A deeper understanding of multi-laser additive manufacturing systems
1
Multi-laser additive manufacturing is revolutionizing how things are made. The industrial 3D printing technique allows multiple lasers to work simultaneously. Sometimes, the lasers are mounted on robotic arms, such as in laser wire or powder-directed energy deposition additive manufacturing, or guided by galvanometer mirrors used for laser powder bed additive manufacturing. Both techniques make printing much faster, increasing manufacturing efficiency.
How the printing paths are planned and how the design is broken into parts and distributed among various laser heads affect how heat spreads and defects form during printing. However, scientists haven’t fully explored how these factors in multi-laser systems work together to create perfectly manufactured parts.
To grasp a stronger insight into the systems, Assistant Professor Azadeh Haghighi, director of the Smarture Lab at UIC, and her team are investigating a new method that uses a type of artificial intelligence called Physics-Informed Neural Networks (PINN). PINN is a deep learning technique that integrates physical laws into a neural network’s training process.
This method looks at how heat builds up and moves during printing. First, it breaks the object into sections, then creates paths for the lasers, adjusting their power and speed. Next, it uses a clustering model – a way of grouping data – to choose the best path-planning strategies, based on how close the lasers are to each other while printing. The model is tested by comparing its results with the PINN-based heat model.
“Because this method can work with different shapes and designs, it helps find the best way to divide and plan the printing process. By linking path planning with how heat spreads, the framework offers a new way to keep temperatures more stable, which makes multi-laser additive manufacturing more reliable,” Haghighi said.
Multi-laser additive manufacturing is already transforming aerospace, defense, and energy sectors — where complex, high-performance components must be built quickly and precisely. By combining AI and physics, Haghighi’s research could unlock the next generation of smart, cooperative manufacturing systems — where multiple robots equipped with lasers or heat sources work together to 3D print structures, learning in real time to minimize defects and control heat. This framework could ultimately guide multi-robot in-space manufacturing, where autonomous systems collaborate to fabricate and repair defect-free large structures on the Moon, Mars, and beyond.
The research is part of the quality-aware cooperative printing direction in her lab, with broader implications for in-space manufacturing, and it was recently published in the journal Manufacturing Letters under the title “A physics-informed neural network framework for decomposition and path planning in multi-laser additive manufacturing.”