Chatter suppression-oriented multi-scale collaborative simulation method and system for thin-walled part cutting

Abstract

The application discloses a kind of thin-walled part cutting multiscale collaborative simulation method and system for chatter suppression, and relates to the field of collaborative simulation.The method of the application comprises S1: obtaining the vector data of cutting force and machining vibration signal to obtain cutting-vibration coupling simulation result;S2: based on the cutting-vibration coupling simulation result, obtain multiscale node stiffness degradation data;S3: based on the multiscale node stiffness degradation data, obtain chatter critical frequency simulation data;S4: based on the chatter critical frequency simulation data, obtain chatter suppression trajectory simulation scheme.The method of the application, by real-time acquisition of cutting force vector and vibration data, combined with node stiffness recursion and workpiece geometry information analysis, fusion modal frequency self-matching identification, realize frequency and processing parameter dynamic correlation, avoid resonance through frequency matching and trajectory adjustment, improve the chatter identification accuracy and control feedback speed, improve the thin-walled part stability guaranteeing ability.

Description

Technical Field

[0001] This invention relates to the field of collaborative simulation technology, specifically to a multi-scale collaborative simulation method for cutting thin-walled parts with chatter suppression and a multi-scale collaborative simulation system for cutting thin-walled parts with chatter suppression. Background Technology

[0002] The field of collaborative simulation technology involves model building, parameter exchange, and coupled computation between multi-physics, multi-scale, and multi-dimensional systems. It aims to break down data barriers between different computing software or modeling tools, improving simulation efficiency and prediction accuracy in product design, process development, and system control. The core of this technology includes heterogeneous modeling collaboration, cross-domain physical coupling, unified temporal and spatial analysis, and multi-scale data mapping, and is widely used in industrial scenarios with high performance and precision requirements, such as aerospace, automotive manufacturing, and machining.

[0003] Among them, the traditional multi-scale co-simulation method for chatter suppression in thin-walled part cutting refers to the problem of chatter easily occurring in parts with high dynamic response characteristics during the cutting process. It relies on the workpiece dynamic characteristics and cutting force model parameters obtained from experiments as input, and predicts the mechanical response through single physical domain modeling. It mainly uses cutting force calculation methods based on empirical formulas, dynamic stability criteria based on frequency domain characteristics, and static structural modal parameter extraction methods for modeling and simulation. It cannot achieve real-time coupling of multi-physical information in the cutting process and dynamic feedback coordination between multi-scale features.

[0004] Traditional solutions focus on static modal parameter extraction and quantitative mechanical models based on empirical formulas, making only unidirectional physical field predictions. They lack dynamic data coupling and structure-process linkage mechanisms, making it difficult to map real-time stiffness changes caused by material removal and to identify the impact of workpiece modal frequencies evolving with processing at different time points. This results in delayed response to chatter-sensitive areas, making it prone to misjudgment in high-dynamic-response processing scenarios. It also fails to identify instantaneous resonance risks, affecting the stability of the process path and the final processing accuracy of thin-walled parts. Summary of the Invention

[0005] In view of this, one objective of the present invention is to provide a multi-scale collaborative simulation method for thin-walled part cutting for chatter suppression, which solves the technical problem that the existing technology has a lag in response to chatter-sensitive areas and is prone to misjudgment in high dynamic response machining scenarios.

[0006] Another objective of this invention is to provide a multi-scale co-simulation system for cutting thin-walled parts for chatter suppression.

[0007] To achieve at least one of the above objectives, the present invention provides the following technical solution:

[0008] One embodiment of the present invention provides a multi-scale co-simulation method for cutting thin-walled parts for chatter suppression, which includes the following steps:

[0009] S1: Obtain vector data of cutting force and machining vibration signal, and perform multibody dynamics coupling calculation based on the vector data and machining vibration signal to obtain cutting-vibration coupling simulation results;

[0010] S2: Based on the cutting-vibration coupling simulation results, obtain the change in material removal volume and the rate of change of cutting force vector at the node, and perform finite element stiffness coefficient update calculation based on the change and the rate of change to obtain multi-scale node stiffness degradation data.

