CN-122008077-A - CBN grinding wheel non-contact grinding superalloy taper hole abrasion prediction method

CN122008077ACN 122008077 ACN122008077 ACN 122008077ACN-122008077-A

Abstract

The invention discloses a method for predicting wear of a conical hole of a CBN grinding wheel in a non-contact grinding superalloy, which relates to the technical field of state monitoring and intelligent prediction in a precise grinding process, and mainly comprises the following steps: and calibrating the dynamic contact arc acoustic propagation model by using a system identification method to obtain dynamic geometric parameters, constructing acoustic vibration geometric coupling characteristics by combining with a multi-mode signal in a grinding process, correcting the acoustic vibration geometric coupling characteristics by using a maximum mean difference algorithm and a self-adaptive window algorithm, training a wear prediction model by using the corrected acoustic vibration geometric coupling characteristics, and carrying out regression analysis on the corrected acoustic vibration geometric coupling characteristics by using the trained wear prediction model to obtain a prediction result. By implementing the grinding wheel abrasion prediction method for high-temperature alloy taper hole grinding, the adaptability, the robustness and the sensitivity of abrasion state characterization under complex variable working conditions can be improved.

Inventors

SUN HAO
PENG FANGYU
LI TENG
WANG HANMEI
HU ZHIXING
WAN KUN

Assignees

武汉数字化设计与制造创新中心有限公司
中国航发南方工业有限公司

Dates

Publication Date: 20260512
Application Date: 20260107

Claims (10)

1. The non-contact type wear prediction method for grinding the conical hole of the superalloy by using the CBN grinding wheel is characterized by comprising the following steps of: S1, acquiring a multi-mode signal in a grinding process, wherein the multi-mode signal in the grinding process comprises a high-frequency stress wave signal, a low-frequency vibration signal and an air sound signal; s2, calibrating the dynamic contact arc acoustic propagation model by using a system identification method to obtain a calibrated dynamic contact arc acoustic propagation model; S3, obtaining dynamic geometric parameters by using the calibrated dynamic contact arc acoustic propagation model, and constructing sound vibration geometric coupling characteristics according to the multi-mode signals and the dynamic geometric parameters in the grinding process; S4, correcting the sound vibration geometric coupling characteristic by using a maximum mean difference algorithm and a self-adaptive window algorithm to obtain a corrected sound vibration geometric coupling characteristic; and S5, carrying out regression analysis on the corrected geometric coupling characteristics of the sound vibration by using the trained abrasion prediction model to obtain abrasion loss, abrasion trend and residual life estimation of the grinding wheel.
2. The CBN grinding wheel non-contact grinding superalloy taper hole wear prediction method according to claim 1, wherein the calculation formula of the dynamic contact arc acoustic propagation model is: , , ; Wherein, the A dynamically varying contact arc length; a square root function; Is the radius of the grinding wheel; Is the taper angle of the conical hole; The depth is the real-time grinding depth; attenuation of multiple reflections of the high-frequency acoustic emission wave on the hole wall; the diffusion attenuation of air sound in the conical space is realized; And The attenuation coefficient of the high-frequency acoustic emission wave reflected by the hole wall for a plurality of times and the diffusion attenuation coefficient of the air sound in the conical space are respectively obtained.
3. The CBN grinding wheel non-contact grinding superalloy taper hole wear prediction method as in claim 1, wherein the dynamic geometry parameters include dynamic contact arc length, taper angle, and grinding depth.
4. The CBN grinding wheel non-contact grinding superalloy taper hole wear prediction method according to claim 1, wherein the process of constructing the sound vibration geometrical coupling feature is: extracting wear-sensitive features from the grinding process multi-mode signal, wherein the wear-sensitive features comprise AE frequency band energy, vibration envelope peak value and air sound spectrum centroid feature; And carrying out weighted fusion on the wear-sensitive characteristic and the dynamic geometric parameter to obtain the sound vibration geometric coupling characteristic.
5. The CBN grinding wheel non-contact grinding superalloy taper hole wear prediction method of claim 1, wherein the sound vibration geometric coupling characteristics include energy and kurtosis of sound emission signals, envelope peak and root mean square values of vibration signals, spectral centroid and modulation index of air sound signals, grinding depth, dynamic contact arc and taper angle.
6. The CBN grinding wheel non-contact grinding superalloy taper hole wear prediction method according to claim 1, wherein the process of correcting the sound vibration geometrical coupling feature is: Taking the early groups of sound vibration geometric coupling characteristics in the normal abrasion state as reference distribution; Calculating the maximum mean value difference value between the geometric coupling characteristic of the sound vibration and the reference distribution in the current time window by using a maximum mean value difference algorithm; When the maximum mean value difference value exceeds a preset threshold value for 3 times continuously, and the self-adaptive window algorithm is utilized to detect abrupt change, judging that the significant drifting occurs, eliminating data of the acoustic geometric coupling characteristics with the significant drifting, adding the latest multiple groups of acoustic geometric coupling characteristics, and obtaining corrected acoustic geometric coupling characteristics.
7. The CBN grinding wheel non-contact grinding superalloy taper hole wear prediction method according to claim 1, wherein the wear prediction model is a Light GBM model or a line ridge regression model.
8. The CBN grinding wheel non-contact grinding superalloy taper hole wear prediction method according to claim 1, wherein the wear prediction model is a Light GBM model, super parameters of the Light GBM model are determined by bayesian optimization, the maximum depth of the tree is not more than 8 layers, and the learning rate is not more than 0.1.
9. The CBN grinding wheel non-contact grinding superalloy taper hole wear prediction method according to claim 1, further comprising comparing and verifying according to the grinding wheel wear amount and an offline microscope measurement result, when an absolute value of a prediction error of the grinding wheel wear amount and the offline microscope measurement result in a full life cycle is not more than The prediction error of the remaining life estimate does not exceed the total life And judging that the predicted result is qualified.
10. A system for CBN grinding wheel non-contact grinding superalloy taper hole wear prediction, the system comprising: The grinding process multi-mode signal module is configured to acquire a grinding process multi-mode signal, wherein the grinding process multi-mode signal comprises a high-frequency stress wave signal, a low-frequency vibration signal and an aero-acoustic signal; The dynamic contact arc acoustic propagation model calibration module is configured to calibrate the dynamic contact arc acoustic propagation model by using a system identification method to obtain a calibrated dynamic contact arc acoustic propagation model; The sound vibration geometric coupling characteristic construction module is configured to acquire dynamic geometric parameters by using the calibrated dynamic contact arc acoustic propagation model, and construct sound vibration geometric coupling characteristics according to the grinding process multi-mode signals and the dynamic geometric parameters; The abrasion prediction model increment training module is configured to correct the sound vibration geometric coupling characteristic by utilizing a maximum mean difference algorithm and a self-adaptive window algorithm to obtain a corrected sound vibration geometric coupling characteristic; and the grinding wheel abrasion prediction module is configured to carry out regression analysis on the corrected geometric coupling characteristics of the sound vibration by utilizing the trained abrasion prediction model to obtain the abrasion loss, abrasion trend and residual life estimation of the grinding wheel.

