A cutting depth and speed control method based on machined surface feature data clustering

Abstract

The present invention discloses a method for controlling cutting depth and cutting speed based on clustering of machining surface feature data; the method comprises the following steps: first, characterizing the machining surface morphology as a data set of waviness and roughness, then correlating the parameters of the cutting depth and cutting speed of the machining process, digitally classifying the stability and instability of the machining system, then using an acceleration sensor to measure the vibration signal of the grinding process system and calculating the parameters of the system working mode, obtaining the natural frequency and damping ratio of the machining system, then comparing the established machining surface feature clustering with a dynamic model associated with system vibration and grinding stability, finally fitting the stiffness and grinding force coefficient of the machining system, selecting the cutting depth and cutting speed within the process range of grinding stability to achieve regional matching of virtual and real combination, and adjusting the cutting depth and cutting speed within the changing area to maximize the machining surface quality and machining efficiency.

Description

Technical Field

[0001] The present invention relates to the field of diamond grinding, and in particular to a cutting depth and cutting speed control method based on machined surface feature data clustering. Background Art

[0002] Precision grinding is a key process in the precision machining of hard die steel. The stability of the machining process directly impacts both accuracy and efficiency. Grinding process stability is constrained by multiple factors, including the machine tool structure, grinding wheel condition, and workpiece structure and characteristics. Current approaches focus on understanding the mechanisms that cause grinding instability, developing dynamic models of the grinding system to determine stable grinding parameters, and optimizing machine tool structure to improve rigidity. However, these approaches have yet to be applied in actual machining processes.

[0003] Modal analysis of precision grinding systems is an important foundation for studying their process performance. Typically, there are two methods, finite element modal analysis, which involves building a finite element model of the system and establishing boundary conditions and loads to solve the problem. Another method is experimental modal analysis, typically using the hammer method. However, both methods examine the dynamic characteristics of the grinding system in a static state, which differs significantly from actual operating conditions.

[0004] The grinding stability region diagram, an important tool for predicting the machining stability of a grinding system, is typically obtained by establishing the system dynamics equation, deriving the system transfer function, and then solving the characteristic equation. The stability boundary in the lobe diagram demarcates the stable and unstable regions and can guide the selection of grinding parameters. However, obtaining the system parameters required to establish the system dynamics equation is cumbersome, making it difficult to apply in actual machining. Summary of the Invention

[0005] The purpose of the present invention is to overcome the shortcomings of poor processing stability and low processing efficiency in diamond grinding of difficult-to-process metals, and to provide a cutting depth and cutting speed control method based on clustering of processing surface feature data.

[0006] In the cutting depth and cutting speed control process of the present invention, the processing surface morphology is first characterized as a data set of waviness and roughness, and then the parameters of the cutting depth and cutting speed of the processing process are associated, and the stability and instability of the processing system are digitally classified. Then, an acceleration sensor is used to measure the vibration signal of the grinding process system and the parameters of the system working mode are calculated to obtain the natural frequency and damping ratio of the processing system. Then, the established processing surface feature clustering is compared with the dynamic model associated with the system vibration and grinding stability. Finally, the stiffness and grinding force coefficient of the processing system are fitted, and the cutting depth and cutting speed are selected within the process range of grinding stability to achieve regional matching of virtual and real combination, and the cutting depth and cutting speed are adjusted within the changing area to maximize the processing surface quality and processing efficiency.

[0007] The present invention is achieved through the following technical solutions:

[0008] A cutting depth and speed control method based on machined surface feature data clustering comprises the following steps:

[0009] Step S1: Grinding is performed on a CNC grinding machine. The metal workpiece is processed according to different cutting depth and cutting speed parameters. The average waviness W of the processed surface of the metal workpiece is measured using a laser confocal measuring instrument. a and average roughness R a ;

[0010] Step S2: The average waviness W corresponding to each set of processing parameters measured in step S1 is a and average roughness R a As a surface morphology feature set, the stability cluster analysis of the machined surface morphology features is carried out, and the machining states are divided into stable and unstable categories according to the results of the cluster analysis;

[0011] Step S3: In the dynamic modeling stage, a dynamic differential equation for the vertical direction of grinding wheel grinding is established. The dynamic grinding force is expressed as the material removal rate. The dynamic equation is Laplace transformed to obtain the transfer function of the system. The denominator of the transfer function is set to 0 to obtain the critical stability equation of grinding. The critical stability equation is solved to obtain the rotation speed N and feed depth h.