[0011] S3: Obtain the geometric position data of the workpiece at the processing time based on the multi-scale node stiffness degradation data, perform a proportional weighted summation calculation based on the multi-scale node stiffness degradation data and the geometric position data to obtain the structural modal frequency weighting coefficient of the workpiece, obtain the flutter sensitive frequency range based on the structural modal frequency weighting coefficient, and perform a fusion calculation based on the structural modal frequency weighting coefficient and the flutter sensitive frequency range to obtain the flutter critical frequency simulation data;

[0012] S4: Based on the flutter critical frequency simulation data, obtain the critical frequency value, obtain the spindle frequency, and compare the critical frequency value and the spindle frequency using a frequency matching algorithm to obtain a flutter suppression trajectory simulation scheme.

[0013] The multi-scale co-simulation method for thin-walled part cutting aimed at chatter suppression, as described above, specifically includes the following steps in step S1:

[0014] S101: Acquire vector data of cutting force collected by a three-dimensional force sensor and machining vibration signals of workpiece surface nodes collected by a vibration sensor, wherein the machining vibration signals include stress data and displacement data. Based on the vector data, the stress data and the displacement data, an alignment method is used to obtain a surface node stress-displacement dataset.

[0015] S102: Perform multibody dynamics coupling calculations based on the surface node stress-displacement dataset to obtain cutting force vibration coupling response data;

[0016] S103: Based on the cutting force vibration coupling response data, obtain the vibration amplitude and frequency of the nodes during the simulation period. Based on the vibration amplitude and frequency, the stress data and the displacement data, obtain the cutting-vibration coupling simulation results.

[0017] The multi-scale co-simulation method for chatter suppression in thin-walled part cutting, as described above, specifically includes the following steps in step S2:

[0018] S201: Based on the workpiece mesh cell information in the cutting-vibration coupling simulation results, obtain the change in the material removal volume in each mesh cell, and calculate the rate of change of the cutting force vector within adjacent time steps through differential operation to obtain the rate of change of the cutting force vector.

[0019] S202: Based on the rate of change of the cutting force vector and the change in the material removal volume of each node, perform finite element stiffness recursive calculation to obtain the updated node stiffness coefficients;

[0020] S203: Obtain multi-scale stiffness degradation evolution data of thin-walled parts based on the nodal stiffness coefficients, and perform multi-scale coupling calculations based on the multi-scale stiffness degradation evolution data to obtain the multi-scale nodal stiffness degradation data.

[0021] The multi-scale co-simulation method for chatter suppression in thin-walled part cutting, as described above, specifically includes the following steps in step S3:

[0022] S301: Obtain the stiffness degradation value of the node based on the multi-scale node stiffness degradation data, obtain the node response coordinates of the workpiece surface position based on the stiffness degradation value and the workpiece displacement record at the current processing time, and obtain the geometric position data of the workpiece surface based on the node response coordinates.

[0023] S302: Based on the geometric position data and the multi-scale node stiffness degradation data, obtain the real-time displacement vector and stiffness change of the node at the processing time, and perform proportional weighted summation calculation based on the real-time displacement vector and the stiffness degradation data to obtain the structural modal frequency weighting coefficient.

[0024] S303: Based on the structural modal frequency weighting coefficient, obtain the dynamic response path of the workpiece at the node during the processing; based on the dynamic response path, obtain the distribution function curve of the response amplitude and frequency; based on the distribution function curve, obtain the flutter sensitive frequency range; and based on the structural modal frequency weighting coefficient and the flutter sensitive frequency range, perform fusion calculation to obtain the flutter critical frequency simulation data.

[0025] The multi-scale co-simulation method for chatter suppression in thin-walled part cutting, as described above, specifically includes the following steps in step S4:

[0026] S401: Obtain the critical frequency value based on the chatter critical frequency simulation data, obtain the spindle frequency and tool feed trajectory coordinates, and obtain path segment frequency matching data based on the critical frequency value, the spindle frequency and the feed trajectory coordinates;

[0027] S402: Based on the path segment frequency matching data, a frequency matching algorithm is used to obtain the difference between the main shaft frequency and the critical frequency value. Based on the comparison between the difference and a set risk threshold, resonance risk path segment data is obtained.

