Font Size: a A A

Compilation Of Non-Stationary Dynamic Cutting Force Spectrum Of CNC Lathe

Posted on:2022-01-20Degree:DoctorType:Dissertation
Country:ChinaCandidate:S X WangFull Text:PDF
GTID:1481306728981569Subject:Mechanical Manufacturing and Automation
Abstract/Summary:PDF Full Text Request
CNC machine tools are an important part of modern manufacturing systems,and reliability has become a key factor restricting their development.Carrying out reliability design and reliability test for the CNC machine tool and its key functional components is an important way to improve the reliability of them,and the load spectrum of the CNC machine tool is an objective basis for guiding the reliability design and reliability test.It has important theoretical value and practical significance to study the method of compiling spectrum of the CNC machine tool.With the improvement of the technical level of CNC lathes and the expansion of the application scope,the load conditions are becoming more and more complex and diverse.The conventional method of using empirical formulas to obtain static cutting force can no longer meet the requirements of compiling the load spectrum;the dynamic cutting force spectrum compiled in the existing literature does not consider the non-stationary changes of the dynamic cutting force caused by tool wear and other factors,that is,the mean value and amplitude value of dynamic cutting force increase with the increase of the cutting time of the tool,which leads the compiled cutting force spectrum cannot fully reflect the real load data.Aiming at the above problems,this thesis systematically studies the compilation method and procedure for nonstationary dynamic cutting force spectrum.For complex working conditions,a method for selecting typical working conditions based on the dynamic characteristics of the load is proposed.The measured non-stationary dynamic cutting force is processed,and the nonstationary generalized extreme value(NGEV)model is used to extrapolate the extreme load.The kernel density distribution model of the mean and amplitude of the cutting force was established,and the dynamic cutting force spectrum of the CNC lathe was constructed.The main research contents of the thesis are listed as follows:(1)A method for selecting typical working conditions based on the dynamic characteristics of the load is proposed.36 sets of cutting load data on CNC lathes are obtained using the built dynamic cutting force test system in this thesis.On this basis,the rainflow counting matrix and interval weighting method are used to obtain the mean value and amplitude of the dynamic cutting force.The calculation model based on exponential type of mean and amplitude is established.According to the load data of 1696 working conditions of the CNC lathe collected in the field test,the mean and amplitude are calculated using the built model,and the distribution diagram of the cutting force is drawn.The cutting force distribution is divided into intervals according to the principle of frequency equidistance.The process parameters corresponding to the average value which is closest to the highest peak in each interval are determined as typical working conditions,and finally 40 sets of typical working conditions of CNC lathes are obtained.After comparative analysis,the typical working conditions selected in this thesis cover a larger,more comprehensive and representative cutting force than the typical working conditions proposed in the existing literature.(2)A noise reduction method for three-dimensional non-stationary cutting force based on fractal dimension is proposed.In view of the large difference in the noise reduction effect of a single noise reduction method on the non-stationary dynamic cutting force at different stages.Considering the characteristics of the three-dimensional cutting load,the Man-Kendall(M-K)and Spearman's Rho(S-R)trend test methods are used to analyze the non-stationarity of the cutting force data,and the non-stationary data is divided into multiple relatively stationary stages.Noise reduction methods such as wavelet threshold,singular value decomposition,and empirical mode decomposition are used to reduce the noise of the cutting force data of multiple relatively stationary stages.The effects of different noise reduction methods are evaluated by comparing the deviates from the average of box-counting dimension of different noise reduction method at each relatively stable stage,and data noise reduction is performed.(3)A quantitative model of the distribution of extreme load for CNC lathes based on adaptive block maximum method(ABMM)and non-stationary generalized extreme value distribution(NGEV)is constructed.Using the generalized extreme value(GEV)distribution method based on the block maximum method(BMM),the extreme value load distribution model of the light-medium-heavy dynamic cutting forceare established in the thesis.The model results show that when the cutting load differs greatly,the number of optimal iblocks obtained by the BMM is also quite different.Therefore,the thesis proposes an extreme value sampling method for non-stationary dynamic cutting force based on the adaptive block maximum method(ABMM).A quantitative model of NGEV for extreme load is constructed and it is used to extrapolate the extreme load of the lathe.(4)A kernel density estimation method based on goodness-smoothness comprehensive evaluation is proposed,which realizes the frequency extrapolation of cutting load.The thesis uses the multi-index estimation goodness test method and the smoothness test method for kernel density curve based on envelope curve to comprehensively evaluate the kernel density estimation results of the cutting force distribution under different bandwidths,and the optimal kernel density estimation results are obtained.Compared with the cutting force distribution model constructed by the existing mixed Weibull distribution method,the cutting force distribution model constructed in this thesis based on the kernel density estimation method can obtain higher estimation accuracy.On this basis,the established optimal mean and amplitude kernel density distribution model and extrapolation coefficients are used to extrapolate the load frequency of cutting force.The two-dimensional spectrum of the mean and amplitude of the cutting force is obtained.And using the mixed congruence method,the numbered load frequency is randomly selected,and finally the random programing load spectrum of CNC lathe cutting force is established.Through the research of this thesis,a method for compiling the non-stationary dynamic cutting force spectrum of CNC lathe is proposed,which provides a theoretical basis for the subsequent work of carrying the reliability design and reliability test of the entire CNC lathe and its key subsystems.
Keywords/Search Tags:CNC lathe, cutting force spectrum, typical working conditions, fractal dimension, non-stationary generalized extreme value distribution, kernel density estimation
PDF Full Text Request
Related items