Robotic 3D Measurement With Applications To Blade-Disk Assemble In Aero-engine

The aero-engine provides power for the aircraft and is one of the core parts of the aircraft,which directly affects the reliability and operational performance of the aircraft.High-performance aero-engines are inseparable from the high-quality design,processing and assembly of their core components.With the continuous development of the design level and fabricating precision of the core components of aero-engine in recent years,the assembly precision has received more and more attention as the final process of aero-engine manufacturing.The blade-disk is the main component of the aero-engine rotor,with complex structure,various sizes,and high quantity of assembly parts,and usually operates under harsh working conditions of high temperature,high pressure and high speed,even small geometric or surface defects may have huge impacts to its performance.However,the existing assembly measurement methods have problems such as lack of robustness and flexibility in data acquisition,lack of rationality in assembly simulation,and lack of correlation in error analysis,which seriously affect the quality and efficiency of blade-disk assembly in aero-engine.Therefore,this paper studies the robotic three-dimensional measurement method of blade disk in aero-engine.The main research work and innovations are as follows:Aiming at the problem of blade scanning data registration caused by sparse workpiece surface features,noise,and empty in scanning data in the process of blade-disk assembly measurement.After analyzing and summarizing the shortcomings of existing registration filtering methods,an outlier removing method based on a consistency search strategy is proposed.The problem of detecting mismatched data converts into a consistency model estimation problem,which provides correct registration point pairs for multi-viewpoints registration of workpiece and provides the calculation basis for the alignment of multi-view scanning data.Finally,the effectiveness of the proposed method is verified by the test of blade scanning data with different outlier ratios and the comparison runs between various outlier removing algorithms.Similar problems to blade-disk,aiming at the problem of rotor-disk scanning multi-view data matching caused by the measurement range limitation of 3D scanning equipment and the complexity of the target with the complex irregular-shaped surface,a multi-view scan registration method is proposed.The matching pairs between multi-view data are constructed using a consistency search strategy,and the multi-view registration results are solved based on the nonlinear optimization algorithm.This method effectively improves the accuracy of multi-view registration and provides high-quality and stable multi-view aligning data for the subsequent three-dimensional error measurement.Finally,the effectiveness of the proposed method is verified by the test of input points with different overlap rates and the comparison of a variety of multi-point cloud joint registration algorithms.Dressing the problem of In the process of blade-disk assembly,the three-dimensional scanning information of a single blade-disk component and the final assembled bladedisk needs to be repeatedly measured.At the meaning time,defective assemblies may appear and need to be re-decomposed and try other batches of assembly parts.To address this low-efficient assembly problem,a three-dimensional surface registration method of assembly surfaces is proposed.By introducing the physical constraints of the assembly and the assembly quality parameters into the registration error function,the method estimates the assembly pose between the blade-disk components with high accuracy,which can provide guidance for the assembly sequence and assembly components selection before assembling,which improves the assembly accuracy and assembly efficiency of the disc.Finally,the experiment verifies the robustness of the algorithm to different input data,and the effectiveness of the proposed method is verified by comparing two key blade-disk assembly indicators.Dressing the problem of the error of the assembly blade-disk,a three-dimensional error distribution optimization method based on the robotic grinding is proposed.The commonly used robot grinding is used as the error trimming method for the bladedisk parts,and the margin distribution requirements of the blade-disk are determined according to the technical characteristics of the robot grinding.Then,the grinding time function based on the robot grinding is constructed,combined with the grinding characteristic curve,and finally The 3D measurement results are optimized,and the effectiveness of the proposed method is verified by experiments.The method effectively shortens the grinding time and improves the grinding efficiency with under basically guaranteed grinding quality,and provides efficient three-dimensional error distribution data for subsequent robot grinding operations.Finally,the experiment verifies the effectiveness of the proposed method through the input test of scanning data of various blade-disk parts and the comparison of the grinding time required for the structure of various point cloud registration algorithms.Finally,taking the assembling blade-disk in aero-engine as three-dimensional assembly measurement objects,we establish a robotic assembly measurement system based on our existing automatic testing line in the lab.Applying the proposed method,scanning data of blade-disk assembly parts are obtained,assembling situation is estimated,and blade-disk errors to each assembly parts are measured.Experiments also verify the effectiveness of the method.