Search

CN-122018430-A - Method for improving gear machining precision

CN122018430ACN 122018430 ACN122018430 ACN 122018430ACN-122018430-A

Abstract

The invention discloses a method for improving gear machining precision, which comprises the steps of cleaning a cutter matrix, modifying the surface and eliminating stress; constructing a dynamic coupling error field model, and dynamically compensating through a feedforward feedback composite strategy; performing dynamic balance grading calibration and active cutting vibration suppression of the cutter; constructing a digital twin model of a physical processing system, and forming closed-loop control through virtual-real data linkage and iterative optimization; according to the invention, the multi-source errors are fused and predicted through the dynamic coupling error field model, so that the self-adaptive fusion of physical constraint and data driving is realized, and the technical problems that the existing error compensation model is limited to a single physical mechanism and the multi-source error coupling is difficult to accurately predict are solved; through the deep coordination of error compensation, process control and digital twin technology, the dynamic adaptation and full-period closed-loop regulation and control of each flow are realized, and the technical problems that the coordination of each link is poor, the parameters are fixed and the adaptation to the change of the working condition is impossible in the prior art are solved.

Inventors

  • GAO NIPING
  • WANG HEYI
  • YANG JIACHENG

Assignees

  • 陕西国防工业职业技术学院

Dates

Publication Date
20260512
Application Date
20260224

Claims (10)

  1. 1. The method for improving the machining precision of the gear is characterized by comprising the following steps of: s1, cleaning, surface modification and stress relief treatment are carried out on a cutter matrix, and machining precision deviation caused by deformation and defects is avoided from a base material layer; S2, constructing and applying a dynamic coupling error field model based on a physical information neural network, carrying out fusion prediction on the geometric error, cutting force, temperature and cutter abrasion loss of a machine tool, and carrying out dynamic compensation through a feedforward feedback composite strategy; s3, implementing dynamic balance calibration and active cutting vibration suppression of the cutter, and weakening the influence of dynamic interference on tooth surface accuracy; S4, constructing a digital twin model of the physical processing system, and forming closed-loop control through virtual-real data linkage and iterative optimization.
  2. 2. The method for improving gear machining accuracy according to claim 1, wherein the step S2 includes the steps of: S21, establishing a partial differential equation set for describing thermal deformation and force-induced deformation in the system processing process, and taking the partial differential equation set as a model physical kernel; s22, constructing an attention mechanism neural network, and inputting time sequence data comprising the curvature of a cutting path and the instantaneous material removal rate; s23, dynamically adjusting the output weights of the physical kernel and the attention mechanism neural network according to the processing state through the self-adaptive fusion module, and fusing the error prediction value output by the physical kernel and the error prediction value output by the attention mechanism neural network. And S24, outputting a dynamic error field distribution diagram which is in a future time domain and covers the tooth surface based on the fusion result.
  3. 3. The method of improving gear machining accuracy of claim 1, wherein the dynamic compensation is calculated by the following formula: σ(t)=σ 0 (t)-k(T,F)·[d·ΔG(t)+e·ΔF(t)+f·ΔT(t)] Wherein, sigma (T) is the error after compensation, sigma 0 (T) is the original error, d, e, F are the weight coefficients of the geometric error, the cutting force error and the temperature error, and delta G (T), delta F (T) and delta T (T) are the real-time error amounts of the geometric error, the cutting force and the temperature.
  4. 4. The method for improving gear machining precision according to claim 2, wherein the step S2 is preceded by the further steps of: Preprocessing the multi-field error source data, removing abnormal values, training and optimizing the model, inhibiting over-fitting through an iterative algorithm, and verifying the model compensation accuracy under various processing working conditions.
  5. 5. The method for improving gear machining precision according to claim 2, wherein the dynamic weight adjustment comprises the steps of dynamically adjusting output weights of a physical kernel and an attention mechanism neural network, wherein the step of enabling the output of the physical kernel to be higher than a first weight value to ensure modeling precision in a stable cutting stage when the cutting force and vibration signal fluctuation quantity are lower than a preset threshold value, and the step of enabling the attention mechanism neural network to be higher than a second weight value in a preset time when the sensor detects that the cutting force or vibration signal suddenly changes to exceed the threshold value.
  6. 6. The method for improving machining accuracy of gears according to claim 1, wherein said active suppression of cutting vibration comprises the steps of: S31, collecting vibration signals in the processing process in real time; S32, carrying out spectrum analysis on the vibration signal, and identifying key vibration frequency bands affecting tooth surface accuracy; S33, dynamically adjusting damping control parameters based on analysis results, and pertinently restraining key vibration frequency bands; S34, adjusting machining parameters in a linkage mode, checking the inhibition effect, and ensuring that the vibration amplitude is in an accuracy allowable range.
  7. 7. The method for improving gear machining accuracy according to claim 6, wherein the performing spectral analysis on the vibration signal includes the steps of: S321, filtering the original vibration signal to remove environmental interference noise; S322, acquiring frequency amplitude distribution of the vibration signal by adopting a fast Fourier transform algorithm; S323, setting an amplitude threshold according to the gear machining precision requirement, and locking a high-frequency vibration frequency band exceeding the threshold.
  8. 8. The method for improving gear machining accuracy according to claim 6, wherein the dynamically adjusting damping control parameters based on the analysis result comprises the steps of: S331, matching initial damping parameters according to the locked vibration frequency band, and starting targeted inhibition regulation; S332, calculating the vibration suppression rate and judging whether the vibration suppression rate meets the preset requirement; s333, if the inhibition rate does not reach the standard, fine tuning damping parameters through an iterative algorithm until the vibration amplitude is reduced to an allowable range; S334, recording optimal parameters and storing the optimal parameters into a database.
  9. 9. The method for improving gear machining accuracy according to claim 1, wherein the step S4 includes the steps of: s41, constructing a digital twin model covering machine tools, workpieces, fixtures and environmental elements; s42, predicting a future processing precision change trend based on the long-term and short-term memory network model; S43, feeding back the online detected tooth surface error data to the twin model, and iteratively optimizing the cutting parameters and the error compensation coefficients.
  10. 10. The method for improving gear machining accuracy according to claim 9, further comprising, after said step S43, dynamic calibration of a digital twin model, comprising the steps of: s44, calculating residual errors of the actual measurement value of the key physical quantity and the twin simulation value in real time; s45, when the residual error continuously exceeds a preset residual error threshold value and exceeds preset times, judging that the twin body is misaligned, and triggering a calibration flow; and S46, the calibration flow is used for preferentially adjusting the dynamic parameters of the machine tool and the abrasion state parameters of the cutter in the twin body through a reverse identification algorithm, so that the matching degree of the simulation value and the actual measurement value is restored to be within a preset residual error threshold value.

