Samia Bouddahab

Within the Ansys Sound team, I worked on a final year project focused on identifying and separating acoustic components in signals using artificial intelligence models. My main tasks included:

- Calculating the Short-Time Fourier Transform (STFT) of signals in accordance with the Ansys Sound ecosystem.

- Creating and augmenting the dataset for model training.

- Applying semantic segmentation with the U-Net model to detect signal components.

- Optimizing and evaluating the model to improve defect detection performance.

- Discovering the acoustic domain and the application of artificial intelligence for sound analysis.

During my internship at ASTRIIS, I explored the field of predictive maintenance and vibration signals, as well as the crucial impact of artificial intelligence for sustainable equipment management. My main responsibilities included:

- Analyzing existing equipment for vibration measurement and documenting their specificities.

- Testing and comparing different vibration sensors, such as AC102, AC115, and ACC PCB accelerometers, to determine their effectiveness in various conditions.

- Conducting practical tests on equipment, like a coffee machine and a fan, to measure and analyze the generated vibrations.

- Programming a new board based on Raspberry Pi and MCC 172 for the acquisition, sampling, and transmission of time-series signals.

- Working with large datasets, optimizing storage using binary formats to ensure maximum efficiency.

- Writing reports and presenting results to technical teams.