[0028] S403: Based on the resonance risk path segment data, the flutter suppression trajectory simulation scheme is obtained.

[0029] The multi-scale co-simulation method for chatter suppression in thin-walled part cutting, as described above, further includes:

[0030] S5: Based on the chatter suppression trajectory simulation scheme, obtain the cutting parameters of the path points, and obtain the workpiece roughness and machining efficiency. Based on the cutting parameters, the workpiece roughness and the machining efficiency, obtain a multi-objective optimization function. Based on the multi-objective optimization function, perform the NSGA-II algorithm to obtain a set of co-simulation optimization parameters.

[0031] The multi-scale co-simulation method for thin-walled part cutting aimed at chatter suppression, as described above, specifically includes the following steps in step S5:

[0032] S501: Based on the chatter suppression trajectory simulation scheme, obtain the path point position, obtain the feed direction and cutting parameters, obtain the path spacing and direction offset based on the path point position and the feed direction, obtain the coordinate gradient change based on the path spacing and the direction offset, and obtain the initial cutting parameter coefficient set of the path point based on the coordinate gradient change and the cutting parameters, wherein the cutting parameters include spindle speed, feed rate and depth of cut;

[0033] S502: Obtain workpiece roughness and machining efficiency; obtain roughness fluctuation rate and efficiency response time based on the workpiece roughness and machining efficiency; obtain a multi-objective optimization function based on the roughness fluctuation rate and efficiency response time; obtain a multi-objective function input parameter set based on the multi-objective optimization function and the initial cutting parameter coefficient set of the path point.

[0034] S503: Based on the input parameter set of the multi-objective function, perform the NSGA-II algorithm to obtain multiple solutions. Based on the multiple solutions, calculate the congestion degree and dominance relationship of the solutions to obtain a set of co-simulation optimization parameters.

[0035] Another embodiment of the present invention provides a multi-scale co-simulation system for thin-walled part cutting aimed at chatter suppression, comprising:

[0036] The first module is used to acquire vector data of cutting force and machining vibration signal, and to perform multibody dynamics coupling calculation based on the vector data and machining vibration signal to obtain cutting-vibration coupling simulation results;

[0037] The second module is used to obtain the change in material removal volume and the rate of change of cutting force vector at the node based on the cutting-vibration coupling simulation results, and to perform finite element stiffness coefficient update calculation based on the change and the rate of change to obtain multi-scale node stiffness degradation data.

[0038] The third module obtains the geometric position data of the workpiece at the processing time based on the multi-scale node stiffness degradation data, performs a proportional weighted summation calculation based on the multi-scale node stiffness degradation data and the geometric position data to obtain the structural modal frequency weighting coefficient of the workpiece, obtains the flutter sensitive frequency range based on the structural modal frequency weighting coefficient, and performs a fusion calculation based on the structural modal frequency weighting coefficient and the flutter sensitive frequency range to obtain the flutter critical frequency simulation data.

[0039] The fourth module is used to obtain the critical frequency value based on the flutter critical frequency simulation data, obtain the spindle frequency, and compare the critical frequency value and the spindle frequency using a frequency matching algorithm to obtain a flutter suppression trajectory simulation scheme.

[0040] The multi-scale co-simulation system for thin-walled part cutting aimed at chatter suppression, as described above, further includes:

[0041] The fifth module is used to obtain the cutting parameters of the path points based on the chatter suppression trajectory simulation scheme, and to obtain the workpiece roughness and machining efficiency. Based on the cutting parameters, the workpiece roughness and the machining efficiency, a multi-objective optimization function is obtained. Based on the multi-objective optimization function, the NSGA-II algorithm is performed to obtain a set of co-simulation optimization parameters.

[0042] A third embodiment of the present invention also provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-described well leakage risk warning method.

[0043] The fourth embodiment of the present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the above-described well leakage risk warning method.

[0044] The fifth embodiment of the present invention also provides a computer program product, the computer program product including a computer program, which, when executed by a processor, implements the above-mentioned well leakage risk early warning method.

