Research On Sampling Location Planning And Decomposition Algorithm Of Surface Topography Of Undevelopable Ruled Surface

The application of complex machined surfaces in production and life is increasingly widespread,and undevelopable ruled surfaces are also receiving more attention.The detection planning of the micro-surface topography quality of the sidemilled ruled surface is not yet mature,and the feature decomposition algorithm of the micro-surface topography still needs to be optimized.In this paper,a systematic analysis and research is carried out,and a new improved algorithm for sampling location planning and feature decomposition is proposed.A new sampling position plan for microscopic surface topography quality is proposed.Firstly,the complex geometric features of non-developable ruled surfaces are analyzed.The factors such as twist angle,opening and closing angle and singular position,variable curvature and variable cutting width are studied.The surface probability model of each factor is qualitatively established.Secondly,all the surface probability models are combined to establish a multi-parameter function,and the judgment threshold is determined,and a prediction model of difficult-to-machine position is established,so that the sampling position of the micro-surface topography can be effectively determined.A comprehensive improved algorithm for the decomposition of microscopic surface topography features is proposed.Firstly,the optimization algorithm uses information entropy as the cost function to adaptively find the optimal wavelet packet basis function and obtain the optimal wavelet packet decomposition tree.Secondly,in view of the traditional soft-hard threshold function,a semi-hard threshold function is proposed between the two,and the measured signal is denoised and reconstructed by wavelet semi-hard threshold function.At the same time,based on the complete ensemble empirical mode decomposition of adaptive noise(CEEMDAN)algorithm,some optimization improvements are made.The upper limit frequency is adaptively selected for the added Gaussian white noise,the band-limited Gaussian white noise is obtained,and the energy difference tracking method is selected as the screening iterative stopping criterion of the optimization algorithm.The improved algorithm is used to decompose the reconstructed signal,and the decomposed feature spectrum signal is analyzed again to ensure the decomposition accuracy of the optimization algorithm,so as to pave the way for subsequent error analysis and error traceability.The simulation experiment and the machining measurement experiment were carried out to verify the sampling position planning and feature decomposition optimization algorithm of the proposed microscopic surface topography quality.On the one hand,the simulation of the difficult-to-machine position prediction model for the3 D model of the "S" specimen,and the actual processing and measurement of the "S" specimen,verified the effectiveness of the proposed micro-surface topography quality sampling position planning.On the other hand,the signal is collected at the difficult processing position,and the comprehensive improved algorithm and empirical mode decomposition(EMD),ensemble empirical mode decomposition(EEMD),CEEMDAN and other decomposition algorithms are compared by simulation and measurement experiments.The decomposition results verify the accuracy,feasibility and efficiency of the proposed optimization algorithm.