The Research On Intellectualized Process Design And Quality Control Of Spot Welding Of Aluminum Alloys

The spot welding of aluminum alloys has gained more and more application inautomobile, airline and spaceflight industries. For its physical characteristics,aluminum alloys have bad weldability and quality controlability and the quality ofnugget fluctuates according to the process parameters. The design of processparameters and quality control of its resistance spot welding is a ill structural taskwith multi-inputs and multi-outputs, which shows typical non-linearity. It is aeffective way to employ AI to design the parameters and control the process quality.Digital Signal Processing is a new technology that employs numericalimplementation to gathering signals, translating, synthesizing, evaluating andidentification to obtain information with computers or special processing devices.For the high speed and precision and convenient interfaces, DSP is fit to handlecomplex signal processing algorithm and obtain broad application.As the quality control of aluminum alloys resistance spot welding is acomplicated systems engineering, many factors and sub-systems including processparameters design of aluminum alloys resistance spot welding system and processcontrol of resistance spot welding system exert an influence on quality ofaluminum alloy resistance spot welding. This paper developed a intelligentdistributed controlling system of aluminum alloys which, depending on thesupervisor CPU, managed and controlled the distributed micro-controllers inprocess. With powerful ability of man-machine interaction and great capacity forRAM/ROM, it is convenient to build a man-machine interaction system and carryout the main control functions. on the other side, the micro-controller directlycontrols the objects on the line and carry out the functions of bottom control anddata aquisitioon. The supervisor and lower CPU can work independently, whilethey can work together, both them constitute the intelligent system of aluminumalloys resistance spot welding. The supervisor CPU carries out the general controlfunctions, which include process parameters design, general control and store andanalyze the data from the lower CPU. The lower CPU completes the dataacquisition and process control on line. This paper focuses on the intelligentprocess parameters design system (IPPDS) of aluminum alloys resistance spotwelding(AARSW) in the supervisor and the intelligent process control ofaluminum alloys resistance spot welding in the lower CPU, supported by NSFC,the number is 50175048..The IPPDS of aluminum alloys resistance spot welding consists of threeparts, process parameters design, machine learning and spot welding reference .Thetask completed in process parameters design module comprises user input,designing and amending spot welding parameters; And that completed in machinelearning module consists of maintaining databases and rules databases and trainingFNN. In spot welding reference module, it is supplied that the help of relativeknowledge of spot welding including material weldability, electrode and analysison flaws. In process parameters design module, various reasoning methodsincluding case based reasoning (CBR), rule base reasoning and fuzzy neuralnetworks (FNN) are applied to intimate fully the ways that people experts in fieldstry to design spot welding parameters.In the CBR sub-module, the case charactersare analyzed in details according to AARSW process particularity, and by referringto which the strategies of searching and amending are established. In this paper,interior of AARSW process was revealed deeply. Under the limit of the first classof characters, thickness and thermal physical properties of materials were selectedas initial indexes, which increased the flexibility of reasoning process and wasmore similar to the way based on experiences that people experts thought. A fuzzyinference amending algorithm was proposed, in which electricresistivity and yieldstrength are used as inputs and welding current and electrode force as outputs,furthermore which meets interior behavior of AARSW process. All of thatincreased the flexibility and accuracy of CBR.The design conception of reasoning module under the idea of new expertssystem based on FNN, tried to intimate the thinking way of people experts andimprove reasoning capability of the system by integrating FNN and ES. Thismodule, under the limit of current waves and electrode styles, aimed to set up ageneral reasoning module which was able to determine corresponding AARSWprocess parameters in terms of various thermal physical properties of various kindsof aluminum alloys. By analyzing FNN and ES, the idea of integrating traditionalES and FNN was applied to IPPDS of AARSW. According to prosperities ofAARSW parameters, the process of reasoning was divided into several modules byclassifying process parameters and relative intelligent techniques were adopted fordifferent modules. In the subsystem for main parameters solution, welding cycleswere used as one of inputs in sub-networks, sign of strength of welding process, toattain the value of welding current. That applied to complex characters of solutionto IPPDS of AARSW well. Output value of welding current was able to be adjustedin the range of current that power can supply to fit the given power, which canoptimize outputs. The influence of material physical prosperities on weldingprocess was added to solution to welding procedure parameters by materialweldability was served as one of inputs, so it was possible to intimate deeply theway of people experts thinking. And that as material distinguishability took placeof the way to distinguish materials by name in similar system before and increasedreasoning flexibility. TMS LF2407A was used as the chief core of the intelligent controller of RSWinverter for the first time to improve the control precision in the lower system ofthe distributed intelligent control system of AARSW, and the hardware sub-systemwhich comprised man-machine interaction module, IGBT driver module andfailure diagnosis and protection module etc. was developed. The man-machineinteraction module consists of LCD display module and membrane keyboardcarries out the functions of process parameters setup, shift of control modules, as awhole , it is the communication module between users and system. The driver andprotection module of IGBT is the key factor of designing RSW inverter, forwhether RSW inverter does work or not directly depends on the work status ofIGBT, which is the main power translated component. This device operates as anisolation amplifier for these modules and provides the required electrical isolationbetween the input and output with an opto-coupler. Short circuit protection isprovided by a built in desaturation detector. A fault signal is provided if the short...