面向复合脆性材料激光热裂切割在线感知与实时智能优化控制

面向复合脆性材料切割存在的切割质量低、智能化程度不足等问题，本项目提出飞秒激光辅助连续激光诱导热裂切割在线感知与实时智能优化控制新方法。通过获取飞秒激光在复合脆性材料内部制备的定向初始裂纹状态信息，建立激光的热源吸收模型及包含初始裂纹信息的非线性裂纹扩展有限元模型，揭示裂纹扩展动力学行为机制；针对切割场景的动态、开放、强反光、存在不确定遮挡等特点，研究基于视频图像、红外热像及声发射信号的多模态信息特征提取方法，建立加工状态与多维度异质信号间的映射关系；通过数据分析筛选，提取影响切割质量的主要工艺参数，建立工艺参数与切割质量之间的映射关系模型，探索端到端最优化智能算法，实现切割过程的实时最优工艺参数组合。为实现复合脆性材料“无脱层”、“无崩边”、“无损伤”的高质高效智能切割提供理论基础和新方法，对解决脆性材料的绿色化、智能化加工有重要的研究意义和应用价值。

Cutting composite-brittle materials have the problems of low cutting quality and insufficient intelligence. This project proposes a new method of online perception and real-time intelligent optimal control of femtosecond laser-assisted continuous laser-induced thermal cracking. By acquiring the directional initial crack state information inside the composite-brittle material prepared by the femtosecond laser, the laser heat source absorption model and the nonlinear crack propagation finite element model containing the initial crack information are established to reveal the crack propagation dynamics behavior mechanism. In view of the dynamic, open, highly reflective and uncertain shielding features of the cutting scene, etc., study the multi-modal information feature extraction methods based on video images, infrared thermal images and acoustic emission signals, and establish the mapping relationship between processing status and multi-dimensional heterogeneous signals. Through data analysis and screening, extract the main process parameters that affect the cutting quality, establish the mapping relationship model between the process parameters and the cutting quality, explore the end-to-end optimization intelligent algorithm, and realize the real-time optimal process parameter combination of the cutting process. This project provides the theoretical basis and new methods for achieving high-quality and efficient intelligent cutting of composite brittle materials without delamination, edge collapse, and damage, which is of great research significance for solving the green and intelligent processing of brittle materials and application value.