Tool grinding wheel dressing device and method based on annular array and dynamic multi-scale convolution

Abstract

The invention discloses a tool grinding wheel dressing device and method based on an annular array and dynamic multi-scale convolution, an annular array sensor monitoring device is arranged at the position of a tool grinding wheel, a sensor is packaged in a flange, a pressure-sound dual-signal annular piezoelectric array sensor is used for collecting signals, and the pressure-sound dual-signal annular piezoelectric array sensor is used for monitoring the tool grinding wheel. The problems that a traditional sensor is poor in anti-interference capability, single in detection signal and single in monitoring area are effectively solved. The multi-scale dynamic convolutional network is used for signal processing, double signals of sound and finishing force are fused, precise classification of the abrasion state and abrasive particle collapse positioning are achieved, a radial displacement self-adaptive compensation instruction is generated, and the problems that in the traditional finishing process, shutdown detection efficiency is low, and manual judgment errors are large can be effectively solved.

Description

Technical Field

[0001] The present invention belongs to the technical field of tool grinding wheel dressing, and particularly relates to a tool grinding wheel dressing device and method based on circular array and dynamic multi-scale convolution. Background Art

[0002] In the field of ultra-precision grinding, ceramic-bonded tool grinding wheels are widely used for the forming dressing of diamond rollers, and their dressing accuracy directly affects the surface quality and fatigue life of key parts such as gears and bearings. However, in the prior art, due to the relatively low hardness of ceramic grinding wheels, they are prone to wear during high-speed dressing, resulting in dressing path deviation and tooth profile error accumulation. Traditional monitoring methods have significant defects: Firstly, it relies on a single sensor (such as force or sound signal). For example, although an acoustic emission sensor can capture the microscopic high-frequency signals of abrasive wear, it is easily interfered by the spindle rotation noise, resulting in the effective signals being submerged and making it difficult to distinguish between grinding wheel wear and mechanical vibration. The pressure signal can monitor the change trend of the dressing force, but since the abrasive grains are irregularly distributed and the surface average stress conditions are inconsistent, accurate judgment cannot be made.

[0003] The sensor position is single, and it cannot effectively capture the wear signals in different circumferential regions of the grinding wheel. The sensor pasted on the end face of the base has more signal noise due to its clearance and vibration, and the stator-rotor type sensor on the spindle has a complex structure and a large manufacturing difficulty.

[0004] Secondly, the traditional judgment method relies on the human ear and the human eye, which belongs to off-line shutdown detection. It judges by the size of the dressing contact sound and the visual shape change of the tool grinding wheel, and cannot capture microscopic defects such as local chipping and progressive wear of the grinding wheel in real time. Moreover, the compensation amount is judged manually, resulting in compensation lag and the accumulation of dressing path deviation.

[0005] Thirdly, existing algorithms (such as wavelet packet decomposition combined with SVM) have insufficient generalization ability under complex working conditions and are difficult to adapt to the time-varying characteristics of grinding wheel wear, resulting in the loss of chipping characteristics in the high-frequency band (>500 kHz), redundant noise in the low-frequency band (<50 kHz), and diamond roller noise in the adjacent frequency band (≈300 kHz). The feature extraction is redundant and the calculation is complex, and the multi-modal data fusion ability is poor.

[0006] Therefore, aiming at the problems of weak anti-interference ability, single collected signal, single signal monitoring area of the existing signal acquisition device, and poor generalization ability of the signal processing algorithm, resulting in low monitoring accuracy, it is necessary to develop a new tool grinding wheel adaptive dressing control system. Summary of the Invention

[0007] The purpose of the present invention is to provide a tool grinding wheel dressing device and method based on circular array and dynamic multi-scale convolution for the above-mentioned deficiencies of the prior art, so as to improve the dressing accuracy and efficiency of the tool grinding wheel.

