Study Of Intelligent Control System For Co₂welding Power Based On DSP

CO₂gas shielded arc welding is a arc welding method that use CO₂as protectivegas.And it is a kind of gas shielded welding processes focused onpromoting.Compared with other arc welding, CO₂welding has characteristics of highefficiency, low cost and rust-resistant and low hydrogen.However,CO₂welding hasbeen the issues of metal splashes large, poor weld shape and Parameter adjustmentdifficult,which has been limited by the factors of welding power source and controltechnology.The key to solving this problem lies in the control of arc weldinginverter.But welding inverter power source is a nonlinear,time-varying systems that isdifficult to establish a precise mathematical model.Therefore, intelligent controltechnology is applied to CO₂welding power control system in order to get a better arcwelding inverter dynamic characteristics, effectively inhibit the spatter and improveweld forming.this is the development direction of welding power control system.This topic adopted DSP as the core of the control system hardware circuit,anddesigned the hardware and software of CO₂welding power according to thecharacteristics of the short circuit transfer.The main contents are:1.On basis of understanding working principle of inverter arc welding powersource,taking DSP as the core of the control system,the main circuit of CO₂weldingpower inverter adopted the full-bridge inverter circuit,and the hardware circuit ofcontrol system of arc welding inverter was designed.2.In this paper,the current waveform control method was used to control short-circuit stage,and the fuzzy neural network controller was used to control the arc stagebased on the analysis of short circuit transfer process of CO₂welding.3.The neural network were used to remember fuzzy rules,and a fuzzy neuralnetwork controller was designed,and variable learning rate BP algorithm was used toadjust the weights and threshold.The controller is simulated in MATLAB,and theresults showed that the effect is good. 4.CO₂welding power control system software adopts modular design，whichIncludes the main program module, A/D conversion subroutine module, short-circuitcontrol subroutine module, arc control subroutine module, fuzzy neural networksubroutine module, as well as arc and closing arc subroutine module.5.CO₂welding power simulation model was established,and the control effects offuzzy controller and fuzzy neural network controller were simulated andcompared.The results show that the fuzzy neural network controller has better dynamicperformance.and it is able to meet the CO₂welding power requirements of thedynamic characteristics and control the stability of the arc voltage.6.Finally, the overall system was simulated and studied,and the output filterinductor influence on the system is discussed.The results showed that this system isdesigned to meet the need for CO₂welding power control.In this topic,control system of CO₂welding power based on DSP was studied.thecurrent waveform control method was used to control short-circuit stage,and the fuzzyneural network controller was used to control the arc stage.And a fuzzy neural networkcontroller was designed.Simulation studies have shown that the control system workswell, and has the value to continue to study.