[0012] Step S4: In the working modal analysis phase, the grinding vibration acceleration signal of the grinding system is measured under the working state. The characteristic frequency is obtained by replacing the frequency response function of the system with the auto-power spectrum density function of the signal. The damping ratio is obtained by the half-power bandwidth method on the auto-power spectrum density function and substituted into the damping ratio in step 1 to draw a preliminary grinding stability region diagram.

[0013] Step S5: Mapping the classification result in step S2 onto the grinding stability region map obtained in step S2 to obtain the grinding force coefficient, the grinding system stiffness and the final grinding stability region map;

[0014] Step S6: Selecting a machining parameter combination in a stable region from the grinding stability region diagram obtained in step S5 to achieve grinding with maximum machining efficiency while maintaining stable machining surface quality.

[0015] In step S3, the calculation method of the grinding wheel speed N and the feed depth h uses Lyapunov's first discriminant method to solve the critical situation. The calculation method can be expressed as:

[0016]

[0017]

[0018] In the formula, ω is the critical unstable angular frequency of the system, ω n is the natural angular frequency of the grinding system, let λ=ω / ωn ;k' m is the grinding force coefficient, k is the system stiffness coefficient, b is the abrasive grain width, ξ is the system damping ratio, and T is the grinding wheel rotation period. This preferred approach is adopted because diamond wheel grinding of metals is a complex physical and chemical process. The actual relationship between grinding wheel speed N and feed depth h is nonlinear, making it difficult to calculate using empirical formulas. This algorithm can more accurately express the relationship between system parameters and critical machining parameters.

[0019] Substitute the obtained natural frequency and damping ratio of the grinding system into the calculation formula of the grinding wheel speed N and the critical feed depth h to give the grinding force coefficient k' m , the initial value of the stiffness coefficient k, and obtain the preliminary grinding stability region diagram.

[0020] In step S4, the grinding vibration acceleration signal is measured using a piezoelectric acceleration sensor. The average normalized power spectrum density of the acceleration signal is calculated. The peak on the curve is the natural frequency. Two points with an amplitude of 1 / 1.414 of the natural frequency on both sides of the natural frequency of the power spectrum density function are taken to identify the damping ratio.

[0021] In step S1, a diamond grinding wheel is used for grinding, and a 45# steel substrate is brazed with large-grain diamond, namely 16# diamond abrasive grains.

[0022] The diamond grinding wheel processes the workpiece surface through radial feed.

[0023] In step S1 , the size of the metal workpiece to be ground is 20 mm×20 mm×5 mm, and the workpiece material is difficult-to-machine metal, such as die steel, titanium alloy, etc.

[0024] In step S1, the laser confocal measuring instrument is used to measure the three-dimensional contour point cloud of the workpiece processing surface using a laser confocal microscope, and then the point cloud is imported into the analysis software. A straight line with a sampling length of L = 1 mm is drawn on the surface, and the surface waviness W is calculated based on the two-dimensional contour on the line. a and roughness R a .

[0025] In step S1, the roughness R of the grinding experiment with different cutting depth and cutting speed parameters is a and waviness W a The data is recorded as surface quality x (i) =(R ai , W ai ), i = 1, 2…m; Since the processing state needs to be divided into stable and unstable categories, the number of clusters K = 2; the centroid point μ of each stable category is randomly selected j ∈R 2 , j = 1, 2 (1-stable, 2-unstable); for each set of experimental surface quality X(i) Calculate the stability category c to which it belongs (i) , i is the processing parameter group;

[0026] c (i) =argmin i ||x (i) -μ i || 2 (3)

[0027] For each stable category c (i) , recalculate the center of mass μ of the stable category

[0028]

[0029] The piezoelectric acceleration sensor is installed on the workpiece fixture to measure the dynamic characteristics of the grinding system in the working state. In the running state, the modal parameters of the affected machine tool will change. Only the response signal of the structure in the working state needs to be measured, which reduces the damage to the system structure and is more convenient in application.