[0008] To solve the above technical problems, the technical solution adopted by the present invention is as follows: A tool grinding wheel dressing device based on a ring array and dynamic multi-scale convolution, comprising a grinding machine, a tool grinding wheel and a diamond roller installed on the grinding machine, and a ring array sensor monitoring device is also provided at the tool grinding wheel; The ring array sensor monitoring device includes a rear end cover of the encapsulation flange, a front end cover of the encapsulation flange and a conductive slip ring. A pressure film sensor and an acoustic emission sensor are also provided inside the front end cover of the encapsulation flange. The signal transmission lines of the pressure film sensor and the acoustic emission sensor are connected to the rotor side of the conductive slip ring through a cable, and the stator side of the conductive slip ring is fixed on the other side of the grinding machine; The front and rear flanges are assembled, pressed and fixed on the tool grinding wheel and installed on the main shaft of the grinding machine to real-time monitor the internal radial pressure and the internal acoustic emission signal in the hole; A straight-hole microchannel for gas exchange is also provided on the outer layer of the pressure film sensor and the acoustic emission sensor.

[0009] An elastic silicone damping layer is also provided inside the acoustic emission sensor, and the elastic silicone damping layer is annularly in contact and arranged inside the acoustic emission sensor.

[0010] The conductive slip ring is also connected with a preamplifier and transmitted to the computer control terminal through a data acquisition card.

[0011] The pressure film sensors are evenly distributed inside the front end cover of the encapsulation flange at intervals of 120°.

[0012] The acoustic emission sensors are evenly distributed inside the pressure film sensors at intervals of 120°.

[0013] A method for a tool grinding wheel dressing device based on a ring array and dynamic multi-scale convolution includes the following steps: (1) Use a digital filter to filter out high-frequency noise >500KHz and low-frequency noise <50KHz, and use differential vibration suppression to suppress ≈300KHz diamond roller noise; Perform sliding smoothing processing on the pressure signal, perform dynamic pulse entropy extraction, and collect sound and pressure signals; (2) Perform short-time Fourier transform on the sound signal and output a feature vector; Perform pressure pulse entropy extraction on the pressure signal, output a feature vector, and fuse it with the sound signal; (3) Perform multi-scale convolution branches, which are divided into three branches: micro wear, medium wear and macro wear, and fuse the branches; (4) Use an attention mechanism for frequency band weighting and de-weighting; (5)The fully connected classification layer outputs grades, namely micro wear grade, medium wear grade, and macro wear grade respectively; (6)Determine the radial compensation amount according to the wear grade; (7)Input the radial compensation amount into the computer to generate updated G-code; (8)Input the updated G-code into the grinding machine control system to adjust the grinding wheel trajectory in real time and achieve adaptive closed-loop control.

[0014] In step (3), Micro wear branch (1*1 convolution kernel): Medium wear branch (3*3 convolution kernel): Macro wear branch (5*5 convolution kernel): Where W n is the convolution kernel weight matrix; b n is the bias term; And the convolution branches are fused and spliced using the following formula: .

[0015] In step (4), the attention mechanism enhances the features of key frequency bands: Where Q and K are the query matrix and the key matrix, which are obtained by transformation; is the feature dimension; is the frequency band attention weight; And the weighted feature is .

[0016] In step (5), the fully connected classification layer outputs grades according to the following formula, Where c is the output grade.

[0017] The beneficial effects of the present invention are: (1) The present invention discloses a tool grinding wheel dressing device and method based on a circular array and dynamic multi-scale convolution. A circular array sensor monitoring device is set at the tool grinding wheel, and the sensor is encapsulated in a flange. A pressure-sound dual-signal circular piezoelectric array sensor is used to collect signals, effectively solving the problems of poor anti-interference ability, single detection signal, and single monitoring area of traditional sensors; a multi-scale dynamic convolution network is used for signal processing, fusing dual signals of sound and dressing force, realizing accurate classification of wear states and positioning of abrasive grain chipping, and generating a radial displacement adaptive compensation instruction, which can effectively solve the problems of low detection efficiency during shutdown and large human judgment error in the traditional dressing process.

[0018] (2) Reduce the shutdown monitoring time of the tool grinding wheel through both structure and algorithm, improve the dressing accuracy and efficiency of the diamond roller, and further improve the processing quality of high-value-added parts such as aerospace gears and precision reducer gears.

[0019] (3) Integrate an elastic silicone damping layer to isolate the vibration noise of the spindle, and the straight-hole microchannel realizes physical cooling through gas exchange, solving the problems of poor anti-interference ability and thermal noise influence of the traditional monitoring system.

