Estimation of Weld Joint Penetration under Varying GTA Pools
Estimation of Weld Joint Penetration under Varying GTA Pools
复制标题
DOI：
--
复制
发表时间：
2013
期刊：
影响因子：
--
通讯作者：
Y. K. Liu;W. Zhang;Y. M. Zhang
中科院分区：
文献类型：
--
作者：
Y. K. Liu;W. Zhang;Y. M. Zhang
焊接接头熔深的传感和控制是自动化焊接中关注的基本问题。对于完全熔透的熔池，背面传感器可以检测到由背面焊缝宽度指定的接头熔深。然而，由于背面传感器在接触传感器以及焊枪和传感器之间的运动匹配方面的局限性，因此最好使用正面传感器。已经进行了广泛的研究，以使用各种前端传感技术(参考文献)来监控焊接过程。1-8)。提取和解释了不同类型的信息来描述焊接过程的状态。在许多已提出的正面传感方法中，焊接熔池几何形状被认为可以提供对焊接过程状态的有价值的见解。在钨极气体保护焊(GTAW)过程中，焊接缺陷和接头熔深等重要信息包含在熔池表面变形中(参考文献)。9-11)。熟练的焊工可以通过直接查看正面熔池来提取有关焊接接头熔深的信息。这意味着可以开发一种先进的控制系统，通过模拟人类焊工的估计和决策过程来精确控制接头熔透。然而，应该检查正面熔池特征参数与接头熔透之间的相关性，以便于在线熔透监测和准确的GTAW工艺熔透控制。最近，一种创新的基于视觉的GTAW工艺传感系统在肯塔基大学焊接研究实验室开发(参考文献)。12)。因此，可以实时重建三维(3D)焊接熔池表面几何形状。进一步发现，三维熔池表面可以用其宽度、长度和凸度来表征，而不是用一大组三维表面坐标(参考文献)来表征。12、13)。因此，可以使用所提出的优化模型和熔池特征参数以可接受的精度来估计接头熔深。然而，为了控制焊缝熔深，需要调整焊接电流。目前尚不清楚，当熔池变化很大时，这些特征参数是否仍然能够以可接受的精度预测焊接接头熔深。这个问题的答案和这一能力的发展是必须回答/解决的基本问题，以便使用熔池作为反馈信息来控制接头熔深。因此，本文研究了不同全熔透条件下接头熔深与特征参数的相关性，发展了在熔池几何形状变化较大的情况下预测接头熔深的能力。
Sensing and control of the weld joint penetration are fundamental issues of concern in automated welding. For a fully penetrated weld pool, the joint penetration specified by its backside bead width could be sensed by a backside sensor. However, a frontside sensor is preferred because of the limitations of the backside sensor in sensor access and motion match between the welding torch and the sensor. Extensive research has been performed to monitor the welding process using various frontside sensing techniques (Refs. 1–8). Different types of information have been extracted and interpreted to describe the state of the welding process. Among the many proposed frontside sensing methods, the weld pool geometry is believed to provide valuable insights into the state of the welding process. Important information, such as weld defects and joint penetration, are contained in the surface deformation of the weld pool in the gas tungsten arc welding (GTAW) process (Refs. 9–11). A skilled welder can extract information about the weld joint penetration by directly viewing the frontside weld pool. This implies that an advanced control system could be developed to precisely control the joint penetration by emulating the estimation and decision-making process of the human welder. However, the correlation between the frontside weld pool characteristic parameters and joint penetration should be examined to facilitate online penetration monitoring and accurate penetration control of the the GTAW process. Recently, an innovative vision-based sensing system for the GTAW process was developed in the University of Kentucky Welding Research Laboratory (Ref. 12). Three-dimensional (3D) weld pool surface geometry could thus be reconstructed in real-time. It was further found that the 3D weld pool surface could be characterized by its width, length, and convexity instead of a large set of 3D surface coordinates (Refs. 12, 13). The joint penetration could thus be estimated using the proposed optimal model and weld pool characteristic parameters with an acceptable accuracy. However, to control the weld joint penetration, the welding current should be adjusted. It is unclear if these characteristic parameters may still be capable of predicting the weld joint penetration in an acceptable accuracy when the weld pool varies substantially. The answer to this question and development of this capability are fundamental issues that must be answered/resolved in order to use the weld pool as feedback information to control the joint penetration. Hence, this paper studies the correlation of the joint penetration with the characteristic parameters under varying full penetration conditions and develops the capability to predict the weld joint penetration despite large variations in weld pool geometry.