Lasers that Learn: AI-enabled intelligent materials processing

Lead Research Organisation: University of Southampton
Department Name: Optoelectronics Research Centre (ORC)

Abstract

Lasers are used for an extremely wide range of manufacturing processes. This is due, in part, to their significant flexibility with respect to parameters such as pulse length, pulse energy, wavelength, and beam size. However, this flexibility comes at a price, namely the significant amount of time that must be dedicated to finding the optimal set of parameters, for each and every manufacturing process or customer specification. The standard practice in industry is the mechanical collection of laser machining data for all parameter combinations, in order to find the optimal combination of parameters. However, this process is both time-consuming and unfocussed, and it can take days or weeks, hence costing unnecessary time and money. Even when the optimal parameters have been determined, small changes, for example in laser power or beam shape, during manufacturing, can result in a final product quality that is below the required standard, once again costing time and money. There will also be instances where the specification is not known in advance due to variability in the manufacturing process. What is needed, therefore, are a series of methodologies for identifying optimal parameters before manufacturing, for providing real-time monitoring and error correction during manufacturing, and for enabling process-control (for example stopping the laser exactly at task completion, or varying the laser power for the final finishing steps).

The research field of machine learning has seen some extremely significant developments in recent years, and it is now widely understood to be a catalyst for a fundamental change across almost all manufacturing industries. The objective of this proposal is to develop the technological and human expertise required for the integration of machine learning approaches into the UK laser-based manufacturing industry and the NHS.

This proposal therefore seeks to leverage state-of-the-art machine learning techniques for solving well-known problems in laser-based manufacturing and materials processing, resulting in improvements in efficiency, reliability, and precision. The results of this proposal will lead to time and money savings for both the UK laser-based manufacturing industry and the NHS. This proposal will cover the application of neural networks for modelling and optimising of femtosecond laser machining, instantly identifying laser-based manufacturing parameters for any customer specification, automatically compensating for residual cavity effects in fibre lasers, enabling targeted delivery of laser light for psoriasis treatment, and laser welding process enhancement in real-time via multi-sensor data.

Planned Impact

Impact Summary

The main beneficiaries of this research lie within the manufacturing and healthcare sectors, each of which represents a major opportunity for financial return (economic benefit), and healthcare provision (societal impact).

1. Economic benefit

Our industrial project partners, Oxford Lasers and SPI Lasers will have direct benefit from this work. Oxford Lasers will have the potential for faster turn-around times on customer orders, along with higher precision and reliability, and of course freeing up valuable technical staff time for more creative duties. This extended capabilities will therefore offer a direct financial benefit to the company. SPI Lasers, similarly, will have the potential for a more precise and reliable fibre laser system, with capabilities that far surpass current abilities. This will directly lead to the opening of new markets (such as high-speed manufacturing of bespoke security features on high-value objectives), and lead directly to company competitiveness and financial benefit. For both companies, financial gains from increased (global) market share could lead to employment of additional UK staff. Whilst integrating the proposed machined learning and laser machining techniques into UK industry is an important objective of this proposal, a spinout from Southampton is also a realistic possibility, if financial backing is secured to take this forward.

2. Healthcare and societal benefit

With the continuously increasing strains on the over-burdened public health systems of developed nations such as the UK, and likewise for developing countries, the role of healthcare, and in particular the diagnosis and treatment of skin conditions, such as skin cancer and psoriasis, is an increasingly important factor. The proposed technique, which combines state-of-the-art machine learning, with directed laser energy transfer, for the identification and targeted irradiation of affected areas of skin, offers a significant potential for cost savings, and improved patient care. If the initial proof-of-principle experiments proposed here are successful, then there really is potential for a genuinely radical breakthrough in the treatment of psoriasis.

3. Training and the next generation of researchers

This programme will be of immense value to the next generation of research scientists and, in particular, those researchers who can span the 'single discipline' problem so prevalent in academia. Scientists involved in this project will become experts with both lasers and machine learning. When combined with healthcare, their skillset will be extremely valuable for both UK academia and laser-based manufacturing companies in the UK.

Publications

10 25 50

publication icon
Courtier A.F. (2022) Predicting the Surface Topography of Stainless Steel Cut by Fibre Laser via Deep Learning in 2022 Conference on Lasers and Electro-Optics, CLEO 2022 - Proceedings

publication icon
Courtier A.F. (2022) Predicting the Surface Topography of Stainless Steel Cut by Fibre Laser via Deep Learning in Optics InfoBase Conference Papers

publication icon
Courtier AF (2021) Modelling of fibre laser cutting via deep learning. in Optics express

publication icon
Grant-Jacob J (2022) Deep learning in airborne particulate matter sensing: a review in Journal of Physics Communications

publication icon
Grant-Jacob J (2023) Real-time control of laser materials processing using deep learning in Manufacturing Letters

publication icon
Grant-Jacob J (2023) Live imaging of laser machining via plasma deep learning in Optics Express

publication icon
Grant-Jacob J (2020) Lensless imaging of pollen grains at three-wavelengths using deep learning in Environmental Research Communications

publication icon
Grant-Jacob J (2021) In-flight sensing of pollen grains via laser scattering and deep learning in Engineering Research Express

publication icon
Grant-Jacob JA (2022) Single-frame 3D lensless microscopic imaging via deep learning. in Optics express

publication icon
Grant-Jacob JA (2024) Evolution of laughter from play. in Communicative & integrative biology

