Simultaneous Physical Field Reconstruction and Property Identification Using Multi-Target Sensing Methods for Metal Additive-Manufacturing and Thin-Walled Machining

使用多目标传感方法进行金属增材制造和薄壁加工的同步物理场重建和属性识别

• 批准号: 1662700
• 负责人: Kok-Meng Lee
• 金额: $ 29.99万
• 依托单位: Georgia Tech Research Corporation
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2017
• 资助国家: 美国
• 起止时间: 2017-06-15 至 2021-05-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1662700&HistoricalAwards=false
• 关键词: Simultaneous; Physical; Field; Reconstruction; Property

Metallic Additive-Manufacturing using laser sintering to fabricate parts layer-by-layer directly from a 3D computer model has emerged and has grown rapidly in the past decade. Such technology can significantly reduce time-to-market, improve product quality, reduce material-waste, and reduce cost. This project will help overcome important roadblocks that currently exist before this technology can be widely adopted. This research will use eddy current-based novel sensing mechanism during manufacturing process that will enable precision in manufacturing along with reduced cost. This technique will allow a robust layer-by-layer quality assessment while the part is being fabricated and the production rates will be expedited eliminating post-build inspection or destructive testing. While the method is in the context of metal additive manufacturing, the underlying system architecture and ideas can be extended to spawn a spectrum of different applications; for example, new medical applications such as magnetic induction tomography for accurate diagnosis of intracerebral hemorrhage. Several industries may benefit, including the automotive, the aerospace, the medical and healthcare, and industrial manufacturing sectors. The project also includes interesting K-12 outreach efforts in addition to curriculum development. The objective of the research is to develop a novel methodology to model a distributed-parameter system, reconstruct its physical fields and infer its system properties from limited measurements for analyzing and controlling its dynamic behaviors. A novel multi-target sensing methodology based on a set of multi-frequency eddy-current sensing will be developed for reconstructing the physical fields which measure geometrical parameters as well as detect surface and subsurface defects during the machining process. The work will also formulate dynamic model of the thin-walled plate during machining, and develop efficient methods for reconstructing its displacement, strain and stress fields from limited displacement measurements in real time for monitoring and controlling the dynamics of the distributed-parameter system. Unlike traditional single-frequency eddy-current sensors the multi-target sensing system can adaptively synthesizes a relatively high-resolution eddy-current pattern between adjacent coils with a combination of appropriate frequencies to simultaneously determine the displacement, thickness and electrical conductivity. The distributed current source (DCS) modeling method for sensor design and analysis is highly efficient and reduces computational complexity. It is expected that the ability to reconstruct the multi-physical fields (electric, magnetic, displacement, force, strain and stress) from the multi-target sensors and field reconstruction algorithms during part fabrication not only will offer intuitive insights into the effects of several critical factors (such as material mechanical properties, boundary geometrical/clamping constraints and damping coefficients) on thermal/machining induced residue stresses that are generally the main cause of thin-walled product distortions, and but also will provide an essential basis to vibration suppression.

在过去的十年里，使用激光烧结直接从3D计算机模型逐层制造零件的金属增材制造技术已经出现并迅速发展。这种技术可以显著缩短产品上市时间，提高产品质量，减少材料浪费，降低成本。该项目将有助于克服目前在该技术被广泛采用之前存在的重要障碍。本研究将在制造过程中使用基于涡流的新型传感机制，以提高制造精度并降低成本。该技术将允许在零件制造时进行可靠的逐层质量评估，并且将加快生产率，从而消除后期检查或破坏性测试。虽然该方法是在金属增材制造的背景下，但底层系统架构和思想可以扩展，以产生一系列不同的应用；例如，磁感应断层扫描等新的医学应用可以准确诊断脑出血。一些行业可能会受益，包括汽车、航空航天、医疗保健和工业制造部门。除了课程开发之外，该项目还包括有趣的K-12推广工作。本研究的目的是开发一种新的方法来模拟分布参数系统，重建其物理场，并从有限的测量中推断其系统特性，以分析和控制其动态行为。本文提出了一种基于多频涡流传感的多目标传感方法，用于加工过程中物理场的重建、几何参数的测量以及表面和亚表面缺陷的检测。本研究还将建立薄壁板加工过程的动力学模型，并开发有效的方法，通过有限的位移测量实时重建薄壁板的位移、应变和应力场，以监测和控制分布参数系统的动力学。与传统的单频涡流传感器不同，多目标传感系统可以自适应合成相邻线圈之间相对高分辨率的涡流图，并结合适当的频率，同时确定位移、厚度和电导率。分布式电流源（DCS）建模方法用于传感器的设计和分析，效率高，计算量小。预计在零件制造过程中，从多目标传感器和场重建算法中重建多物理场（电、磁、位移、力、应变和应力）的能力不仅将提供对几个关键因素(如材料机械性能、边界几何/夹紧约束和阻尼系数)对热/加工引起的残余应力的影响，这些应力通常是薄壁产品变形的主要原因，但也将为振动抑制提供必要的基础。

项目成果

Machine Perception Based on Eddy Current for Physical Field Reconstruction of Conductivity and Hidden Geometrical Features

• DOI: 10.1109/tii.2019.2910857
• 发表时间: 2019-04
• 期刊: IEEE Transactions on Industrial Informatics
• 影响因子: 12.3
• 作者: Min Li;Kok-Meng Lee

An Online Tool Temperature Monitoring Method Based on Physics-Guided Infrared Image Features and Artificial Neural Network for Dry Cutting

基于物理引导红外图像特征和人工神经网络的干切削刀具温度在线监测方法

• DOI: 10.1109/tase.2018.2826362
• 发表时间: 2018-10-01
• 期刊: IEEE TRANSACTIONS ON AUTOMATION SCIENCE AND ENGINEERING
• 影响因子: 5.6
• 作者: Lee, Kok-Meng;Huang, Yang;Lin, Chun-Yeon
• 通讯作者: Lin, Chun-Yeon

State-space Model and Analysis of Motion-induced Eddy-current based on Distributed Current Source Method

• DOI: 10.1109/aim.2019.8868829
• 发表时间: 2019-07
• 期刊: 2019 IEEE/ASME International Conference on Advanced Intelligent Mechatronics (AIM)
• 作者: Bingjie Hao;Xiaoshu Liu;Kok-Meng Lee;Kun Bai

Eddy-Current Dynamic Model for Simultaneous Geometrical and Material Parameter Measurements of Magnetic Materials

用于同时测量磁性材料几何和材料参数的涡流动态模型

• DOI: 10.1115/dscc2018-9211
• 发表时间: 2018
• 期刊: Proceedings of the ASME 2018 Dynamic Systems and Control Conference
• 作者: Hao, Bingjie;Lee, Kok-Meng;Bai, Kun
• 通讯作者: Bai, Kun

Distributed Current Source Method for Modeling Magnetic and Eddy-Current Fields induced in Biological Object

用于模拟生物物体中感应的磁场和涡流场的分布式电流源方法

• DOI: 10.1109/aim.2019.8868493
• 发表时间: 2019
• 期刊: Proceedings of the International Conference on Advanced Intelligent (AIM 2019
• 作者: Lin, Chun-Yeon;Lee, Kok-Meng;Chen, Yuan-Liang;Huang, Shih-Cheng
• 通讯作者: Huang, Shih-Cheng