Penetration Control Of Gmaw Based On The Characteristic Of Frontal Welding Pool

With the rapid development of manufacturing industry,automatic welding has become a trend to replace manual welding.However,as the first process of thick plate multi-layer and multi-pass welding are still required to complete it by manual welding at present.The main reason is that it is difficult to control the penetration state in the process of backing welding,which has become the bottleneck of the development of backing welding automation.As an important parameter reflecting the quality of backing welding,the penetration state of welding seam is closely related to welding heat input.The adaptive neural fuzzy control technology is adopted to penetration control system modeling,to determine the requirements of welding current and welding speed when the penetration status does not satisfy.One of the necessary conditions for fusion control modeling is the real-time acquisition of fusion state.This paper determines the fusion state based on the passive visual sensing method.So the front to the back of a molten pool characteristics and penetration state dynamic corresponding relationship can provide basic data for penetration automatic control.Based on the passive vision sensing method,this paper conducts the GMAW backing welding test,and analyzes the change rule of front weld pool feature and penetration state when welding parameters change and the dynamic corresponding relationship between them.The penetration control model of GMAW backing welding based on welding current and welding speed is established.The main contents and results are as follows:(1)Depend on the problem that arc light interferes with the molten pool area in GMAW welding proeess,an image process algorithm of molten pool is developed to reliably extract het characteristic parameters of frongtal molten pool area,molten pool length and width.(2)Carry out GMAW backing welding test to adjust welding current and welding speed change step size,and analyze the dynamic change rule and correspongding relationship between front weld pool feature and back melt penetration state when welding parameters change.It can be concluded from the result analysis that the transition time between the characteristic parameters of the front melting pool and the width of the back melting pool decreases with the increase of the step length dynamically adjusted by the welding parameters.The lag time between the penetration state of the back melting pool and the characteristic of the front melting pool is relatively stable,and the lag time is about 0.5s.(3)An adaptive neural fuzzy fusion penetration control model was established,in which the difference between the actural fusion width on the back side and the given fusion width on the back side was taken as the input,and the adjusting of welding current and welding speed was taken as the output.By adjusting the welding current or welding speed,the fusion state on the back side could be controlled within the fusion penetration range.Tests were carried out on the non-penetration,penetration and over-penetration states as the initial penetration states,and the penetration,penetration and over-penetration states as the initial penetration states,and the penetration control model was verified.The test results showed that the welding parameters could be adjusted according to the output of the established penetration control model,and the back penetration state could be accurately controlled within the penetration range.