The penetration status of welding seam is an important index for evaluating the welding forming quality, most of the traditional methods for predicting the penetration status are based on the vision sensing method, by extracting the two-dimensional or three-dimensional geometric features of weld pool, establishing the relationship between geometric features of weld pool and the penetration status. In this paper, a prediction method of Cold Metal Transfer(CMT) penetration status based on the temperature distribution in the local area of weld pool is proposed, this method does not need to obtain the full field visual information of weld pool, and can realize the prediction of CMT penetration status by detecting the temperature distribution in the local area of weld pool. Based on color CCD, the temperature field detection system of weld pool is designed, and the relatively simple temperature measurement equation is fitted through the calibration experiment. Taking the temperature in the local area of weld pool as the input and the penetration status as the output, by using Support Vector Machine(SVM) for classification, establishing the penetration status prediction model, the experimental results show that the model can effectively predict the CMT penetration status.

焊缝的熔深状态是评价焊接成形质量的重要指标，大多数传统的熔深预测方法都是基于视觉传感方法，通过提取熔池的二维或三维几何特征来进行的。 ，建立焊缝几何特征与熔深状态之间的关系。提出了一种基于焊缝局部区域温度分布的冷金属熔深渗透状态预测方法，该方法不需要获取焊缝的全场可视信息，可以实现通过检测焊缝池局部区域的温度分布来预测CMT渗透状态。基于彩色CCD，设计了焊池温度场检测系统，并通过标定实验拟合出较为简单的温度测量方程。实验结果表明，以焊缝池局部温度为输入，以熔深状态为输出，通过支持向量机（SVM）进行分类，建立了熔深状态预测模型，实验结果表明该模型可以有效地预测熔深。 CMT渗透状态。