Modeling And Analyse Of Microscopic Geometic Morphology Of Spiral Bevel Gear In Gear Grinding Process

Surface roughness is an important physically parameter that showing the grind-processing quality, and is also an important indicator of geometric surface morphology. Due to the foreign technology confidentiality, currently there is no information published about the calculating formula for grind-processing surface roughness Ra value of spiral bevel gears. At the same time, the extremely complicated teeth profile of the spiral bevel gears causes many variables involved and makes it difficult to measure and control the Ra value in real-time, which had restricted the degree of productive automation.Firstly, this paper introduces the modeling methods in grinding process and summarizes the domestic and international status quo of spiral bevel gears in grind-processing, and briefly introduces researching situation on the surface roughness at present stage. The experimental conditions and experimental program of the surface roughness of grinding spiral bevel gear tooth is also introduced in this paper. The method of orthogonal experiment is adopted to experimental program, which has reduced the experimental number times and saved the experiment expense, and also obtained the quite precise experimental result.Through dealing with data of the orthogonal experiment table, the result indicates: the influence of velocity ( v_s)of the grinding wheel to the surface roughness ( R_a )is most remarkable , next is the deep tooth feed( a_f), and the last by the wheel feed rate( v_f).The relationship among the surface roughness and all kinds of machining conditions is complicated and nonlinear, and also is affected by many unknown factors. The method of the surface roughness of the grinding spiral bevel gear tooth surface with the combine of the computer technology and artificial neural networks has been adopted in this paper. Namely when modeling, the correlation function in NN toolbox of MATLAB is used to establish the BPNN whose inputs are three factors in grinding( v_s, v_f,a_f)and outputs is Ra. Based on the experimental data, the BPNN is trained and tested.。It shows that the BPNN can get a more precise measuring result of grinding forces in grinding spiral bevel and hypoid gears. At the last part of the paper, an assumption closer to the actual structure wheel is established on the basis of forefathers. Through establishing the mathematical model of wheel and tooth surface, grinding processing is simulated by using the simulative techniques of digital computer, the morphology of grinding wheel and tooth surface are simulated by using Matlab language. Based on the studying of the grinding wheel structure, the obtaining of random distribution of abrasive wheel is an important factor of formatting low roughness to the grinding spiral bevel gear tooth surface.This study identified the relationship among three facts in grinding,the structure of grinding wheel and the surface roughness of spiral bevel gear tooth . It is preparation for calculation of the dynamic mesh force and the corresponding dynamic response regularity, and it is also the basis of the establishment of performance model for the real dynamic of gear transmission and of the determining of the model parameters.