LUMIN: AI-Enhanced Intelligent Manufacturing with Advanced Laser and Sensor Systems

Project ID: 2531bd1687

(You will need this ID for your application)

Research Theme: Manufacturing The Future

Research Area(s): Manufacturing Technologies
Advanced Materials
Engineering

UCL Lead department: Mechanical Engineering

Department Website

Lead Supervisor: Chu Lun Alex Leung

Project Summary:

Achieving the UN Sustainable Development Goals requires cleaner, smarter, and more efficient manufacturing. Laser-based material processing—such as additive manufacturing (AM) and laser welding—offers a powerful route to net-zero by enabling near-net-shape production and joining of dissimilar materials with minimal waste and energy use. However, process-induced imperfections—such as pores, spatter, and lack of fusion—can compromise component performance, leading to costly rework, scrap, and even in-service failure.

Traditional process qualification in laser AM and welding is slow, expensive, and resource-intensive, often taking months or years to qualify a single process across varying materials, geometries, and environmental conditions. Extensive experimentation, metallography, and data analysis are required, and existing trial-and-error or purely simulation-based methods can no longer keep pace with industrial needs. This project aims to understand, detect, and minimise defect formation in laser material processing by integrating artificial intelligence (AI) with multimodal sensing, enabling a new paradigm of data-driven process qualification.

By leveraging few-shot learning approaches, AI models trained on diverse historical datasets and literature can generalise to new processing conditions, rapidly predicting melt pool dynamics, defect formation, and optimal parameters. Only a small number of new experiments will be needed to fine-tune the models, enabling qualification within days rather than weeks, thereby reducing cost, energy use, and carbon footprint. Working in collaboration with IPG Photonics and industrial partners, the successful candidate will gain hands-on experience in laser processing, advanced sensing, in situ X-ray imaging, and AI-based data analytics.

Project Objectives:

  1. Curate and label high-dimensional, multi-modal datasets integrating text, signal, and image data acquired during laser processing.
  2. Uncover key defect formation mechanisms through data-driven feature extraction and correlation analysis.
  3. Utilise synchrotron X-ray imaging ground truth to develop and validate predictive AI models for anomaly prediction.
  4. Deploy and assess machine-learning models for real-time anomaly detection and process control.