Weld penetration state and penetration depth real-time prediction method based on visual characteristics of molten pool

Abstract

The invention discloses a weld penetration state and penetration depth real-time prediction method based on visual characteristics of a molten pool, a set of molten pool visual sensing system is designed based on a color CCD, and two-dimensional characteristics of the molten pool can be extracted in real time in the welding process. The area, the length and the width of a molten pool are used as input, the penetration state of a weld joint is used as output, a GMAW penetration state real-time prediction model is established based on a support vector machine (SVM), and experimental results showthat the model can effectively predict the penetration state of the weld joint in the welding process in real time. Similarly, the area, the length and the width of the molten pool serve as input, the weld penetration serves as output, a GMAW weld penetration real-time prediction model is established based on the BP neural network, and experimental results show that the model can effectively predict the weld penetration in the welding process in real time. The design is simple, and operation is easy and convenient; the calculation result of the constructed model is verified, and the accuracyis high.

Description

technical field [0001] The invention relates to a real-time prediction method of welding penetration state and penetration depth based on the visual characteristics of molten pool, and belongs to the technical field of welding fusion. Background technique [0002] During the welding process, the penetration depth of the weld seam is an important parameter that many scholars care about. Therefore, the real-time and accurate prediction and control of the penetration state and penetration depth of the weld seam have always been a hot spot in the field of welding research. Many studies in the early stage focused on detecting the changes of the penetration state and penetration depth of the weld under different welding conditions. [0003] In the prediction of weld penetration state, visual sensing method is the most widely used detection method. In the prior art, there is a simple and flexible visual sensing system that collects multiple frames of keyhole images on the back of ...

AI Technical Summary

Problems solved by technology

However, these detection methods have corresponding shortcomings. The molten pool oscillation method has a delay, and it is difficult to achieve real-time detection.

Ultr

Images