多层单道GMA增材制造成形特性及熔敷尺寸控制

发布时间：2018-06-27 07:59

本文选题：多层单道GMA增材制造 + 成形形貌　；参考：《哈尔滨工业大学》2014年博士论文

【摘要】：熔化极气体保护电弧(Gas Metal Arc， GMA)增材制造(AdditiveManufacturing，AM)采用电弧作为热源、焊丝作为填充材料，根据零件三维模型逐层堆敷直至形成金属零件。当前对GMA-AM过程的研究主要集中在成形工艺的可行性、成形件组织和力学性能上，对成形过程稳定性、成形件表面质量和尺寸精度的研究比较匮乏。本文以多层单道GMA-AM为研究对象，以提高成形过程稳定性和成形精度为目标，深入研究了GMA直接成形金属零件过程的工艺特性、成形件表面质量，并在此基础上对成形尺寸的视觉传感检测及闭环控制等关键问题进行了系统研究，为确保成形过程质量、成形尺寸精度奠定了基础。首先研究了不同熔敷工艺参数下单层单道成形形貌，得到了单层单道良好成形形貌的工艺规范区间。接着探索了熔敷电流、堆积速度、热输入对多层单道成形形貌的影响。研究表明，熔敷电流是影响多层单道成形形貌的决定性因素，电流大于200A时，通常在第二到四层出现熔池表面严重失稳、甚至流淌，成形件形貌呈现坍塌；多层单道获得良好成形形貌的电流区间为100~180A；随熔敷电流的增大，得到良好成形形貌的最大热输入减小。采用二次回归旋转组合设计的方法建立了熔敷工艺参数与熔敷层尺寸的模型关系。该模型可预估不同工艺参数下的熔敷层层宽和层高尺寸，为三维模型分层切片提供了依据。针对GMA-AM过程起弧端熔敷尺寸高且宽、熄弧端高度低的特点，提出了相应的控制策略。封闭路径成形时，相同层间采用起弧和熄弧端搭接的方式弥补高度尺寸缺陷。对非封闭路径零件，一种策略是相邻层间采用交错式堆积方式；另一种策略是相邻层间采用同向式堆积方式，起弧点增大堆积速度、减小送丝堆积速度比，熄弧点降低熔敷电流、堆积速度、增大送丝堆积速度比。进一步研究了多层单道GMA-AM表面质量影响因素。激光视觉传感系统对多层单道表面质量进行测量，，建立了成形件表面质量评价体系。考察了层间温度、定送丝堆积速度比、变送丝堆积速度比等因素对成形件表面质量的影响。结果表明，熔敷层表面质量是层间温度、层高、电流、堆积速度多重因素作用的结果。相同熔敷工艺参数下，层间温度越高，表面质量越差。送丝堆积速度比为定值时，随电流增加，表面质量下降。保持电流和层间温度不变，堆积速度增加，表面质量提高；保持堆积速度和层间温度不变，电流增加，表面质量下降。为实时检测GMA-AM过程熔敷层尺寸，设计了一套用于熔敷尺寸控制的双被动视觉传感系统，可同时获得熔敷层宽度和焊枪到熔敷层表面的高度图像。采用高斯滤波去噪、边缘提取、Hough变换拟合等图像处理技术，开发了GMA-AM熔敷层尺寸实时处理算法，实现了熔敷层宽度、高度参数的在线准确检测，为过程建模和控制器设计打下基础。采用单变量阶跃响应传递函数方法建立了工艺参数与熔敷层尺寸的单输入单输出动态模型，分析了系统的时域响应特点。成形过程存在非线性、时滞等特点，简单的数学模型难以精确描述。随机试验数据辨识了堆积速度与熔敷层尺寸的非线性Hammerstein模型，该模型可用于控制系统仿真。针对GMA-AM过程多变量、多干扰的特点，以熔敷层宽度为被控变量，堆积速度为控制变量，设计了单神经元自学习(Proportional SummationalDifferential，PSD)控制器。仿真结果和干扰试验验证了控制器性能。参数自学习PSD控制器在熔敷层定宽度和变宽度控制中均可以获得良好的控制效果。同时为了满足熔敷层高度的稳定性要求，对熔敷层上表面到焊枪喷嘴的距离进行了宏观监测和自适应控制。最后，应用GMA-AM技术成形了航天发动机燃烧室壳体、环形机匣、液体火箭发动机推力室等模拟件，成形件表面质量良好。双视觉传感系统检测和PSD控制器控制结果表明，熔敷层宽度控制精度小于0.4mm。
[Abstract]:In order to improve the stability and the precision of the forming process , the paper studies the process characteristics and surface quality of the forming process by using the multi - layer single GMA - AM as the research object , and then studies the key problems such as the visual sensing and closed - loop control of the forming process .