[0045] Compared with the prior art, the beneficial effects that the at least one technical solution adopted by the present invention can achieve include at least the following:

[0046] The method of this invention is based on the joint acquisition of real-time cutting force vector and vibration data, combined with dynamic nodal stiffness recursion and synchronous analysis of workpiece geometric information, and integrates modal frequency self-matching identification to achieve dynamic correlation between frequency and machining motion parameters. It dynamically tracks structural state changes at each machining moment, avoids resonance risks through frequency matching and real-time trajectory adjustment, strengthens the fusion of information between multiple physical fields during the cutting process, improves the analytical depth of chatter formation mechanism and the timeliness of control measures, makes the overall simulation prediction more compatible with real-time machining conditions, significantly improves the accuracy of chatter risk identification and control feedback speed, and optimizes the ability to ensure the stability of complex thin-walled parts during process development. Attached Figure Description

[0047] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0048] Figure 1 This is a flowchart of the multi-scale co-simulation method for thin-walled part cutting for chatter suppression according to the present invention;

[0049] Figure 2 This is another flowchart of the multi-scale co-simulation method for thin-walled part cutting for chatter suppression according to the present invention;

[0050] Figure 3 This is a flowchart of step S1 in the method of the present invention;

[0051] Figure 4 This is a flowchart of step S2 in the method of the present invention;

[0052] Figure 5 This is a flowchart of step S3 in the method of the present invention;

[0053] Figure 6 This is a flowchart of step S4 in the method of the present invention;

[0054] Figure 7 This is a flowchart of step S5 in the method of the present invention;

[0055] Figure 8 A schematic diagram of a multi-scale co-simulation system for cutting thin-walled parts to suppress chatter.

[0056] Figure 9 Another schematic diagram of a multi-scale co-simulation system for cutting thin-walled parts for chatter suppression;

[0057] Figure 10This is a schematic diagram of a computer device in an embodiment of the present invention. Detailed Implementation

[0058] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0059] Unless otherwise defined, all technical and scientific terms used in this invention have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains; the terminology used in this invention is for the purpose of describing particular embodiments only and is not intended to limit the invention; the terms “comprising” and “having”, and any variations thereof, in the specification and the foregoing description of the invention are intended to cover non-exclusive inclusion.

[0060] The term "embodiment" as used in this invention means that a specific feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described in this invention can be combined with other embodiments.

[0061] The specific term "exemplary" used in this invention means "serving as an example, embodiment, or illustration." Any embodiment illustrated as "exemplary" is not necessarily to be construed as superior or better than other embodiments. Although various aspects of embodiments are shown in the accompanying drawings, the drawings are not necessarily drawn to scale unless specifically indicated otherwise.

[0062] In the description of this invention, "a plurality of" means two or more (including two), unless otherwise explicitly specified.

[0063] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other. The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0064] Figure 1 This is a flowchart of the multi-scale co-simulation method for thin-walled part cutting for chatter suppression according to the present invention.

[0065] like Figure 1 As shown, an embodiment of the present invention provides a multi-scale co-simulation method for cutting thin-walled parts to suppress chatter, which includes the following steps:

[0066] S1: Obtain vector data of cutting force and machining vibration signal. Specifically, obtain vector data of cutting force through a three-dimensional force sensor, collect machining vibration signal captured by a vibration sensor, and perform multibody dynamics coupling calculation based on vector data and machining vibration signal to obtain cutting-vibration coupling simulation results.

[0067] S2: Based on the cutting-vibration coupling simulation results, the change in material removal volume and the rate of change of cutting force vector at the node are obtained. Specifically, based on the cutting-vibration coupling simulation results, the node state changes are analyzed, the change in material removal volume is detected, the rate of change of cutting force vector data between adjacent time steps is calculated, and the finite element stiffness coefficient is updated based on the change and the rate of change to obtain multi-scale node stiffness degradation data.

