Acoustic Emission Dataset for Multi-Laser LPBF Systems: Supporting DUAL DISCO Signal Processing Technique

This dataset contains acoustic emission (AE) signals collected during experiments with multi-laser Laser Powder Bed Fusion (LPBF) systems. The data supports the research presented in the paper titled "DUAL DISCO: A Novel Approach to Acoustic Emission Monitoring in Multi-Laser LPBF Systems." The dataset is designed to facilitate the development and validation of advanced signal processing techniques, specifically the DUAL DISCO method, which aims to disentangle and analyze AE signals from simultaneous laser operations.

Contents:

Raw_data.zip: This file contains the AE signals used for training. The data was recorded using two condenser microphones positioned around the LPBF build plate, capturing signals from both sequential and simultaneous laser operations across various melting regimes (conduction and keyhole modes).

Raw_data_test.zip: This file includes the AE signals used for testing, recorded under different experimental conditions to evaluate the generalization capabilities of signal processing algorithms.

params.xlsx: This spreadsheet provides the ground truth labels for the training data, indicating the melting regime (conduction or keyhole) for each signal in Raw_data. 

params_test.xlsx: This spreadsheet contains the ground truth labels for the test data, similarly indicating the melting regime for each signal in Raw_data_test.

Applications:

This dataset is intended for researchers and practitioners in the field of additive manufacturing and signal processing. It can be used to:

Develop and test new algorithms for AE signal processing in multi-laser LPBF systems.

Explore the acoustic characteristics of different melting regimes.

Enhance the understanding of process monitoring techniques in additive manufacturing.

Acknowledgments:

The dataset was collected using the AddUp FormUp 350 machine and is part of a research project supported by the Bern Economic Development Agency. Special thanks to Thomas Rytz for technical support.