CN-122008076-A - Laser trimming path real-time compensation method and system integrating online monitoring

CN122008076ACN 122008076 ACN122008076 ACN 122008076ACN-122008076-A

Abstract

The invention relates to the technical field of precise grinding of aero-engine blades and discloses a real-time compensation method and a real-time compensation system for a laser trimming path fused with on-line monitoring, wherein the real-time compensation method for the laser trimming path fused with on-line monitoring comprises the steps of collecting monitoring data of a multi-source heterogeneous sensor for laser trimming of a grinding wheel and planning the laser trimming path; the method comprises the steps of preprocessing data, aligning time and space, extracting multi-mode features, selecting features, fusing the multi-mode features, evaluating the current processing state of grinding wheel laser trimming, predicting the processing state of the grinding wheel laser trimming, performing multi-objective compensation optimization on a basic path planning result, performing path correction and smoothing processing, executing a compensation path instruction, monitoring an executing process in real time and evaluating the executing effect, and performing self-adaptive feedback.

Inventors

MEI FENG
YANG YUJIAN
ZHOU GUANGHUA

Assignees

苏州强芯智能装备科技有限公司

Dates

Publication Date: 20260512
Application Date: 20260309

Claims (10)

1. The real-time compensation method for the laser trimming path integrated with the online monitoring is characterized by comprising the following steps of: Collecting monitoring data of a multisource heterogeneous sensor for grinding wheel laser trimming, planning a laser trimming path to obtain a basic path planning result, and carrying out data preprocessing and space-time alignment to obtain standard multisource monitoring data; Based on a standard multisource monitoring dataset, carrying out multi-mode feature extraction and feature selection to obtain a multi-mode key feature set; based on the multi-mode key feature set, carrying out multi-mode feature fusion, and evaluating the current processing state of the grinding wheel laser trimming to obtain a comprehensive evaluation result of the current processing state; Predicting the machining state of the laser trimming of the grinding wheel based on the comprehensive evaluation result of the current machining state, and outputting a comprehensive prediction result; Based on the comprehensive evaluation result and the comprehensive prediction result of the current processing state, performing multi-objective compensation optimization on the basic path planning result to obtain a comprehensive path compensation amount; based on the comprehensive path compensation amount and the basic path planning result, performing path correction and smoothing processing to obtain a compensation path instruction; Executing the compensation path instruction, monitoring the execution process in real time, evaluating the execution effect, and performing self-adaptive feedback to obtain the complete data record and the finishing report of the processing process.
2. The method for compensating for a laser trimming path in real time by fusing on-line monitoring according to claim 1, wherein the collecting the monitoring data of the multi-source heterogeneous sensor of the laser trimming of the grinding wheel comprises: Configuring a multi-sensor monitoring system, wherein the multi-sensor monitoring system comprises a laser displacement sensor, a visual imaging system, an acoustic emission sensor and an infrared temperature sensor; Establishing a calibration curve between a distance measurement value and a real distance by using a standard gauge block for a laser displacement sensor, performing camera internal parameter calibration on a visual imaging system by using a calibration plate, establishing a rotation matrix and a translation vector between a camera coordinate system and a processing coordinate system by hand-eye calibration, testing frequency response characteristics and sensitivity parameters of the acoustic emission sensor by using a known acoustic source, and performing temperature calibration on an infrared temperature sensor according to emissivity characteristics of a measured material to obtain calibration parameters of a multi-sensor monitoring system; The multi-source heterogeneous sensor monitoring data synchronously acquire the monitoring data according to respective sampling frequencies, time stamps are made when the data are acquired, and multi-source original monitoring data are output.
3. The method for real-time compensation of laser trimming path integrated with on-line monitoring according to claim 1, wherein the performing of data preprocessing and space-time alignment comprises: Filtering and denoising the multi-source heterogeneous sensor monitoring data and extracting preliminary features, and outputting preprocessed monitoring data; based on calibration parameters of the multi-sensor monitoring system and the preprocessed monitoring data, carrying out space-time alignment on the multi-source monitoring data, establishing a unified time grid, carrying out time delay compensation on time stamps of the sensors, mapping spatial positions measured by different sensors to a unified processing coordinate system, and generating a standard multi-source monitoring data set.
4. The method for real-time compensation of laser trimming path integrated with on-line monitoring according to claim 1, wherein the performing multi-modal feature extraction comprises: extracting geometric deviation features from the morphology point cloud data, registering the point cloud data with a target trimming molded surface, calculating the normal distance from each measuring point in the point cloud to the target molded surface to generate shape deviation field distribution, extracting global features of deviation, and extracting spatial distribution features of deviation; extracting surface quality features from visual image data, extracting texture feature parameters by adopting a gray level co-occurrence matrix method, extracting the outline boundary of a grinding wheel by edge detection to evaluate the boundary integrity, and analyzing and identifying a burn area by a color space; extracting material removal process characteristics from acoustic emission signals, extracting main frequency components and time domain changes thereof, calculating accumulated energy and energy change rate of signals in each time window, identifying acoustic emission events by setting amplitude threshold values, and counting event characteristics; And extracting heat accumulation effect features from the temperature data, extracting standard deviation of space statistics features, extracting time evolution features of temperature, and extracting heat affected zone features.