[0068] S3: Obtain the geometric position data of the workpiece at the processing time based on the multi-scale node stiffness degradation data, perform a proportional weighted summation calculation based on the multi-scale node stiffness degradation data and the geometric position data to obtain the structural modal frequency weighting coefficient of the workpiece, obtain the flutter sensitive frequency range based on the structural modal frequency weighting coefficient, and perform a fusion calculation based on the structural modal frequency weighting coefficient and the flutter sensitive frequency range to obtain the flutter critical frequency simulation data;

[0069] S4: Obtain the critical frequency value based on the flutter critical frequency simulation data, and obtain the spindle frequency. Specifically, the spindle frequency includes the spindle speed and feed rate. Based on the critical frequency value and the spindle frequency, a frequency matching algorithm is used to compare and identify the resonance risk trajectory, thereby obtaining a flutter suppression trajectory simulation scheme.

[0070] The method of this invention avoids resonance through frequency matching and trajectory adjustment, significantly improving the accuracy of flutter risk identification and control feedback speed, and optimizing the ability to ensure the stability of complex thin-walled parts during process development.

[0071] Figure 2 This is another flowchart of the multi-scale co-simulation method for thin-walled part cutting for chatter suppression according to the present invention.

[0072] like Figure 2 As shown, another embodiment of the present invention provides a multi-scale co-simulation method for thin-walled part cutting oriented towards chatter suppression. The method in this embodiment is largely the same as the method in the above embodiment, except that after step S4, the method further includes:

[0073] S5: Obtain the cutting parameters of the path points based on the chatter suppression trajectory simulation scheme. Specifically, based on the path coordinate information in the chatter suppression trajectory simulation scheme, obtain the corrected path point cutting parameter configuration, and obtain the workpiece roughness and machining efficiency. Based on the cutting parameters, workpiece roughness and machining efficiency, obtain a multi-objective optimization function. Based on the multi-objective optimization function, perform the NSGA-II algorithm, and use the path point coordinate offset and cutting parameters as decision variables to perform optimization, and obtain the set of co-simulation optimization parameters.

[0074] This method avoids resonance risks through frequency matching and real-time trajectory adjustment, strengthens the fusion of information between multiple physical fields during the cutting process, improves the analytical depth of chatter formation mechanism and the timeliness of control measures, makes the overall simulation prediction more compatible with real-time machining conditions, significantly improves the accuracy of chatter risk identification and control feedback speed, and optimizes the ability to ensure the stability of complex thin-walled parts during process development.

[0075] The cutting-vibration coupling simulation results include cutting force distribution characteristics, vibration acceleration amplitude variation and energy transfer relationship; multi-scale nodal stiffness degradation data include elastic modulus variation, damage distribution state and local damage factor; chatter critical frequency simulation data include chatter critical frequency range, resonance sensitive node information and main mode frequency characteristics; chatter suppression trajectory simulation scheme includes resonance risk avoidance path, frequency safety range and trajectory adjustment point set; co-simulation optimization parameter set includes surface roughness value, machining efficiency index and path point cutting parameter configuration.

[0076] Figure 3 This is a flowchart of step S1 in the method of the present invention; Figure 4 This is a flowchart of step S2 in the method of the present invention; Figure 5 This is a flowchart of step S3 in the method of the present invention; Figure 6 This is a flowchart of step S4 in the method of the present invention; Figure 7 This is a flowchart of step S5 in the method of the present invention.

[0077] like Figure 3 As shown, step S1 specifically includes the following steps:

[0078] S101: Acquire the vector data of cutting force collected by the three-dimensional force sensor and the machining vibration signal of the workpiece surface node collected by the vibration sensor. The machining vibration signal includes stress data and displacement data. Based on the vector data, stress data and displacement data, the time stamp matching method is used to align them to obtain the surface node stress-displacement dataset.