Description

Method for improving gear machining precision Technical Field The invention relates to the technical field of gear machining, in particular to a method for improving gear machining precision. Background The gear is used as a core component of a mechanical transmission system, the machining precision of the gear directly determines the transmission efficiency, the noise level and the service life of equipment, and the requirements of the high-end equipment field on the precision of the gear are improved to the ISO3-4 level. The existing gear machining precision regulating and controlling method mostly adopts a sectional type regulating and controlling strategy, has obvious technical bottlenecks, and is difficult to meet the requirements of high-precision and large-batch machining. On the one hand, the existing error compensation model is limited to a single physical mechanism, such as thermal deformation or geometric error or pure data driving modeling, and lacks of physical constraint and data self-adaptive fusion capability. On the other hand, the error compensation and the processing technology control and digital twin technology have poor synergy, the vibration suppression and the tool dynamic balance calibration are mostly controlled by adopting fixed parameters, and are disjointed with the error compensation flow, so that the change of working conditions can not be dynamically adapted. Disclosure of Invention Aiming at the defects in the prior art, the invention aims to provide a method for improving the gear machining precision so as to solve the problems in the prior art. In order to achieve the above purpose, the present invention adopts the following technical scheme: The method for improving the gear machining precision comprises the following steps: s1, cleaning, surface modification and stress relief treatment are carried out on a cutter matrix, and machining precision deviation caused by deformation and defects is avoided from a base material layer; S2, constructing and applying a dynamic coupling error field model based on a physical information neural network, carrying out fusion prediction on the geometric error, cutting force, temperature and cutter abrasion loss of a machine tool, and carrying out dynamic compensation through a feedforward feedback composite strategy; s3, implementing dynamic balance calibration and active cutting vibration suppression of the cutter, and weakening the influence of dynamic interference on tooth surface accuracy; S4, constructing a digital twin model of the physical processing system, and forming closed-loop control through virtual-real data linkage and iterative optimization. Optionally, the step S2 includes the steps of: S21, establishing a partial differential equation set for describing thermal deformation and force-induced deformation in the system processing process, and taking the partial differential equation set as a model physical kernel; s22, constructing an attention mechanism neural network, and inputting time sequence data comprising the curvature of a cutting path and the instantaneous material removal rate; s23, dynamically adjusting the output weights of the physical kernel and the attention mechanism neural network according to the processing state through the self-adaptive fusion module, and fusing the error prediction value output by the physical kernel and the error prediction value output by the attention mechanism neural network. And S24, outputting a dynamic error field distribution diagram which is in a future time domain and covers the tooth surface based on the fusion result. Optionally, the dynamic compensation is calculated by the following formula: σ(t)=σ0(t)-k(T,F)·[d·ΔG(t)+e·ΔF(t)+f·ΔT(t)] Wherein, sigma (T) is the error after compensation, sigma 0 (T) is the original error, d, e, F are the weight coefficients of the geometric error, the cutting force error and the temperature error, and delta G (T), delta F (T) and delta T (T) are the real-time error amounts of the geometric error, the cutting force and the temperature. Optionally, before the step S2, the method further includes: Preprocessing the multi-field error source data, removing abnormal values, training and optimizing the model, inhibiting over-fitting through an iterative algorithm, and verifying the model compensation accuracy under various processing working conditions. Optionally, the dynamic weight adjustment comprises the steps of dynamically adjusting the output weights of the physical kernel and the attention mechanism neural network, wherein the dynamic weight adjustment comprises the steps of giving the physical kernel output a higher weight value than a first weight value to ensure modeling accuracy in a stable cutting stage when the cutting force and vibration signal fluctuation quantity are lower than a preset threshold value, and increasing the attention mechanism neural network output weight to be higher than a second weight value in preset time when t