[0020] (4) Adopt pressure and sound circular array sensors, which are evenly staggered and distributed at 120° intervals, covering the entire circumferential area of the grinding wheel, solving the problems of single signal and insufficient area coverage of traditional single sensors.

[0021] (5) Extract wear grade features of different branches through different convolution kernels, use the attention mechanism to weight the key frequency bands, and enhance feature partitioning, solving the problem of weak generalization ability of the existing wavelet packet decomposition and support vector machine algorithms. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 is a schematic diagram of the overall device structure of the present invention; Figure 2 is a cross-sectional view of the circular array sensor monitoring device of the present invention; Figure 3 is a layout schematic diagram of the circular array sensor; Figure 4 is a flow control diagram of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0023] The following specific embodiments illustrate the implementation manners of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification.

[0024] Please refer to Figure 1It should be noted that the structures, proportions, sizes, etc. shown in the drawings of this specification are only used to cooperate with the content disclosed in the specification for those familiar with this technology to understand and read, and are not used to limit the limiting conditions for the implementation of the present invention. Therefore, they do not have substantial technical significance. Any modification of the structure, change of the proportional relationship, or adjustment of the size, without affecting the effects that the present invention can produce and the purposes that can be achieved, should still fall within the scope that can be covered by the technical content disclosed in the present invention. At the same time, the terms such as "upper", "lower", "left", "right", "middle", and "one" cited in this specification are only for the convenience of clear narration and are not used to limit the scope for the implementation of the present invention. The change or adjustment of their relative relationship, without substantial change in the technical content, should also be regarded as the scope in which the present invention can be implemented.

[0025] The present invention provides a tool grinding wheel dressing device and method based on a ring array and dynamic multi-scale convolution, as Figures 1 to 4 shown.

[0026] The tool grinding wheel dressing device based on a ring array and dynamic multi-scale convolution includes a grinding machine 1, a tool grinding wheel 4 and a diamond roller 5 installed on the grinding machine 1, and a ring array sensor monitoring device is also provided at the tool grinding wheel 4; the ring array sensor monitoring device includes a rear end cover 11 of the encapsulation flange, a front end cover 12 of the encapsulation flange and a conductive slip ring 17, and a pressure film sensor 13 and an acoustic emission sensor 19 are also arranged inside the front end cover 12 of the encapsulation flange. The signal transmission lines of the pressure film sensor 13 and the acoustic emission sensor 19 are connected to the rotor side of the conductive slip ring 17 through a low-noise cable 16, and the stator side of the conductive slip ring 17 is fixed on the other side of the grinding machine! to solve the problem of wire winding caused by the rotation of the main shaft.

[0027] The front and rear flanges are assembled and fixed on the tool grinding wheel 4 by screwing and installed on the grinding machine main shaft 3 through bolts 18 to real-time monitor the internal radial pressure and internal acoustic emission signal in the hole.

[0028] A straight-hole microchannel 15 for gas exchange is also arranged outside the pressure film sensor 13 and the acoustic emission sensor 19. The main shaft rotates at high speed for gas exchange to conduct physical cooling, reducing the influence of high temperature on the pressure-sensitive material and thermal noise interference; and the conductive slip ring 17 is also connected with a preamplifier 8 and transmitted to the computer control terminal 10 through a data acquisition card 9. The pressure film sensors 13 are evenly distributed at intervals of 120° inside the front end cover 12 of the encapsulation flange, and the stress changes generated when the tool grinding wheel 4 contacts the diamond roller 5 are collected through piezoelectric signals to characterize the pressure changes under different dressing amounts; the acoustic emission sensors 19 are evenly distributed at intervals of 120° inside the pressure film sensors 13, and the abrasive wear and chipping sound signals during the dressing process are directly monitored by being integrated inside the tool grinding wheel 4.

[0029] An elastic silicone damping layer 20 is further provided inside the acoustic emission sensor 19, and the elastic silicone damping layer 20 is disposed in a circular contact inside the acoustic emission sensor 19. The elastic silicone damping layer 20 is in direct contact with the acoustic emission sensor 19 to isolate the rotation noise of the grinding machine spindle 3 and the vibration noise of the grinding machine.

