Online Sensing Of Aero-engine Shaft Deformation And Feed Speed Optimization

Aero-engine shaft is an important load-bearing component.The certain type of aero-engine shaft studied in this thesis is a typical thin-walled slender shaft with low rigidity.In the cutting process,generally smaller process parameters are used to reduce the deformation of the shaft,but the processing efficiency is low.Based on the "mechanism + data" method,an offline stiffness regression model and an online turning force prediction model are established and combined in this thesis to quickly and accurately sense the real-time deformation of the slender shaft.The fuzzy control method is used to optimize the feed speed to improve the processing efficiency while ensuring that the processing error does not exceed the set threshold.An offline stiffness regression model is established based on the finite element substructure analysis method and support vector regression algorithm.The unit where the cutter-workpiece engagement area is located is regarded as a substructure,and the third-order stiffness matrix at each contact is solved by using ANSYS APDL programming.Based on the support vector regression algorithm,regression models are established for different elements of the third-order matrix,and the genetic algorithm based on the elite retention strategy is used to optimize the model hyperparameters.Finally,the effects of stiffness simulation error and stiffness regression error on the prediction of elastic part deformation are analyzed through simulation verification and theoretical derivation.An online turning force prediction model is established based on the principle of spindle drive and linear regression algorithm.According to the spindle drive equation under the nocutting process and in the cutting process,the spindle current-tangential cutting force model is established,and the complete cutting force prediction model is obtained according to the empirical relationship between the three-direction cutting force.In order to realize the rapid online sensing of part deformation,the offline stiffness regression model is combined with the online cutting force prediction model.The results of the constant-feed turning experiment show that the diameter error caused by the predicted part deformation is consistent with the change trend of the measured diameter error,and the largest proportion reaches 82.47%.The part deformation prediction value at the section where the largest diameter error occurs in the constant feed turning experiment is used as a threshold,and is compared with the real-time part deformation prediction value,as the input of the fuzzy controller.And fuzzy optimization rules for feed speed is formulated based on the processing experience of field technicians.The experimental results of segmented variable feed cutting show that the machining efficiency is increased by 10.92% under the premise of ensuring the machining accuracy.The online monitoring software and hardware system of the machining process is built,and the functions of online sensing of part deformation and feed speed optimization are integrated.Communication with the CNC system is established to obtain the real-time position coordinates of the tool,and current signals are collected through current sensors and Beckhoff equipment.Software function modules and display modules are developed based on C#multithreading technology and Any CAD API respectively.The construction of software and hardware platforms at the processing site of the aero-engine shaft and turbine guide is conducive to the process decision-making of the on-site processing personnel.