Research On In-situ Measurement Technology Of Shell Thickness Based On Laser Trigonometry

There are a large number of thin-walled shell components in launch vehicle,whose wall thickness is too thin or too thick could reduce the structural strength or the payload of the rocket respectively.In order to assure the processing quality,it is necessary to measure the three-dimensional distribution of the shell thickness.However,among the existed shell thickness measurement methods,the offline method represented by the laser triangulation brings difficulty to accurately compensate the thickness error,and the in-situ method represented by ultrasonic waves is still inefficient due to its technical characteristics of point-by-point and contact measurement.Therefore,it is necessary and inevitability to study the in-situ measurement method of shell thickness based on laser triangulation.The core of the proposed measurement method is to clamp a fixture with two laser displacement sensors on the spindle of the milling machine,which can realize continuous measurement of the three-dimensional distribution of the shell thickness by the motion of the milling machine.Obviously,the diversity of surface morphology and the randomicity of installation location limit the accuracy of the sensor line direction,which becomes the key restrictive factors of the thickness measurement accuracy.Focusing on the topic of improving the accuracy of in-situ measurement method for cylindrical shell thickness in this dissertation,some key technical problems are discussed in a manner of combination of mathematical analysis with experimental investigation.The specific contents are as follows:A geometric model for in-situ measurement of shell thickness based on two laser displacement sensors is established.The influence of installation errors of part,sensors on thickness measurement results is analyzed.Then,an error model of laser displacement sensor considering the geometrical characteristics of the object is established based on the laser scattering characteristics of complex surface,which can be used to qualitative investigate the effects of parameters such as curvature,and inclination angle on single sensor and thickness measurement results.The results show that for two laser sensors who are symmetrically installed on both sides of the measured shell,its principle errors caused by the geometrical characteristics of the object surface such as inclination angle and surface curvature can be compensated or neglected,while the accuracy of the position relationship between the sensors,the milling machine and the measured part is the key factor to determine the accuracy of thickness measurement.To solve the problem of the contradiction between the measuring precision of the thickness and the pose accuracy of parts,a measurement model of cylindrical shell thickness with double offset parameters is established.By using the phase difference calculation method and difference algorithm based on relevant theory,the unknown parameters including the offset of shell installation eccentricity and sensor measurement line are estimated for extracting the corresponding thickness from the measured values of two laser displacement sensors.The uncertainty of the proposed method is analyzed by Monte Carlo method.The simulation results show that the model can improve the in-situ measurement accuracy of shell thickness.Adjusting the collinearity of laser beams is a common core technology in thickness measurement based on two laser displacement sensors.To solve this problem,a laser beam spatial vision positioning method based on double-optical prism is proposed,which can determine the relative spatial attitude between the two sensor measurement lines according to the central point coordinates of the four spots in the image.With the help of Lagrange nonlinear decomposition theory,the image histogram can be decomposed iteratively according to the minimum class variance objective function for extracting the segmentation threshold regarded as the optimal boundary of the spot region and improving the fitting accuracy of the laser spot center.The error of collinearity detection method of laser beam is analyzed.It can be seen that the proposed method can satisfy the necessary accuracy by comparing with collinearity tolerance of sensor in shell thickness measurement.To realize the denoising of shell thickness data without noise frequency information,an improved adaptive filter method based on modal decomposition algorithm is proposed.With some theoretic analyses and improvement for empirical mode decomposition and variational mode decomposition algorithms,they are combined to decompose the thickness data into several intrinsic mode functions.Then,the concept of instantaneous energy probability is presented to select the demarcation points that can judge whether intrinsic mode components are the useful signals or noise information with the Hellinger probability distribution distance,thus the reconstruction and filtering of the signals are realized.The simulation and experimental results of shell thickness data show that the denoising performance of the proposed method is better than that of the existing wavelet filtering method without choosing the basis function and decomposition level.Finally,the collinear detection method of laser beam and the shell thickness measurement method based on double laser displacement sensors are experimentally validated by builting the corresponding experiment platform.The effectiveness of the proposed method is verified by the measurement of the standard gauge block and the eccentricity of the measured parts.