5. The method for real-time compensation of laser trimming path integrated with on-line monitoring according to claim 1, wherein the performing multi-modal feature integration comprises: Establishing a feature fusion rule base based on expert knowledge; And carrying out feature fusion and state evaluation by adopting a lightweight model, wherein the state evaluation model adopts a gradient lifting decision tree or a support vector machine, and outputting shape deviation scores, uniformity removal scores, heat accumulation grades, processing stability scores and surface quality grades.
6. The method for real-time compensation of laser trimming path integrated with on-line monitoring according to claim 1, wherein the predicting the processing state of laser trimming of the grinding wheel comprises: Establishing a simplified physical model of material removal, adopting an energy threshold criterion, removing the material when the energy density received by a unit area of a laser action area exceeds a material removal threshold, wherein the removal depth is in a direct proportion relation with the energy density, and predicting the material removal amount; establishing a simplified physical model of heat conduction, dividing the surface of the grinding wheel into a plurality of lumped units by adopting a lumped parameter model, establishing a heat balance equation of each unit, and carrying out numerical solution on the heat balance equation of each unit to predict the evolution of a temperature field; Establishing a simplified physical model of thermal deformation, calculating thermal strain of each position based on temperature field distribution, establishing a stress-strain relation based on elastic modulus and poisson ratio of the material to solve thermal stress distribution, and predicting thermal stress and thermal deformation of the grinding wheel; predicting a state evolution trend by adopting a data driving model; And fusing the prediction results of the physical model and the data driving model to obtain a comprehensive prediction result.
7. The method for real-time compensation of laser trimming path integrated with on-line monitoring according to claim 1, wherein the performing multi-objective compensation optimization on the basic path planning result comprises: Establishing a multi-objective compensation optimization problem, wherein optimization objectives comprise minimizing shape deviation, maximizing removal uniformity, controlling temperature level and keeping execution smoothness, and compensation variables comprise position adjustment quantity of path points, scanning speed adjustment quantity and laser power adjustment quantity; solving a collaborative optimization problem by adopting a hierarchical iteration strategy; Performing feasibility checking and constraint processing of the compensation scheme; The priority of compensation and execution policy are calculated.
8. The method for compensating for a laser trimming path in real time with integrated on-line monitoring according to claim 1, wherein the performing the path correction and smoothing process comprises: generating a compensated path point coordinate sequence, and superposing the position adjustment quantity of each path point in the path on the original coordinates to obtain compensated new coordinates, wherein the position adjustment is performed along the normal direction of the surface of the grinding wheel; smoothing the compensation path, and adopting a B spline fitting method to fit the compensated path point as a control point or a data point to generate a B spline curve; performing dynamic constraint checking and optimization of the path, and calculating the speed, acceleration and jerk of each axis when the path is executed; The trim path is segmented using segmented track generation and look-ahead buffering.
9. The method for real-time compensation of laser trimming path integrated with on-line monitoring according to claim 1, wherein the performing adaptive feedback comprises: continuously collecting monitoring data, evaluating an execution effect, comparing the current actual state with an expected state, and calculating a state tracking error; Implementing a self-adaptive feedback control adjustment compensation strategy, setting a feedback control gain, and performing incremental adjustment according to a state tracking error, wherein the adjustment mode comprises incremental adjustment of compensation quantity, fine adjustment of process parameters and update of prediction model parameters; Realizing closed-loop data flow and system self-learning to form a complete closed loop from monitoring data acquisition to execution control; and establishing an abnormality detection and grading abnormality processing mechanism, and starting the grading abnormality processing mechanism when an abnormal state is detected, wherein the grading abnormality processing mechanism comprises that primary abnormality adopts parameter fine adjustment and enhancement monitoring, secondary abnormality immediately reduces laser power to a safe level, and tertiary abnormality immediately pauses processing and starts a safety protection program.
10. A real-time compensation system for a laser trimming path fused with on-line monitoring, which is used for executing the real-time compensation method for the laser trimming path fused with on-line monitoring according to any one of claims 1 to 9, and comprises the following steps: The multi-sensor monitoring module is used for collecting multi-source heterogeneous sensor monitoring data of grinding wheel laser trimming, planning a laser trimming path to obtain a basic path planning result, and carrying out data preprocessing and space-time alignment to obtain standard multi-source monitoring data; The feature extraction module is used for carrying out multi-mode feature extraction and feature selection based on the standard multi-source monitoring data set to obtain a multi-mode key feature set; the state evaluation module is used for carrying out multi-mode feature fusion based on the multi-mode key feature set and evaluating the current processing state of the grinding wheel laser trimming to obtain a comprehensive evaluation result of the current processing state; The state prediction module predicts the processing state of the grinding wheel laser trimming based on the comprehensive evaluation result of the current processing state and outputs the comprehensive prediction result; the compensation optimization module is used for performing multi-objective compensation optimization on the basic path planning result based on the comprehensive evaluation result and the comprehensive prediction result of the current processing state to obtain a comprehensive path compensation quantity; The path generation module is used for carrying out path correction and smoothing processing based on the comprehensive path compensation quantity and the basic path planning result to obtain a compensation path instruction; And the execution control module is used for executing the compensation path instruction, monitoring the execution process in real time, evaluating the execution effect, and carrying out self-adaptive feedback to obtain the complete data record and the trimming report of the processing process.

