Non-contact ultrasound testing of batteries

Hy-Met are a company using ultrasound for non-destructive testing of batteries. These ultrasound signals contain information on the health of a battery. Hy-Met aim to understand real acoustic data. However, data for particular defects is scarce. The challenge is to understand the reflection and transmission of ultrasound signals within the layers of battery, especially with defects present in certain regions of the battery. In this report, we seek to (i) understand properties of current models, (ii) apply data analysis techniques to extract relevant features from real data, (iii) understand reflection and transmission via probabilistic and combinatorial models. These may lead to the generation of synthetic data for further analysis and training.