The Process Simulation Of Solid-state Laser Precision Cutting Machine

Laser cutting technology include a series of characters such as the widely cutting range,the rapidly cutting speed, the narrowed cutting cap, the nicely cutting quality and the largelycutting flexibility, all of which has been made it widely used in modern industry. With theextensive application of industrial technology and the introduction of CAD/CAM, lasercutting technology is moving in the direction of automation. Now a days, many of theautomotive industry have been brought in automatic laser cutting line to improve productquality. In the production process, the study of laser cutting processing is essential, which isconnected to the product quality. Based on some study results, this paper do the research ofthe laser cutting processing simulation which is combine with the neural network knowledge.In this paper, the experiments have been made in the KJG150300500w YAG metalcutting machine. At this machine, the adjustable parameters include laser output voltage U,laser pulse width Th, pulse frequency f and laser table’s movement speed v (ie, processingspeed v). In the test, the roughness quality metrics is slit roughness Ra. On the experiments, itmainly do the following studies:Firstly, Summary the current situation of laser cutting development and research at homeand abroad, and lead to the purpose and significance of this research, and put the required useof research methods and tools.Second, do a detailed theoretical explanations of laser cutting technology. With thecombining of the relevant laser cutting technology, summarize the cutting processing factorsand the affecting parameters of cutting process and influence law of those parameters.Thirdly, Based on the parameters analysis, find out the relationship of voltage U, widthTh, frequency f and processing speed v with the surface roughness Ra through single factorexperiment. Then design4factors4level orthogonal design schemeL1(446)and find out thesignificant of factors and the contribution of slit roughness through intuitive and varianceanalysis of test results. then find out the regression equation with surface roughness and eachparameters through the quadratic general rotary unitized design, and use the results to carryon error fitting and error prediction.Finally, use neural network to carry the text results on error fitting and error prediction ineach of the BP network and RBF Network. Then compare the error analysis results of theregression equation, BP net work and RBF network, and find out the smallest error precisionin the three method. Then use Simulink to modeling and simulation on the smallest errorprecision. Through these studies, this paper draws pictures of the influence figure of surfaceroughness Ra with voltage U, width Th, frequency f and processing speed v, get the optimalparameters of this test, and the smallest error precision simulation model. Those studies notonly effect on the processing of predict section roughness based on the given parameters, butalso have some significance on automatically selecting parameters base on given sectionroughness.