The high reliance on imported lasers and related products for quality laser processing has led to key bottleneck technologies in China's high-end laser processing manufacturing equipment. Addressing the core needs of efficient processing + quality control + cost optimization in precision laser manufacturing, Professor Wu Di's research team has integrated multidisciplinary technologies such as laser physics, image/signal processing, and machine learning. They innovated a completely new online detection + precise energy optimization technology in the field of AI + advanced laser manufacturing, overcoming the limitations of traditional methods like low detection accuracy and poor real-time performance. This has achieved the industrial application of an intelligent control system covering process development - process monitoring - quality control.
Team Leader Wu Di, a university-applied professor and master's supervisor, has been selected for the Shanghai Sailing Program for Young Sci-Tech Talents, the third level of Jiangsu Province's High-Level Talent Program (333 Project), and Nantong's Jianghai Talent municipal recruitment program. He has led over 10 national/provincial ministerial and enterprise-commissioned R&D projects, with cumulative research funding of nearly 6 million RMB, and won the Bronze Award at the 2024 2nd National Postdoctoral Innovation and Entrepreneurship Competition. He has published over 50 high-level journal papers in the field of laser manufacturing and applied for more than 10 invention patents.
Relying on the Shanghai Class III Peak Discipline platform, Wu Di's team has long focused on systematic innovative research in in-situ monitoring and intelligent control of advanced welding processes. Collaborating with a leading listed company, the team tackled the 2024 Shanghai Greater Zero Bay Key Technology Unveiling the List and Appointing the Commander project – Laser Welding Process Monitoring and Control System Based on Multi-source Sensing (total funding: 4.5 million RMB). They successfully independently developed a complete set of domestically produced integrated laser intelligent processing equipment for welding-monitoring-control.
The core technological innovation lies in the deep integration of the welding execution module with intelligent monitoring and dynamic control modules, paired with a self-developed, intellectually property-backed laser welding intelligent diagnosis system powered by process mechanism + AI. This system primarily consists of three optical sensing technologies: multi-spectral sensing, active vision, and Optical Coherence Tomography. By integrating four key steps – process mechanism, full-process monitoring, AI learning models, and central monitoring – it forms a closed-loop technical system of processing is monitoring, monitoring is control, solving the industry pain points of traditional laser equipment being heavy on processing, light on diagnosis, and difficult to control.
Core Technology 1: Rapid Surface Defect Detection Based on Spectral-Visual Multi-modal Sensing
Developed a rapid surface defect detection system based on multi-spectral and visual multi-modal sensing. With a high-speed acquisition capability of 250kHz, it provides real-time feedback on subtle fluctuations in three signals: laser process plasma, reflected light, and molten pool radiation. Combined with visual sensing of the oscillation behavior of the surface molten pool keyhole, it finely reflects the defect formation process in spatial and temporal dimensions. Unique AI algorithms enhance the detection accuracy for surface defects.
Core Technology 2: Real-time Detection of Internal Penetration Based on In-situ OCT Sensing
Developed a real-time penetration monitoring system based on high-precision OCT technology, enabling measurement throughout the entire welding process (pre-welding, during welding, post-welding). It achieves a detection accuracy of 20µm, processing time under 15ms, and a keyhole depth detection range of up to 12mm. By combining OCT data processing algorithms and machine learning models, it can accurately predict penetration depth and lack-of-fusion defects, with a precision error ≤ 0.1mm.
Core Technology 3: Process Development Optimization and Quality Control for Typical Laser Welding Scenarios
Established a process knowledge database covering over 1000 different welding scenarios. Developed a data-mechanism fusion-driven platform for efficient optimization of laser welding processes and quality control. This reduces reliance solely on manual experience and costly process trial-and-error, ultimately achieving the goal of zero-defect welds.
The domestically developed integrated laser intelligent processing equipment for welding-monitoring-control has been successfully applied in precision manufacturing fields such as new energy vehicles, power batteries, and 3C electronics. It has achieved over a 50% increase in production efficiency and over a 30% reduction in processing costs, with an estimated new output value exceeding 50 million RMB. To date, the team's 6 invention patents/software copyrights and over 20 high-level papers provide solid technical support for the equipment's core technological innovation and industrial application.
Relevant research findings by Professor Wu Di have been published in Journal of Manufacturing Processes, a Q1 TOP journal in the laser manufacturing field.
Developed laser intelligent processing equipment based on a multi-source sensing system