[0030] By mapping the feature classification results onto the grinding stability region diagram, the grinding force coefficient and the grinding system stiffness can be obtained. At this time, the grinding stability region diagram of the diamond grinding process system is obtained. The grinding stability region diagram curve represents the processing state of the grinding wheel grinding die steel in this study. The area above the curve is the unstable area, and the area below the curve is the stable area. Selecting processing parameters in the stable area can achieve stable and efficient grinding processing.

[0031] Compared with the prior art, the present invention has the following advantages and effects:

[0032] 1. This invention achieves stable and efficient grinding by applying the working modal calculation method to the grinding of difficult-to-machine metals. Compared with traditional modal analysis methods, it has the advantages of simpler procedures and lower costs.

[0033] 2. The present invention does not need to explore the complex mechanism between stable processing and unstable processing. The cluster analysis method is applied to the processing state classification, and it is classified into stable processing and unstable processing according to the intrinsic relationship between the waviness Wa and roughness Ra data of the processed surface.

[0034] 3. Through this method, the present invention enables the grinding process system to achieve a greater material removal rate while ensuring the surface quality as much as possible, thereby improving processing efficiency and achieving high-efficiency and high-surface quality processing of high-hardness metal materials. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 Schematic diagram of the stable and efficient grinding method of the grinding wheel according to an embodiment of the present invention.

[0036] Figure 2 This is a graph showing the surface corrugation profile of the grinding die steel according to an embodiment of the present invention.

[0037] Figure 3 This is a roughness profile curve of the grinding die steel surface according to an embodiment of the present invention.

[0038] Figure 4 This is a graph showing the cluster analysis results of surface waviness and roughness according to an embodiment of the present invention.

[0039] Figure 5 FIG. 4 is a power spectrum density diagram of the grinding vibration acceleration signal according to an embodiment of the present invention.

[0040] Figure 6 This is a diagram of the grinding stability region of an embodiment of the present invention.

[0041] Figure 7 This is a graph showing the contour of the surface corrugation processed according to the actual verification of an embodiment of the present invention.

[0042] Figure 8 This is a roughness profile curve diagram of the processed surface for actual verification of an embodiment of the present invention. DETAILED DESCRIPTION

[0043] The present invention will be described in further detail below with reference to the embodiments and drawings, but the embodiments of the present invention are not limited thereto.

[0044] Figures 1 to 5 Taking single-point diamond grinding of 20mm×20mm×5mm mold steel as an example, a piezoelectric accelerometer was used to measure the grinding vibration signal. The principle of the present invention's stable and efficient diamond abrasive wheel grinding method based on machined surface characterization clustering is explained in detail, thereby verifying the technical effects of the present invention.

[0045] like Figure 1 As shown, the present invention discloses a cutting depth and cutting speed control method based on machining surface feature data clustering, comprising the steps of:

[0046] During the grinding experiment, the single-point diamond grinding wheel used was a 45# steel matrix brazed with large-grain diamonds. The diamond abrasive grain size was 16#. The workpiece dimensions were 20mm x 20mm x 5mm, and the workpiece material was mold steel. The single-point diamond grinding wheel processed the workpiece surface using radial feed, with a single feed depth of h in the Y-axis direction and a total feed depth of 2 times h.

[0047] Morphology characterization stage: The 3D contour point cloud of the workpiece surface is detected by laser confocal microscopy and imported into the analysis software. A straight line with a sampling length of L = 1 mm is drawn on the surface, and the surface waviness W is calculated based on the 2D contour on the line. a and roughness R a ,like Figure 2 , Figure 3 shown.

