三维形貌原位数字全息测量的物理行为建模与粒子量化

This project focuses on the basic theory and the key technology around in-situ digital holography measurement for the three-dimensional heterogeneous topography of complex surface. The multiple wavelengths theory for three-dimensional topography measurement with the constraints of topological change, variable view field and vibrating surface is investigated. The physical modeling of measurement, the measurement error tracing, the precision response law and the error compensation are also investigated. The region segmentation of complex surface for the measurement planning is proposed. The imaging quality and the measurement range are taken as the objective functions. The optimization of configuration space with the constraints of optical imaging, measurement space, motion space, collision detection and geometrical feature is investigated for in-situ measurement. This project proposes the particle description based three-dimensional topography representation theory with facilitating the performance analysis. The normal velocity flow of particle is derived by using the variable theory. Then the feature based particle adaptive distribution and the measured data quantification are realized by solving the Hamilton-Jacobi partial differential equation. With the application of above theories, the project group sets up the software and hardware experimental platform. The measurement error is controlled below 0.1 micron and the geometric description and the physical feedback are integrated into the particle model. The machining quality is evalued in-situ during the digital manufacturing process. The outcomes of this project will provide the theoretical basis and technical support for the error assessment, the performance analysis and the on-line compensation for the digital design, machining and measurement integration.

针对复杂曲面三维异构加工形貌，本项目围绕原位数字全息测量的基础理论与关键技术展开研究。研究拓扑突变、变视场与振动扰动约束的三维形貌耦合波长测量理论、及测量过程的物理行为建模，探索误差溯源、测量精度响应规律及数字全息的误差补偿方法；建立面向测量的复杂曲面最优分割方法，以成像质量与视场范围为目标，建立多领域混合约束(成像、可测空间、运动约束、碰撞与几何特征等)的原位测量位形规划理论；提出基于粒子描述的三维形貌隐式表征方法，利用变分理论推导法向子空间内的粒子梯度流，通过求解哈密尔顿-雅克比偏微分方程实现基于特征的粒子自适应分布与测量数据的高精度量化。应用上述理论，搭建原位测量实验平台，测量工件的几何描述与物理反馈统一于粒子模型，实现测量误差小于0.1微米与数字化制造过程中的加工质量评价。研究成果可为数字化设计-加工-测量一体化的误差评定、性能分析及在线补偿的闭环制造模式提供理论基础与技术手段。

制造业全球化竞争的背景下，高、精、尖装备的制造不断向现代测量方法、测量手段及测量极限提出新的挑战，数字化制造的高精度、多尺度与异构形貌的快速、高效的三维测量及量化表征是制约数字化设计-加工-测量一体化发展的瓶颈问题，针对上述问题，本项目主要研究测量加工一体化中的关键核心技术。主要研究内容包括：复杂加工表面或加工件的测量过程物理行为建模、散乱环境测量点云数据的特征识别、原位测量过程中的精度创成原理、大范围测量过程的区域主动分割与最优位姿规划、异构形貌原位测量实验平台的搭建及应用研究。项目发表国际期刊与会议论文共12篇，申请专利与软著与专利共4项，培养研究生6名。重要理论成果包括散乱环境点云的生成、根据点云数据进行加工件的特征识别、根据点云特征的单个零件区域分割及机器人视觉伺服操作。