Experimental Research And Simulation Of Fiber Laser Cutting On LY12 Aluminum Alloy

Laser cutting is a very popular modern special processing technology,which has some advantages such as narrow kerf width,processing range,high efficiency,high machining accuracy,widely used in industrial production.In recent years,laser cutting technology has a continuous development and aluminum alloy is increasingly used in aerospace and other fields.But due to the reason for the aluminum alloy is low laser absorption rate,laser cutting is less used in the field of aluminum alloy processing.And because of less study in laser cutting technology about aluminum alloy,the study of laser cutting technology of aluminum alloy is very important.First,from the aspects of theory,this paper introduces the laser cutting technology in the machining principle,characteristics,development trend,main influence factors and related content of fiber laser,and summarizes the evaluation index of laser cutting quality.Secondly,according to the predetermined testing scheme,this paper conducts the laser cutting experiment on 2mm thick of LY12 aluminum alloy by G6020 F fiber laser cutting machine and measures the test pieces cut surface roughness,cutting seam width and length of hang slag by using the measuring instrument.Through single factor experiment,respectively,this paper analyzes the influence rule of the factors such laser power,cutting speed and nitrogen pressure on cutting quality.Through the orthogonal experiment scheme,this paper in detail analyzes the influence of the laser power,cutting speed and nitrogen pressure on cut surface roughness,cutting seam width and the length of hang slag.And this paper determines the optimal combination of factors on comprehensive quality by the matrix analysis method.Finally,with the aid of MATLAB neural network toolbox,established laser cutting machining process simulation model based on BP neural network.By using the model,the cut surface roughness,cutting seam width and length of hang slag which under different processing parameters are simulated and predicted.Finally found,this paper's established the BP neural network model has good fitting precision and prediction precision.BP network was optimized by genetic algorithm.And by using the optimized GA-BP network,the cut surface roughness,cutting seam width and length of hang slag of laser cutting are simulated and predicted.Compared to two kinds of network prediction results,GA-BP network has better prediction precision,and has a certain reference value for the prediction of new samples.