[0048] Cluster analysis stage: Roughness R of grinding experiments with different process parameters a and waviness W a The data is recorded as surface quality x (i) =(R ai , W ai ), i = 1, 2...m. Since the processing state needs to be divided into two categories, stable and unstable, the number of clusters K = 2. Randomly select the centroid point μ of each stable category j ∈R 2 , j = 1, 2 (1-stable, 2-unstable). For each set of experimental surface mass x (i) Calculate the stability category c to which it belongs (i) , i is the processing parameter group.

[0049] c (i) =argmin i ||x (i) -μ i || 2 (3)

[0050] For each stable category c (i) , recalculate the center of mass μ of the stable category, such as Figure 4 shown.

[0051]

[0052] Dynamic modeling stage: Establish the dynamic differential equation of single-point diamond grinding in the vertical direction, express the dynamic grinding force of single-point diamond with material removal rate, perform Laplace transform on the dynamic equation to obtain the transfer function of the system, set the denominator of the transfer function to 0 to obtain the critical stability equation of grinding, and solve the critical stability equation to obtain the rotational speed N and feed depth h.

[0053] Working modal analysis stage: Measure the grinding vibration acceleration signal of the grinding system under working state, use the auto-power spectrum density function of the signal to replace the frequency response function of the system to obtain the characteristic frequency, and obtain the damping ratio by the half-power bandwidth method on the auto-power spectrum density function, such as Figure 5 Substitute the speed and feed depth into the calculation formula to draw a preliminary grinding stability area diagram.

[0054] Actual processing stage: Mapping the feature classification results to the grinding stability region diagram can obtain the grinding force coefficient and grinding system stiffness. At this time, the grinding stability region diagram of the single-point diamond grinding die steel system is obtained, such as Figure 6The grinding stability region curve represents the processing state of the single-point diamond grinding die steel in this study. The area above the curve is the unstable region, and the area below the curve is the stable region. Selecting processing parameters in the stable region can achieve stable and efficient grinding.

[0055] Select 4 sets of parameters for processing in the stable and unstable regions of the grinding stability region diagram, and use these 4 sets of processing tests as verification. Figure 7 , Figure 8 The specific results are as follows:

[0056]

[0057] The grinding parameter combination in the stability zone not only has better processing quality but also can achieve greater processing efficiency and realize grinding with maximum removal rate as required.

[0058] In summary, the present invention establishes a stable and efficient single-point diamond grinding method without the need to study the complex surface formation principles of diamond abrasive grinding systems. Cluster analysis can be used to classify the characteristics of the machined surface topography, eliminating the need for multiple sensors to operate online, improving machining efficiency and quality while reducing labor and time costs.

[0059] As described above, the present invention can be implemented well.

[0060] The implementation methods of the present invention are not limited to the above-mentioned embodiments. Any other changes, modifications, substitutions, combinations, and simplifications made without departing from the spirit and principles of the present invention should be considered as equivalent replacement methods and are included in the scope of protection of the present invention.

Claims

1. A cutting depth and speed control method based on machined surface feature data clustering, characterized in that: The following steps are involved: Step S1: Grinding is performed on the grinding system of a CNC grinder. The metal workpiece is processed according to different cutting depth and cutting speed parameters. The average waviness W of the processed surface of the metal workpiece is measured using a laser confocal measuring instrument. a and average roughness R a ; Step S2: The average waviness W corresponding to each set of processing parameters measured in step S1 is a and average roughness R a As a surface morphology feature set, the stability cluster analysis of the machined surface morphology features is carried out, and the machining states are divided into stable and unstable categories according to the results of the cluster analysis; Step S3: In the dynamic modeling stage, a dynamic differential equation for the vertical direction of the grinding wheel is established. The dynamic grinding force is expressed as the material removal rate. The dynamic differential equation is Laplace transformed to obtain the transfer function of the grinding system. The denominator of the transfer function is set to 0 to obtain the critical stability equation of grinding. The critical stability equation is solved to obtain the rotation speed N and feed depth h. Step S4: In the working modal analysis phase, the grinding vibration acceleration signal of the grinding system is measured under the working state. The characteristic frequency is obtained by replacing the frequency response function of the grinding system with the autopower spectrum density function of the signal. The damping ratio is obtained by the half-power bandwidth method on the autopower spectrum density function and substituted into the signal in step S3 to draw a preliminary grinding stability region map. Step S5: Mapping the classification result in step S2 onto the grinding stability region map obtained in step S4 to obtain the grinding force coefficient, the grinding system stiffness and the final grinding stability region map; Step S6: Selecting a machining parameter combination in a stable region from the grinding stability region diagram obtained in step S5 to achieve grinding with maximum machining efficiency while maintaining stable machining surface quality.

