Online monitoring: Selected projects in laser processing

In situ and real-time monitoring of AM process

Additive manufacturing (AM) is prevented further penetration into wider industries due to the lack of process reproducibility. This is due to the high non-linearity in the laser-material interaction. This issue raises the demand for in situ and real-time quality monitoring that is not available today at an industrial scale. The systems must embrace not only material heating, melting, and solidification, but also phase transformations and residual stresses during the cooling and solidification of the workpiece. Those make the workpiece quality sensitive to a vast number of process parameters, and any slight change of one parameter may significantly impact the quality or mechanical properties.
Research at Empa focused on developing a new online monitoring system that are able to detect laser regimes (conduction mode, lack of fusion, keyhole), defect formation (pores, cracks) as well as microstructures for any AM metal processes.
The three major achievements are to detect online (a) the different AM regimes as well as pores formation with high confidences, (b) the different stables and unstable keyhole process leading to defect formation and (c) detect defects (pores) removal.

 

Selected publications
  • Shevchik S.A., Le-Quang T., Meylan B., Vakili-Farahani F., Olbinado M.P., Rack A., Masinelli G., Leinenbach C., and Wasmer K., “Supervised Deep Learning for Real-Time Quality Monitoring of Laser Welding with X-Ray Radiographic Guidance", Scientific Report, Vol. 10, paper ID: 3389, 2020,
    https://doi.org/10.1038/s41598-020-60294-x

  • Shevchik S.A., Masinelli G., Kenel C., Leinenbach C., and Wasmer K., “Deep Learning For In Situ and Real-Time Quality Monitoring in Additive Manufacturing Using Acoustic Emission”, IEEE Transactions on Industrial Informatics, Vol. 19, Issue 9, pp: 5194-5203, 2019, https://doi.org/10.1109/TII.2019.2910524

  • Wasmer K., Le-Quang T., Meylan B., Vakili-Farahani F., Olbinado M.P., Rack A., and Shevchik S.A., “Laser Processing Quality Monitoring by Combining Acoustic Emission and Machine Learning: A High-Speed X-Ray Imaging Approach”, Procedia CIRP, Vol. 74, Issue 2018, pp: 654–658, 2018, https://doi.org/10.1016/j.procir.2018.08.054  

  • Le-Quang T., Shevchik S.A., Meylan B., Vakili-Farahani F., Olbinado M.P., Rack A., and Wasmer K., “Why Is In Situ Quality Control of Laser Keyhole Welding a Real Challenge?” Procedia CIRP, Vol. 74, Issue 2018, pp: 649–653, 2018,
    https://doi.org/10.1016/j.procir.2018.08.055

  • Shevchik S.A., Kenel C., Leinenbach C. and Wasmer K., “Acoustic Emission for In Situ Quality Monitoring in Additive Manufacturing Using Spectral Convolutional Neural Networks”, Additive Manufacturing, Vol. 21, Issue May 2018, pp: 598-604, 2018,
    https://doi.org/10.1016/j.addma.2017.11.012

     

     

In situ and real-time monitoring of laser welding processes

Despite the fact that laser welding process has been extensively studied, the mechanism responsible for defects formation in the weld joint, in particular porosity, is not fully understood. It is partially due to the complexity of the laser-material interaction and the short lifetimes of the transient (opening and collapse of the keyhole in its unstable state) and events (melting, solidification, pore formation).
Research at Empa focused on developing develop an in situ and real-time monitoring method by combining various types of sensors (optical, acoustic, and optoacoustic) and advanced machine learning algorithms.
The major achievement was the development of an online system able detect with high confidence and in a few milliseconds various welding states such as conduction mode, keyhole with and without porosity, as well as transient such as and unstable keyhole.

 

Selected publications
  • Pandiyan V., Drissi-Daoudi R., Shevchik S.A., Masinelli G., Logé R., and Wasmer K., "Analysis of Time, Frequency and Time-Frequency Domain Features From Acoustic Emissions During Laser Powder-Bed Fusion Process" Procedia CIRP, Vol. 94, Issue 2020, pp: 392–397, 2020, https://doi.org/10.1016/j.procir.2020.09.152

  • Shevchik S.A., Le-Quang T., Meylan B., Vakili-Farahani F., Olbinado M.P., Rack A., Masinelli G., Leinenbach C., and Wasmer K., “Supervised Deep Learning for Real-Time Quality Monitoring of Laser Welding with X-Ray Radiographic Guidance", Scientific Report, Vol. 10, paper ID: 3389, 2020, https://doi.org/10.1038/s41598-020-60294-x

