Pipeline welding is a very important process in the construction of oil pipeline. The traditional welding technology has been replaced by half or fully automatic welding process gradually. The development of welding equipment with the function of weld seam tracking has become an urgent need to develop pipeline automatic welding technology. In this paper, such key technologies as signal acquisition and processing, deviation extraction, real-time control are studied.Establish the swing arc sensor static theory model, namely the relationship between input (the height of the welding torch) and output (the welding current) based on constant wire feeding and voltage slow drop characteristics of MIG welding system, according to the working principle of the swing arc sensor. Through multiple linear regression method with welding experimental data, derive the practical model of arc sensor, then linear process the relationship between welding current and welding torch height approximately.Derive the integral difference AS equation of the deviation e by the left and right integral difference method. Aiming to the problem of linearize the relationship height and current has larger error when H<10mm or H>35mm or the accurate arc sensing mathematical model is difficult to establish, this paper presents a method to establish model and optimize the integral differential deviation extraction algorithm based on BP neural network, establish and verify the model through the experimental data. Make welding experiments of different groove angles with different lateral deviation and of different height deviation. Collect welding current signal and offline process non-short current signal and short current signal by using two different kinds of filtering methods, then compare the actual deviation and theoretical deviation solved by â–³S equation of e,provide support for follow-up tracking control.Design the real-time weld seam track system by using the R232C communication between PLC and Lab View. First Lab View reads the torch swing position information to judge how to process the current signal, start left integral or right integral or solve the difference, use equation to solve the deviations, then transform the deviation value into PLC recognition data and write it in PLC to adjust the welding torch swing, eliminate the deviation. |