In this paper , the morphology of single - layer single - channel formed by different melting process parameters is studied . The influence of welding current , deposition rate and thermal input on the morphology of multi - layer single - channel forming is obtained . The results show that the welding current is the decisive factor affecting multi - layer single - channel forming topography .
the current range of the multi - layer single track to obtain the good forming topography is 100 - 180A ;
With the increase of the welding current , the maximum thermal input of good forming topography is reduced .

The model relationship between the parameters of the welding process and the size of the cladding layer is established by the method of quadratic regression rotation combined design . The model can predict the width and height dimension of the cladding layer under different process parameters , and provide the basis for the layered slicing of the three - dimensional model .

In the process of GMA - AM , a corresponding control strategy is proposed for the characteristics of high arc end cladding size and low arc end height . When the closed path is formed , the arc - extinguishing and arc - extinguishing ends of the same layer are used to make up for the height - dimensional defects .
the other strategy is that the adjacent layers adopt the same - direction stacking mode , the arc - starting point increases the stacking speed , the wire - feeding stacking speed ratio is reduced , the arc - quenching point reduces the deposition current and the stacking speed , and the wire - feeding stacking speed ratio is increased .

The influence factors of the surface quality of multi - layer single - channel GMA - AM are studied . The effect of laser vision sensing system on the surface quality of the multi - layer single - channel GMA - AM is studied . The results show that the surface quality of the layer is the result of multiple factors such as inter - layer temperature , high - layer , current and stacking velocity .
the bulk speed and the interlayer temperature are kept unchanged , the current is increased , and the surface quality is reduced .

In order to detect the size of the cladding layer of GMA - AM process in real time , a set of dual passive vision sensing system is designed to control the width of the cladding layer and the height of the welding gun to the surface of the cladding layer . The real - time processing algorithm of GMA - AM is developed by using Gaussian filter denoising , edge extraction , Hough transform fitting and so on . The online accurate detection of the width and height parameters of the cladding layer is realized , which lays the foundation for the process modeling and controller design .

A single - input single - output dynamic model of the process parameters and the size of the cladding layer is established by a single - variable step - response transfer function method , and the time - domain response characteristics of the system are analyzed . The nonlinear and time - delay characteristics of the forming process are difficult to accurately describe . The stochastic test data identifies the nonlinear Hammerstein model of the stacking velocity and the size of the cladding layer , and the model can be used for the control system simulation .

Based on the characteristics of multi - variable and multi - interference in GMA - AM process , a single neuron self - learning ( PSD ) controller is designed with the width of the cladding layer as the controlled variable and the stacking velocity as the control variable . The simulation results and the interference test have verified the controller performance .

Finally , by using GMA - AM technology , the simulation parts such as combustion chamber shell , annular casing and liquid rocket engine thrust chamber are formed by GMA - AM technology . The surface quality of the former is good . The results of the double vision sensing system and the control of PSD controller show that the control accuracy of the width of the cladding layer is less than 0.4mm .
【学位授予单位】：哈尔滨工业大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：TH16

【参考文献】