  • Shevchik S.A., Le-Quang T., Vakili-Farahani F., Neige F., Meylan B., Zanoli S., and Wasmer K., “Laser Welding Quality Monitoring Via Graph Support Vector Machine With Data Adaptive Kernel”, IEEE Access, Vol. 7, Issue 1, pp: 93108 - 93122, 2019, https://doi.org/10.1109/ACCESS.2019.2927661

  • 24.    Wasmer K., Le-Quang T., Meylan B., Vakili-Farahani F., Olbinado M.P., Rack A., and Shevchik S.A., “Laser Processing Quality Monitoring by Combining Acoustic Emission and Machine Learning: A High-Speed X-Ray Imaging Approach”, Procedia CIRP, Vol. 74, Issue 2018, pp: 654–658, 2018, https://doi.org/10.1016/j.procir.2018.08.054

     

     

     

Control of laser welding process

At present, the most commonly reported approach to increase the repeatability of the laser process quality is the application of traditional regulators, such as proportional-integral (PI) or proportional-integral-derivative (PID) controllers. These methods allow tracking the desired weld quality using measurements of the surface temperature of the process zone (PZ) as feedback. Unfortunately, since they are based on the linearization of the non-linear laser dynamics, they can only operate in a narrow range of the process parameters.
Research at Empa focused on exploring the problem of monitoring AM processes and the opportunity it provides:  a data-driven approach towards laser control. Specifically, we take advantage of recent advances in Machine Learning developments to derive efficient representations of the AM process from the high-dimensional sensory input and use them to generalize previous experiences to new situations. An example of a developed controller capable of autonomously learns a control law achieving a predefined quality independently from the starting conditions and without prior knowledge of the process dynamics.

 

Selected publications
  • Masinelli G., Le-Quang T., Zanoli S., Wasmer K., and Shevchik S.A., “Adaptive Laser Welding Control: A Reinforcement Learning Approach", IEEE Access, Vol. 8, pp: 103803 - 103814, 2020, https://doi.org/10.1109/ACCESS.2020.2998052

  • Masinelli G., Shevchik S.A., Pandiyan V., Le-Quang T., and Wasmer K., “Artificial Intelligence for Monitoring and Control of Metal Additive Manufacturing”, Presented at Additive Manufacturing for Products and Applications Conference 2020 (AMPA), Zürich, Switzerland, 1-3 September, 2020 and in Proceedings of Additive Manufacturing in Products and Applications - AMPA2020 – Industrializing Additive Manufacturing, M. Meboldt and C. Klahn (eds.), Zürich, Switzerland, pp: 205-220, Springer International Publishing, 2020, ISBN 978-3-030-54333-4, https://doi.org/10.1007/978-3-030-54334-1_15

     

     

In situ and real-time monitoring of laser osteotomy

Laser processes are also used for medical applications such as laser ablation for osteotomy (surgery that cuts and reshapes your bones). However, during the ablation, it is of utmost importance to differentiate the type of ablated tissue (skin, fat, muscle and bones) and laser power.
Research at Empa focused on developing an in situ and real-time monitoring system able to classify with high confidence both the tissue and laser power.
The main achievements is our system classifies with high confidence both the type of tissue and laser power.

 

Selected publications
  • Shevchik S.A., Kenhagho H.N., Le-Quang T., Neige F., Meylan B., Guzman R., Cattin P., Zam A., and Wasmer K., "Machine Learning Monitoring for Laser Osteotomy", Journal of Biophotonics, Vol. 14, Issue 4, paper ID: e202000352, pp: 1-11, 2021,
    https://doi.org/10.1002/JBIO.202000352

  • Kenhagho H.N., Shevchik S.A. Saeidi F., Neige F., Meylan B., Rauter G., Guzman R., Cattin P., Wasmer K., Zam A., “Characterization of Ablated Bone and Muscle for Long-Pulsed Laser Ablation in Dry and Wet Conditions”, Materials, Special Issue Advances in Laser Technologies and Applications, Vol. 12, Issue (8), paper ID: 1338, pp: 1-16, 2019, https://doi.org/10.3390/ma12081338

  • Nguendon H.K., Faivre N., Meylan B., Shevchik, S.A., Rauter G., Guzman R., Cattin P.C., Wasmer K., and Zam A., “Characterization of Ablated Porcine Bone and Muscle Using Laser-Induced Acoustic Wave Method for Tissue Differentiations”, Presented at Society of Photographic Instrumentation Engineers (SPIE 2017), Medical Laser Applications and Laser-Tissue Interactions VIII, Vol. 10417, paper ID: 104170N, Munich, Germany, 26 – 29 June, 2017, http://dx.doi.org/10.1117/12.2286121