[0079] Specifically, based on the cutting force vector data acquired by the three-dimensional force sensor, force signal sampling needs to be performed at fixed time intervals during data recording. Taking a frequency of 1000Hz as an example, 1000 sets of three-axis cutting force values ​​are recorded per second, representing the main cutting force ( ), feed direction cutting force ( ) and normal force ( Simultaneously, by arranging strain gauges or laser vibration sensor arrays on the surface of thin-walled parts, stress distribution and displacement response data of different nodes (e.g., numbered A1 to A20) on the workpiece surface are acquired. The synchronous acquisition frequency of node data is set to 1000Hz. Based on this, a timestamp matching method is used to compare the cutting force vector data with the node stress and displacement data at the corresponding time points one by one. The comparison is based on millisecond-level timestamps. If the time error is within ±0.5ms, it is determined to be data from the same sampling period; otherwise, the abnormal data is discarded. For example... At t=2.513s, the cutting force recorded by the force sensor is [120.3N, 65.8N, 30.4N], while the stress recorded by node A5 is σ=85.2MPa and the displacement is δ=0.023mm, with timestamps of 2.5132s and 2.5131s respectively. This meets the matching tolerance requirement, and the corresponding relationship is bound. After the nodes are matched, a surface node stress-displacement dataset is formed. The dataset is recorded in the following format: node number, timestamp, triaxial cutting force, node stress value, and node displacement value. Some sample data are listed in the table below:

[0080] Table 1: Surface Nodal Stress-Displacement Matching Data (Partial)

[0081]

[0082] As shown in Table 1, the stress-displacement response relationship is obtained by matching node data. The further refinement process is as follows: First, the three-dimensional force sensor outputs triaxial force data at a constant frequency and records it according to timestamps. Assuming a recording time period of 10 seconds, the total number of records is 10,000. Each data record is in the format shown in Table 1. The corresponding time is Secondly, in the nodal response data record, the displacement and stress values ​​of each node are also recorded simultaneously. The nodal response records are managed using numbering, and the data structure is as follows: Where j is the node number and k is the number of samples, the final matching process performs the following operations: For any timestamp Search for the data in node j that satisfies timestamp ,in If multiple values ​​meet the criteria, the one with the smallest error is selected for binding. Upon successful matching, a complete record is created in the dataset, and the data fields are merged. For data pairs with errors exceeding the threshold, they are discarded and recorded as missing. The missing rate is calculated as the proportion of the number of failed matching records to the total number of records. For example, if 63 out of 10,000 cutting force records fail to match, the missing rate is 0.63%, which is within a reasonable range. The node numbering should cover areas with significant changes in workpiece surface features, such as edge nodes and slotted nodes, and should be prioritized. During the cutting process, thin-walled aluminum alloy parts (such as 6061-T6) are used as workpieces with dimensions set to 100mm×80mm×3mm. Twenty stress response monitoring points are arranged along the long side. The total number of complete data pairs acquired synchronously during the acquisition period is 193,740, providing accurate input for subsequent vibration co-simulation.

[0083] S102: Perform multibody dynamics coupling calculation based on the surface node stress-displacement dataset to obtain cutting force vibration coupling response data. Specifically, based on the surface node stress-displacement dataset, input the cutting force vector and vibration signal into the multibody dynamics model respectively, analyze the vibration response of the nodes corresponding to the change in cutting force, extract key coupling nodes based on the peak force of the nodes, and obtain cutting force vibration coupling response data.