Description

Laser trimming path real-time compensation method and system integrating online monitoring Technical Field The invention relates to the technical field of precise grinding of aero-engine blades, in particular to a laser trimming path real-time compensation method and system integrating on-line monitoring. Background Aero-engine blades are used as core components of engines, and the geometric accuracy and surface quality of the aero-engine blades directly influence the performance, efficiency and safety of the engines. In the precise grinding process of the blade, the shape precision and the surface state of the grinding wheel are key factors for determining the processing quality. Along with the continuous improvement of the requirements of the aviation industry on the blade processing precision, the traditional grinding wheel dressing method has difficulty in meeting the processing requirements of high precision and high efficiency. As an advanced grinding wheel dressing method, the laser dressing technology has the advantages of non-contact, high precision, strong controllability and the like, and is widely focused and applied in the field of grinding of aeroengine blades. However, existing laser trimming techniques face many challenges in practical applications. The laser absorptivity and the removal threshold values of different positions are different due to the non-uniformity of the grinding wheel materials, uniform trimming is difficult to achieve by adopting uniform process parameters, dynamic disturbance in the machining process such as machine tool vibration, thermal deformation, material performance fluctuation and other factors can influence the stability of trimming precision, and the heat accumulation effect in the laser trimming process can cause local overheating of the grinding wheel, so that the strength of a bonding agent is reduced, the material structure is changed, even thermal cracks are generated, and the service performance and the service life of the grinding wheel are seriously influenced. Most of the current laser trimming systems adopt an offline path planning and open-loop control mode, and lack the capability of real-time monitoring and dynamic adjustment of the trimming process. The method cannot respond to the state change in the processing process in time, and consistency and stability of finishing quality are difficult to ensure. Particularly, when facing the requirements of complex profile grinding wheels or high-precision trimming, the limitation of the prior art is more remarkable, and the development of an intelligent laser trimming method capable of sensing the processing state in real time and dynamically adjusting the trimming strategy is urgently needed. Disclosure of Invention The invention provides a laser trimming path real-time compensation method and system integrating on-line monitoring, which solve the technical problem that stable and high-precision trimming is difficult to realize in the prior art in the face of the influence of multiple factors such as grinding wheel material non-uniformity, dynamic disturbance in the processing process, heat accumulation effect and the like. The invention provides a laser trimming path real-time compensation method integrating on-line monitoring, which comprises the following steps: Collecting monitoring data of a multisource heterogeneous sensor for grinding wheel laser trimming, planning a laser trimming path to obtain a basic path planning result, and carrying out data preprocessing and space-time alignment to obtain standard multisource monitoring data; Based on a standard multisource monitoring dataset, carrying out multi-mode feature extraction and feature selection to obtain a multi-mode key feature set; based on the multi-mode key feature set, carrying out multi-mode feature fusion, and evaluating the current processing state of the grinding wheel laser trimming to obtain a comprehensive evaluation result of the current processing state; Predicting the machining state of the laser trimming of the grinding wheel based on the comprehensive evaluation result of the current machining state, and outputting a comprehensive prediction result; Based on the comprehensive evaluation result and the comprehensive prediction result of the current processing state, performing multi-objective compensation optimization on the basic path planning result to obtain a comprehensive path compensation amount; based on the comprehensive path compensation amount and the basic path planning result, performing path correction and smoothing processing to obtain a compensation path instruction; Executing the compensation path instruction, monitoring the execution process in real time, evaluating the execution effect, and performing self-adaptive feedback to obtain the complete data record and the finishing report of the processing process. In a preferred embodiment, the collecting the multi-source heterogeneous sensor monitoring data of the