 
Title Visualization 1.mp4 
Description The 5th laser pulse (highlighted in red) was deliberately displaced from its undisturbed location to all possible positions on the virtual workpiece. In almost all cases the RL agent was able to successfully complete machining of the target pattern regardless of this disturbance. 
Type Of Art Film/Video/Animation 
Year Produced 2022 
URL https://opticapublishing.figshare.com/articles/media/Visualization_1_mp4/19077152/1
 
Title Visualization 1.mp4 
Description The 5th laser pulse (highlighted in red) was deliberately displaced from its undisturbed location to all possible positions on the virtual workpiece. In almost all cases the RL agent was able to successfully complete machining of the target pattern regardless of this disturbance. 
Type Of Art Film/Video/Animation 
Year Produced 2022 
URL https://opticapublishing.figshare.com/articles/media/Visualization_1_mp4/19077152
 
Title Visualization 2.mp4 
Description The 5th laser pulse (highlighted in red) was deliberately displaced from its undisturbed location to all possible positions on the virtual workpiece. In almost all cases the RL agent was able to successfully complete machining of the target pattern regardless of this disturbance. 
Type Of Art Film/Video/Animation 
Year Produced 2022 
URL https://opticapublishing.figshare.com/articles/media/Visualization_2_mp4/19535962/1
 
Title Visualization 2.mp4 
Description The 5th laser pulse (highlighted in red) was deliberately displaced from its undisturbed location to all possible positions on the virtual workpiece. In almost all cases the RL agent was able to successfully complete machining of the target pattern regardless of this disturbance. 
Type Of Art Film/Video/Animation 
Year Produced 2022 
URL https://opticapublishing.figshare.com/articles/media/Visualization_2_mp4/19535962
 
Title Visualization 3.mp4 
Description The 5th laser pulse (highlighted in red) was deliberately displaced from its undisturbed location to all possible positions on the virtual workpiece. In almost all cases the RL agent was able to successfully complete machining of the target pattern regardless of this disturbance. 
Type Of Art Film/Video/Animation 
Year Produced 2022 
URL https://opticapublishing.figshare.com/articles/media/Visualization_3_mp4/19535965
 
Title Visualization 3.mp4 
Description The 5th laser pulse (highlighted in red) was deliberately displaced from its undisturbed location to all possible positions on the virtual workpiece. In almost all cases the RL agent was able to successfully complete machining of the target pattern regardless of this disturbance. 
Type Of Art Film/Video/Animation 
Year Produced 2022 
URL https://opticapublishing.figshare.com/articles/media/Visualization_3_mp4/19535965/1
 
Title Visualization 4.mp4 
Description The 5th laser pulse (highlighted in red) was deliberately displaced from its undisturbed location to all possible positions on the virtual workpiece. In almost all cases the RL agent was able to successfully complete machining of the target pattern regardless of this disturbance. 
Type Of Art Film/Video/Animation 
Year Produced 2022 
URL https://opticapublishing.figshare.com/articles/media/Visualization_4_mp4/19535971/1
 
Title Visualization 4.mp4 
Description The 5th laser pulse (highlighted in red) was deliberately displaced from its undisturbed location to all possible positions on the virtual workpiece. In almost all cases the RL agent was able to successfully complete machining of the target pattern regardless of this disturbance. 
Type Of Art Film/Video/Animation 
Year Produced 2022 
URL https://opticapublishing.figshare.com/articles/media/Visualization_4_mp4/19535971
 
Title Visualization 5.mp4 
Description The 5th laser pulse (highlighted in red) was deliberately displaced from its undisturbed location to all possible positions on the virtual workpiece. In almost all cases the RL agent was able to successfully complete machining of the target pattern regardless of this disturbance. 
Type Of Art Film/Video/Animation 
Year Produced 2022 
URL https://opticapublishing.figshare.com/articles/media/Visualization_5_mp4/19535974
 
Title Visualization 5.mp4 
Description The 5th laser pulse (highlighted in red) was deliberately displaced from its undisturbed location to all possible positions on the virtual workpiece. In almost all cases the RL agent was able to successfully complete machining of the target pattern regardless of this disturbance. 
Type Of Art Film/Video/Animation 
Year Produced 2022 
URL https://opticapublishing.figshare.com/articles/media/Visualization_5_mp4/19535974/1
 