2. The cutting depth and speed control method based on machined surface feature data clustering according to claim 1, characterized in that: In step S3, the calculation method of the grinding wheel speed N and the feed depth h uses Lyapunov's first discriminant method to solve the critical situation. The calculation method can be expressed as: Where, ω is the critical unstable angular frequency of the grinding system, ω n is the natural angular frequency of the grinding system, let λ=ω / ω n ;k' m is the grinding force coefficient, k is the stiffness coefficient of the grinding system, b is the abrasive width, ξ is the damping ratio of the grinding system, and T is the grinding wheel rotation period.

3. The cutting depth and speed control method based on machined surface feature data clustering according to claim 2, characterized in that: Substitute the obtained natural angular frequency and damping ratio of the grinding system into the calculation formula of the grinding wheel speed N and the critical feed depth h to give the grinding force coefficient k' m , the initial value of the stiffness coefficient k, and obtain the preliminary grinding stability region diagram.

4. The cutting depth and speed control method based on machined surface feature data clustering according to claim 1, characterized in that: In step S4, the grinding vibration acceleration signal is measured using a piezoelectric acceleration sensor. The average normalized power spectrum density of the acceleration signal is calculated. The peak on the curve is the natural frequency. Two points with an amplitude of 1 / 1.414 of the natural frequency on both sides of the natural frequency of the power spectrum density function are taken to identify the damping ratio.

5. The cutting depth and speed control method based on machining surface feature data clustering according to claim 1, characterized in that: In step S1 , a diamond grinding wheel is used for grinding.

6. The cutting depth and speed control method based on machined surface feature data clustering according to claim 5, characterized in that: The diamond grinding wheel processes the workpiece surface through radial feed.

7. The cutting depth and speed control method based on machined surface feature data clustering according to claim 1, characterized in that: In step S1 , the material of the metal workpiece to be ground is die steel or titanium alloy.

8. The cutting depth and speed control method based on machined surface feature data clustering according to claim 1, characterized in that: In step S1, the laser confocal measuring instrument is used to measure the three-dimensional contour point cloud of the workpiece processing surface using a laser confocal microscope, and then the point cloud is imported into the analysis software. A straight line with a sampling length of L = 1 mm is drawn on the surface, and the average waviness W of the surface is calculated based on the two-dimensional contour on the line. a and average roughness R a .

9. The cutting depth and speed control method based on machined surface feature data clustering according to claim 1, characterized in that: In step S1, the average roughness R of the grinding experiment with different cutting depth and cutting speed parameters is a and the average waviness W a The data is recorded as surface quality x (i) =(R ai ,W ai ), i = 1, 2…m; Since the processing state needs to be divided into two categories, stable and unstable, the number of clusters K = 2; the centroid point μ of each stable category is randomly selected j ∈R 2 , j = 1, 2 (1-stable, 2-unstable); for each set of experimental surface mass x (i) Calculate the stability category c to which it belongs (i) , i is the processing parameter group; c (i) =argmin i ||x (i) -m i || 2 (3) For each stable category c (i) , recalculate the center of mass μ of the stable category 10. The cutting depth and speed control method based on machined surface feature data clustering according to claim 4, characterized in that: The piezoelectric acceleration sensor is mounted on a fixture of the metal workpiece.

Images