[0030] In the present invention, a digital filter is used to filter out high-frequency noise > 500KHz and low-frequency noise < 50KHz, and differential vibration suppression is used to suppress the diamond roller noise of ≈ 300KHz; the pressure signal is smoothed by sliding, dynamic pulse entropy extraction is performed, and sound and pressure signals are collected; the short-time Fourier transform is performed on the sound signal to output a 64-dimensional feature vector; pressure pulse entropy extraction is performed on the pressure signal to output a 16-dimensional feature vector, and they are fused into an 80-dimensional feature vector.

[0031] A method for a tool grinding wheel dressing device and method based on a circular array and dynamic multi-scale convolution includes the following steps: (1) Use a digital filter to filter out high-frequency noise > 500KHz and low-frequency noise < 50KHz, and use differential vibration suppression to suppress the diamond roller noise of ≈ 300KHz; Smooth the pressure signal by sliding, perform dynamic pulse entropy extraction, and collect sound and pressure signals; (2) Perform short-time Fourier transform on the sound signal to output a feature vector; Perform pressure pulse entropy extraction on the pressure signal to output a feature vector, and fuse it with the sound signal; (3) Perform multi-scale convolution branches, which are divided into three branches: micro wear, medium wear, and macro wear, and fuse the branches; Micro wear branch (1*1 convolution kernel): Medium wear branch (3*3 convolution kernel): Macro wear branch (5*5 convolution kernel): Where W n is the convolution kernel weight matrix; b n is the bias term; And the convolution branches are fused and spliced using the following formula: .

[0032] (4) Use the attention mechanism for frequency band weighting and downweighting; Where Q and K are the query matrix and the key matrix, which are obtained by transformation; is the feature dimension; band attention weight; and the weighted feature is .

[0033] (5) The fully connected classification layer performs level output, and the outputs are respectively the micro wear level, the medium wear level, and the macro wear level; the fully connected classification layer performs level output according to the following formula, where c is the output level.

[0034] (6) Determine the radial compensation amount according to the wear level.

[0035] (7) Input the radial compensation amount into the computer to generate G-code update.

[0036] (8) Input the updated G-code into the grinder control system to adjust the grinding wheel trajectory in real time and achieve adaptive closed-loop control.

[0037] The present invention uses a pressure-acoustic emission double-layer annular piezoelectric array sensor to collect pressure and sound signals during the grinding process in cooperation with a conductive slip ring equipped with a low-noise cable. The structure is simple, the circumferential signal measurement is uniform and complete, and it can effectively solve the problems of weak anti-interference ability, single collected signal, and single signal monitoring area in the existing detection field; when using a multi-scale convolution-attention network (MSCAN) to fuse the spectrogram of the acoustic emission signal and the dressing force data, extract the grinding wheel wear level map, realize the accurate classification of the wear state and the positioning of abrasive grain chipping, and generate a radial displacement adaptive compensation instruction, which can effectively solve the problems of low detection efficiency during downtime and large human judgment error in the traditional dressing process; reduce the downtime monitoring time of the tool grinding wheel in terms of both structure and algorithm, improve the dressing accuracy and efficiency of the diamond roller, and further improve the processing quality of high-value-added parts such as aerospace gears and precision reducer gears.

[0038] If terms such as "first" and "second" are used in this patent to limit components, those skilled in the art should be aware that the use of "first" and "second" is only for the convenience of describing the present invention and simplifying the description, and these terms have no special meaning.

[0039] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art of this industry should understand that the present invention is not limited by the above embodiments. What is described in the above embodiments and the specification only illustrates the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and these changes and improvements all fall within the scope of the invention claimed. The scope of protection claimed by the present invention is defined by the appended claims and their equivalents.

[0040] In the description of the present invention, it should be understood that the orientation or positional relationship indicated by terms such as "front", "rear", "left", "right", "center", etc. is based on the orientation or positional relationship shown in the drawings. It is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and thus cannot be understood as a limitation to the protected content of the present invention.

[0041] In the ranges disclosed herein, the endpoints and any values are not limited to the precisely recited ranges or values, and these ranges or values should be understood to include values approaching these ranges or values. For numerical ranges, between the endpoint values of each range, between the endpoint values of each range and a single point value, and between single point values, they can be combined with each other to obtain one or more new numerical ranges, and these numerical ranges should be regarded as specifically disclosed herein.

