Detection And Processing Of Space Weld Tracking Signal On Rotating Arc

Along with the development of ship manufacturing, petroleum chemical industry, construction and other fields in our country, the demand of large welding structure is bigger and bigger, just relying on manual welding already can’t meet the demand, therefore, automatic welding is also arises at the historic moment.The automation of welding process not only greatly improve the quality of the weld, but also improve the production efficiency. Welding environment is complicated, however, must adopt the mobile robot which is lightweight, easy to carry, convenient for welding in all directions. In addition, because of welding seam in different space position, considering the influence of the molten pool on the welding current is conducive to the recognition of space weld seam tracking signal, and thus improve the welding quality.In this paper,wheeled mobile robot is regarded as experiment platform, PC104as control center, the cross slide block and the car wheels as actuators, rotating arc sensor and hall current sensor as information sensing system, welding power, shielding gas, feeding wire unit and so on as welding system. The main research contents are:the performance test of robot servo system;the collection system of rotating arc sensor signal and the real-time filtering processing of the welding current signal; the identification of deviation information for rotating arc space weld; design tracking controller of welding seam with90°V shape groove flat welding,90°fillet welding, the90°V shape groove vertical welding.Set the hardware platform of rotating arc sensor signal acquisition system, write software that the acquisition system requires, which is convenient to real-time welding current signal acquisition. Because of the complexity of the welding environment,with producing many interference factors in welding process, which has an serious impact on the welding current signal.In order to extract effective groove information, this paper adopt combined filtering method,including the improvement of the minimum value filtering method, the adaptive median filtering method, the weighted average filtering method and mean filtering method,which are used to remove interference, then greatly improve the welding current waveform, truly reflect the welding groove information,and beneficial to exact the deviation information of welding torch. Researh on three space positions’ welding including90°V shape groove flat welding,90°fillet welding, the90°V shape groove vertical welding, the interference of molten pool is a factor that can’t allow to ignore. Respectively discuss processing method of welding current signal that are affected by dissolving apply hot metal under three circumstances, adopt adjustment method to compensate the signal with weighted factor and weighted array,finally eliminate the noise. And for each welding torch position,each has an experiment on torch in middle,on left completely, on right completely, verified the rationality of the adjustment factor and effectiveness of deviation extraction method.For automatic tracking of space weld, this paper design an improved self-learning fuzzy controller, analyse its working principle, adopt modified function to fix the parameters of the fuzzy controller, write control algorithm, and take experiments on space welding seam tracking.The result of experiment demonstrates that the algorithm is desirable, which is advantageous to realize the real-time weld seam tracking and ensure the welding quality.