Description Deep learning holds significant promise for laser machining due to its capacity for process optimization and real-time feedback control. Through its ability to analyse vast amounts of data, deep learning algorithms can discern intricate patterns in laser machining processes, facilitating the optimization of parameters such as laser power, beam focus, and spot size. This optimisation leads to enhanced efficiency, precision, and quality in material processing. Moreover, deep learning enables real-time monitoring of machining conditions and immediate adjustments based on dynamic feedback, ensuring consistent performance and minimizing defects. By leveraging the power of deep learning, laser machining can achieve unprecedented levels of accuracy and productivity, driving advancements across various industrial sectors. This project has demonstrated several world-firsts in the area of laser-based manufacturing.
Exploitation Route Firstly, the research findings and methodologies developed through this funding could be shared through academic publications, conferences, and workshops, enabling researchers and engineers in the field to adopt and build upon these advancements. Additionally, the trained deep learning models and algorithms could be provided allowing manufacturers and laser machining facilities to integrate them into their systems for process optimization and real-time control. Moreover, collaborations between academia, industry, and government entities could foster the development of standardized tools and frameworks, facilitating broader adoption and implementation of deep learning techniques in laser machining applications. Overall, the outcomes of this funding have the potential for innovation and drive widespread improvements in laser machining processes across various sectors.
Sectors Manufacturing

including Industrial Biotechology

 
Description Work from this project, in terms of deep learning and its application to manufacturing, has led to several important and ongoing collaborations with industrial partners in the UK and abroad.
First Year Of Impact 2022
Sector Manufacturing, including Industrial Biotechology
Impact Types Economic

 
Title distribution_of_angles_v=15.txt 
Description Distribution of angles for v=15 m/min, experimental and predicted. 
Type Of Material Database/Collection of data 
Year Produced 2021 
Provided To Others? Yes  
URL https://opticapublishing.figshare.com/articles/dataset/distribution_of_angles_v_15_txt/15050136
 
Title distribution_of_angles_v=15.txt 
Description Distribution of angles for v=15 m/min, experimental and predicted. 
Type Of Material Database/Collection of data 
Year Produced 2021 
Provided To Others? Yes  
URL https://opticapublishing.figshare.com/articles/dataset/distribution_of_angles_v_15_txt/15050136/1
 
Title distribution_of_angles_v=20.txt 
Description Distribution of angles for v=20 m/min, experimental and predicted. 
Type Of Material Database/Collection of data 
Year Produced 2021 
Provided To Others? Yes  
URL https://opticapublishing.figshare.com/articles/dataset/distribution_of_angles_v_20_txt/15050142/1
 
Title distribution_of_angles_v=20.txt 
Description Distribution of angles for v=20 m/min, experimental and predicted. 
Type Of Material Database/Collection of data 
Year Produced 2021 
Provided To Others? Yes  
URL https://opticapublishing.figshare.com/articles/dataset/distribution_of_angles_v_20_txt/15050142
 
Title distribution_of_pixel_intensities_experimental_predicted.txt 
Description Distribution of pixel intensities for v=15, 20 m/min, experimental and predicted 
Type Of Material Database/Collection of data 
Year Produced 2021 
Provided To Others? Yes  
URL https://opticapublishing.figshare.com/articles/dataset/distribution_of_pixel_intensities_experimenta...
 
Title distribution_of_pixel_intensities_experimental_predicted.txt 
Description Distribution of pixel intensities for v=15, 20 m/min, experimental and predicted 
Type Of Material Database/Collection of data 
Year Produced 2021 
Provided To Others? Yes  
URL https://opticapublishing.figshare.com/articles/dataset/distribution_of_pixel_intensities_experimenta...
 
Title experimental_and_predicted_labels_confusion_matrix.txt 
Description Experimental and predicted cutting speeds for a confusion matrix 
Type Of Material Database/Collection of data 
Year Produced 2021 
Provided To Others? Yes  
URL https://opticapublishing.figshare.com/articles/dataset/experimental_and_predicted_labels_confusion_m...
 
Title experimental_and_predicted_labels_confusion_matrix.txt 
Description Experimental and predicted cutting speeds for a confusion matrix 
Type Of Material Database/Collection of data 
Year Produced 2021 
Provided To Others? Yes  
URL https://opticapublishing.figshare.com/articles/dataset/experimental_and_predicted_labels_confusion_m...
 
Title percentage_of_dark_spots_experimental_and_predicted.txt 
Description Percentage of dark spots for v=15, 20 m/min, experimental and predicted. 
Type Of Material Database/Collection of data 
Year Produced 2021 
Provided To Others? Yes  
URL https://opticapublishing.figshare.com/articles/dataset/percentage_of_dark_spots_experimental_and_pre...
 
Title percentage_of_dark_spots_experimental_and_predicted.txt 
Description Percentage of dark spots for v=15, 20 m/min, experimental and predicted. 
Type Of Material Database/Collection of data 
Year Produced 2021 
Provided To Others? Yes  
URL https://opticapublishing.figshare.com/articles/dataset/percentage_of_dark_spots_experimental_and_pre...