Claims

1. A dressing device for a tool grinding wheel based on a circular array and dynamic multi-scale convolution, characterized in that: It includes a grinding machine, a tool grinding wheel and a diamond roller installed on the grinding machine, and an annular array sensor monitoring device is also provided at the tool grinding wheel; The annular array sensor monitoring device includes a rear end cover of the encapsulation flange, a front end cover of the encapsulation flange and a conductive slip ring. A pressure film sensor and an acoustic emission sensor are also arranged inside the front end cover of the encapsulation flange. The signal transmission lines of the pressure film sensor and the acoustic emission sensor are connected to the rotor side of the conductive slip ring through a cable, and the stator side of the conductive slip ring is fixed on the other side of the grinding machine; The front and rear flanges are assembled, pressed and fixed on the tool grinding wheel and installed on the main shaft of the grinding machine to monitor the internal radial pressure and the internal acoustic emission signal in real time; A straight-hole microchannel for gas exchange is also arranged on the outer layer of the pressure film sensor and the acoustic emission sensor.

2. The dressing device for a tool grinding wheel based on a circular array and dynamic multi-scale convolution according to claim 1, wherein: An elastic silica gel damping layer is also arranged inside the acoustic emission sensor, and the elastic silica gel damping layer is annularly contacted and arranged inside the acoustic emission sensor.

3. The tool grinding wheel dressing device based on the annular array and the dynamic multi-scale convolution according to claim 1, wherein: The conductive slip ring is also connected with a preamplifier and transmitted to the computer control terminal through a data acquisition card.

4. The tool grinding wheel dressing device based on the annular array and the dynamic multi-scale convolution according to claim 1, wherein: The pressure film sensors are evenly distributed inside the front end cover of the encapsulation flange at intervals of 120°.

5. The dressing device for a tool grinding wheel based on a circular array and dynamic multi-scale convolution according to claim 4, wherein: The acoustic emission sensors are evenly distributed inside the pressure film sensors at intervals of 120°.

6. The method of the tool grinding wheel dressing device based on the annular array and the dynamic multi-scale convolution according to any one of claims 1 to 5, characterized in that, It includes the following steps: (1) Use a digital filter to filter out high-frequency noise >500KHz and low-frequency noise <50KHz, and use differential vibration suppression to suppress the ≈300KHz diamond roller noise; Smoothly process the pressure signal by sliding, extract the dynamic pulse entropy, and collect the sound and pressure signals; (2) Perform short-time Fourier transform on the sound signal and output the feature vector; Extract the pressure pulse entropy of the pressure signal, output the feature vector, and fuse it with the sound signal; (3) Perform multi-scale convolution branches, which are divided into three branches: micro wear, medium wear and macro wear, and fuse the branches; (4) Use the attention mechanism to perform frequency band weighting and de-weighting; (5) The fully connected classification layer performs level output, and outputs as micro wear level, medium wear level and macro wear level respectively; (6) Determine the radial compensation amount according to the wear level; (7) Input the radial compensation amount into the computer to generate G code update; (8) Input the updated G code into the grinding machine control system to adjust the grinding wheel trajectory in real time and realize adaptive closed-loop control.

7. The method of the tool grinding wheel dressing device based on the annular array and the dynamic multi-scale convolution according to claim 6, characterized in that, In step (3), Micro wear branch (1*1 convolution kernel): Medium wear branch (3*3 convolution kernel): Macro wear branch (5*5 convolution kernel): where, W n is the convolutional kernel weight matrix; b n is the bias term; And the convolution branches are fused and spliced using the following formula: 。 8. The method of the tool grinding wheel dressing device based on the ring array and the dynamic multi-scale convolution according to claim 6, characterized in that, In step (4), the attention mechanism enhances the features of key frequency bands: Wherein, Q and K are query matrix and key matrix, which are obtained by transformation; is the feature dimension; band attention weight; And the weighted feature is 。 9. The tool grinding wheel dressing device and method based on the annular array and the dynamic multi-scale convolution according to claim 6, characterized in that, In step (5), the fully connected classification layer performs level output according to the following formula, In the formula, c is the output level.

Images