faster measurement for in-situ monitoring.

Approach

• Simulation for data generation to train the network
• Configuration and training of the network
• Simulation for validation of the network

Benefit

The successful outcomes of this project could provide a groundbreaking method for analyzing physical data in areas where, due to limited knowledge, individual interpretation was until now not possible. Specifically, it enables software-based optimization to accelerate the measurement process in electrode manufacturing.