A Prediction Method Of Seam Deviation Based On Magnetic-control Rotating Arc Sensor

There is a big gap of the level of welding automation between our country and developedcountries. Improve the level of welding automation and intelligent is one of the keytechnologies to realize seam auto tracking. But the effect of the common seam trackingmethod in application is still unsatisfactory, the complex and expensive system is limited inpractice, there is no a low-cost and high-precision seam tracking method. An intelligent seamtracking method based on magnetic-control arc sensor as a satisfy requirement of seamtracking way, study it is of great significance. On the basis of the magnetic-control rotatingarc sensor, using the exponential smoothing forecast method and principal componentanalysis method respectively solve the weld deviation, and seam tracking test is carried out,the results show that two kinds of prediction algorithm enhances the precision of seamtracking, for space complex trajectory of weld deviation identification provides a newforecasting method. Main research contents are as follows:1、Designed magnetic-control rotating-arc sensor which could satisfy requirement ofseam tracking system.The two trajectories of magnetic-control rotating arc and magnetic-control swing arcsimulated result show the former effect used in seam tracking was good，and described indetail rotating arc principle of magnetic-control rotating-arc sensor, researched and analysedthe mechanism of magnetic-control rotating-arc sensor, through analysed the simulation resultof electromagnetic-field， designed magnetic-control rotating-arc sensor which satisfyrequirement of seam tracking system.2、Established mathematical model for the magnetic-control rotating arc sensor, andverified the correctness of the model.On the basis of GMAW system mathematical model and HALMOY welding wire meltingmodel, established the mathematical model of magnetic-control rotating-arc sensor, used themodel to simulate variation law of welding current, arc length and welding wire length withtime. Lots of welding tests showed that the simulation waveforms of welding current andactual welding current were consistent, therefore, established the mathematical model wascorrect and feasible.3、 Put forward two kinds of algorithm that Exponential smoothing prediction methodand principal component analysis method, and use them to solve weld deviation respectively.The magnetic-control rotating arc sensing predictive control system consists of fourparts, respectively, internal models, predictive model, reference trajectory and predictive control. Using exponential smoothing prediction method and principal component analysismethod respectively solve the weld deviation, and describes the implementation of specificoperation steps, and through comparing analysis of two kinds of methods that exponentialsmoothing forecast method can meet the demand of seam tracking, and principal componentanalysis method can carry out high precision of seam tracking.4、 Set experiment platform that welding seam tracing system based on themagnetic-control rotating arc sensor.Set experiment platform that welding seam tracing system based on the magnetic-controlrotating arc sensor, realized weld auto tracking. Using the proposed prediction algorithm toimprove the precision of seam tracking, for the space complex trajectory of weld deviationidentification provides a new forecasting method.