Study On The Molten Pool Shape Of TIG Welding With Magnetic Control Based On Image Processing Technology

With the development of science and technology, welding automation and intellectualization has become an important developing direction of welding technology. As a gradually improving welding technology,Magnetron-welding technology can effectively improve the mechanical properties of the welds and the welding quality overall by interfering the arc, drop and the molten. The main purpose of study on magnetron-welding is to explorer the mechanism of liquid metal flow in molten under the electromagnetic force, in order to lay the foundation for improving the welding mechanism under the effect of magnetic and promoting the application of magnetron-welding.The research taking TIG welding as study subjects, the high speed camera system was used to obtain the real-time image of the molten pool edge grayscale change obviously under the different welding parameters and the magnetic parameters. In order to obtaining the a clear pool image by using Matlab software, the digital image processing such as median filtering, image enhancement, edge detection, edge fitting have being used. Contrast the measured weld width and the calculated weld width after image processing, the maximum error is 0.09 mm, error control within 3%, the result shows that the image processing algorithm is feasible, and then set up the molten pool form geometry parameters to analyze the characteristic of the molten pool shape under the different magnetic parameters.The result shows: Under the action of longitudinal magnetic field,the charged particles in welding arc is doing high speed rotary motion under the action of the lorentz force, plasma power, thermal expansion drive, lead to the welding arc expand in the surface of the molten pool and the distribution curve of temperature spread, then the range of metal under the heat source in the surface of the welding workpieces grow, increase the metal quantity which above the melting point. So in macroscopic view, the molten pool increase, namely the maximum width, the maximum length, area and perimeter of molten pool increased with the magnetic parameters increased.Penetration is one of the important parameters to show the welding quality. Established a penetration forecasting model by using BP nerve network can provide beneficial exploration to describe the relationship between the Surface morphology parameters of molten pool images and the welding quality. Based on this reason, this paper using welding speed, excitation current, welding current,the calculated weld width after image processing as input and the predict values of penetration as output to establish a prediction model by using BP nerve network. The welding experiment indicated the maximum error between the measured penetration value and prediction value is 0.0254 mm, error control within 5%,it shows that the model can predict the penetration values under the different parameters of the TIG welding with magnetic control.