The Study On Control Algorithm Of The Filler Rate In Submerged Are Welding Filled With Non-heated Wire

When the high-strength steel, controlled rolling steel and other strong thermalsensitivity material are welded by submerged arc welding, because of excessiveheat in the molten pool, the base metal is damaged by overheating. Then the grainstructure of heat affected zone becomes coarse, easy to produce the phenomenonof embrittlement. Therefore, the reasonable control for energy distribution ofmolten pool can reduce the overheating damage to the base metal. The submergedarc welding filled with non-heated wire is a technological method that can regulatethe process for the heat distribution of molten pool and reduce overheating damageto the base metal by adding non-heated wire (filling wire without power). For theprocess of submerged arc welding filled with non-heated wire, the reasonablefiller quantity of non-heated wire is a major factor determining the success orfailure for this technological method. The study of the technologicalimplementation method was carried out by a set of devices for the submerged arcwelding filled with non-heated wire. The main research contents include thefollowing points.(1) A set of devices that can realize the technology of the submerged arcwelding filled with non-heated wire was established. The devices include the wirefeeder of non-heated wire and microcontroller control system that can realizetechnology process of the submerged arc welding filled with non-heated wire. Thewire feeder designed by Pro-E solid modeling method can regulate the non-heatedwire insertion position, angle, etc. The filler quantity of non-heated wire wascontrolled by the method of regulating the armature voltage of wire feed motor.The dual digital control system that takes80C196KB microcontroller as the core isapplicable for the process control of the submerged arc automatic welding filledwith non-heated wire by setting the threshold of the molten pool length as ajudgement.(2) The heat distribution mechanism of molten pool in the process ofsubmerged arc welding filled with non-heated wire was studied. The doubleellipsoid moving heat source model was adopted as a heat source model. On thebasis of heat transfer, a3-dimensional transient numerical simulation analysis offinite element model for molten pool thermal field of submerged arc welding filledwith non-heated wire was established. The temperature acquisition system thattakes thermocouple, XMC5000data collection and computer as the core wasestablished. The function of real time data collection, display and storage for the temperature in welding process was achieved by the man-machine interface whichwas built by MCGS software. The compared results between the finite elementmodel simulation output and the measured results show that the simulation resultsof thermal cycle curve and temperature value of the verification regional point arebasically closer to experimental results. Also show that the temperature fieldmodel of molten pool can apply to dynamic analysis of the thermal field in theprocess of submerged arc welding filled with non-heated wire welding. Contrast tothe numerical simulation results when non-heated wire was added and not, the heatdistribution mechanism of molten pool in the process of submerged arc weldingfilled with non-heated wire was verified.(3) A adaptive control method of the filler rate of non-heated wire wasexplored. Firstly, the impact law of the filler quantity of non-heated wire on formfactors, microstructures and mechanical properties of weld metal through theexperiment were studied. Based on the balance equation of molten pool heatdynamic distribution for submerged arc welding filled with non-heated wire, thecontrol equation for the filler rate of non-heated wire was formulated. Secondly,the weld metal shaping, microstructure and mechanical properties was taken as aevaluation index, a mapping relationship between arc voltage U, welding current I,welding speed v and the filler rate of non-heated wire vlwas constructed by a largenumber of process test. Finally, a mapping model that takes U, I, v as the inputvector and vlas the target vector by the pattern recognition method based ongenetic algorithm of BP neural network (GABP) was achieved. The linearcorrelation coefficient between the simulation results and the measured valuecome up to0.99188, which indicates that the established GABP control algorithmcan calculate the filler rate value of the non-heated wire.