Description

CBN grinding wheel non-contact grinding superalloy taper hole abrasion prediction method Technical Field The invention relates to the technical field of state monitoring and intelligent prediction in a precise grinding process, in particular to a method for predicting wear of a tapered hole of a CBN grinding wheel in a non-contact grinding superalloy. Background The ceramic bond CBN grinding wheel is a key tool for precisely grinding the inner hole of difficult-to-process materials such as high-temperature alloy, and the like, but the abrasion state directly influences the processing precision, the surface quality and the cost. The existing grinding wheel abrasion monitoring technology mainly comprises the following steps: (1) Monitoring methods based on contact sensors, such as acoustic emission sensors, strain gauges, etc., mounted on the spindle or workpiece holder. The method is inconvenient to install, signals are easy to interfere with the whole vibration and structure transmission path of the machine tool, and the signal to noise ratio is low. (2) An indirect evaluation method based on a single signal, such as judgment by spindle motor current or power. The method is insensitive to micron-sized wear and cannot be applied to taper hole grinding in which cutting parameters dynamically change due to geometric changes. (3) The offline measuring method based on machine vision is low in efficiency because the machining is required to be interrupted, and is not suitable for a closed inner hole grinding environment filled with cooling liquid. For grinding of an inner conical hole, a grinding area is closed, the contact arc length of a grinding wheel and a workpiece dynamically changes along with axial feeding, so that the propagation path of sound waves and vibration signals and the attenuation characteristic of the sound waves and the vibration signals change in a nonlinear manner, and the abrasion state cannot be effectively represented by the characteristics based on a fixed threshold value or a single signal. Therefore, there is a need for an online monitoring method for grinding wheel wear that can accommodate complex sound field changes, does not require contact mounting, and can achieve high-precision prediction. Disclosure of Invention The invention aims to provide a CBN grinding wheel non-contact grinding superalloy taper hole abrasion prediction method which can improve adaptability, robustness and sensitivity of abrasion state representation under complex variable working conditions. The invention provides a CBN grinding wheel non-contact grinding superalloy taper hole abrasion prediction method, which comprises the following steps: S1, acquiring a multi-mode signal in a grinding process, wherein the multi-mode signal in the grinding process comprises a high-frequency stress wave signal, a low-frequency vibration signal and an air sound signal; s2, calibrating the dynamic contact arc acoustic propagation model by using a system identification method to obtain a calibrated dynamic contact arc acoustic propagation model; S3, obtaining dynamic geometric parameters by using the calibrated dynamic contact arc acoustic propagation model, and constructing sound vibration geometric coupling characteristics according to the multi-mode signals and the dynamic geometric parameters in the grinding process; S4, correcting the sound vibration geometric coupling characteristic by using a maximum mean difference algorithm and a self-adaptive window algorithm to obtain a corrected sound vibration geometric coupling characteristic; and S5, carrying out regression analysis on the corrected geometric coupling characteristics of the sound vibration by using the trained abrasion prediction model to obtain abrasion loss, abrasion trend and residual life estimation of the grinding wheel. The invention also provides a CBN grinding wheel non-contact grinding superalloy taper hole wear prediction system, which comprises the following modules: The grinding process multi-mode signal module is configured to acquire a grinding process multi-mode signal, wherein the grinding process multi-mode signal comprises a high-frequency stress wave signal, a low-frequency vibration signal and an aero-acoustic signal; The dynamic contact arc acoustic propagation model calibration module is configured to calibrate the dynamic contact arc acoustic propagation model by using a system identification method to obtain a calibrated dynamic contact arc acoustic propagation model; The sound vibration geometric coupling characteristic construction module is configured to acquire dynamic geometric parameters by using the calibrated dynamic contact arc acoustic propagation model, and construct sound vibration geometric coupling characteristics according to the grinding process multi-mode signals and the dynamic geometric parameters; The abrasion prediction model increment training module is configured to correct the sound vibration geometric coupling