[0084] Based on the surface nodal stress-displacement dataset, the corresponding cutting force vector value in each matched set of data is... The nodal stress σ and displacement δ at the same time stamp are extracted as input parameters and applied to the established multibody dynamics model to simulate the small rigid body displacement and elastic deformation of the workpiece in three-dimensional space caused by force. The simulation iterates in units of 0.001 seconds. In each iteration, a cutting force vector corresponding to a time stamp is read and applied as an external force to the selected node surface of the workpiece model. The loading point is preferentially set based on the stress peak point. For example, if the stress value at node A6 exceeds 90 MPa and the displacement value reaches 0.03 mm at a certain time point, then this node will be selected in subsequent iterations. In the analysis, it is defined as the initial key coupling node and serves as the initial loading node. Subsequently, during the simulation analysis, the transmission path of its force response is traced, and the stress propagation effect on adjacent nodes caused by it is recorded to form a response chain. The model simulates 200 nodes, with 20 nodes forming a local element. Within each element, the nodes are connected to each other using spring-damped elements for force transmission modeling. For example, after node A6 is subjected to a loading force, when its force data propagates to adjacent nodes such as A5 and A7, the change in its vibration response Δδ is recorded and compared with the original displacement δ. If Δδ / δ≥0... If 15 is found, the adjacent node is also included in the current coupling region, gradually forming the coupling region boundary. For nodes within the coupling boundary, their vibration response curves are extracted, and the dominant frequency characteristics are obtained through FFT transformation. It is then determined whether the dominant frequency value is within the vibration-sensitive range of the cutting process. For example, in the machining of thin-walled aluminum alloy parts, the natural frequency of the structure is often located in the range of 180Hz to 220Hz. If the dominant frequency value of a node is within this range, it is marked as a strongly coupled node, ultimately forming a set of strongly coupled nodes. Each node includes its number, stress point number, maximum stress value, maximum displacement value, dominant frequency value, and vibration response. Six attribute fields, including maximum amplitude, are used for subsequent statistical analysis. Within a typical simulation cycle, 35 strongly coupled nodes were recorded, covering 5 local structural units. The nodes exhibited similar frequency amplification characteristics in their responses, and the vibration responses showed significant time delays. Synchronous analysis was then performed in the time dimension to obtain the time difference in node responses excited by changes in cutting force. By comparing the phase changes of responses from different nodes at different time periods, the propagation velocity of the response wavefront on the structure was identified. The propagation velocity of the structural response was obtained by dividing the maximum response time difference Δt between nodes by the distance Δx between nodes. With a distance of 6mm between A6 and A9, and a maximum time response difference of 0.0021s, then... Finally, the parameters are structured and organized to form cutting force vibration coupling response data. The data content is stored in time series for subsequent input.

[0085] S103: Based on the cutting force vibration coupling response data, obtain the vibration amplitude and frequency of the nodes during the simulation period. Based on the vibration amplitude and frequency, the stress data and the displacement data, that is, summarize and organize the characteristic data of multiple mechanical and vibration related parameters, obtain the cutting-vibration coupling simulation results.

[0086] Specifically, based on the cutting force vibration coupling response data, the vibration amplitude and frequency parameters of key nodes are gradually extracted during the time-series simulation of the multibody dynamics model throughout the entire simulation cycle. During execution, the displacement response δ(t) of each key coupling node in each simulation time step is sampled first, and the δ(t) curve is recorded at a sampling interval of 0.001 seconds. Then, the main vibration frequency is extracted from the δ(t) sequence using a fast Fourier transform. With response amplitude Nodes whose response amplitude exceeds twice the average amplitude within the main vibration frequency range [180, 220] Hz are selected and defined as excitation anomalous nodes. The distribution density of anomalous nodes in the time domain is statistically analyzed, and the response trend of nodes in the direction of cutting force application (e.g., the main cutting force direction) is further tracked. If the response amplitude of a node in the main cutting direction accounts for more than 70%, it is defined as a directional main response node. For each main response node, the three components corresponding to its stress point are extracted. Correlation analysis was performed on its vibration frequency and amplitude separately, and the correlation criterion was set as the Pearson correlation coefficient. If the absolute correlation coefficient between any directional component and the vibration frequency is... If the force component in that direction is determined to be the main excitation component, then this main excitation component is compared with the maximum stress value of the node. If the node stress exceeds 1.2 times the reference stress σ0 and its change trend is synchronized with that of the main excitation component (i.e., the stress increases synchronously when the applied force increases and decreases synchronously when the force decreases), then this node is recorded as the main node for cutting vibration. The reference stress σ0 is set as the median value within the range of the average stress of the entire node group ± one standard deviation. For example, if the average stress value is 82 MPa and the standard deviation is 6 MPa, then σ0 = 82 MPa, and the judgment threshold is 98.4 MPa. Subsequently, the main node is established... The coupling relationship curve between the cutting force change and the vibration response is obtained by fitting the functional trend of vibration response and force change at intermediate sampling points using interpolation. The slope of the curve is statistically analyzed in different time periods. If the slope of the curve changes abruptly (i.e., the change Δk exceeds three times the average value of the previous time period), the node is marked as being in an unstable response stage. Finally, the stress, displacement, amplitude, frequency, main excitation direction and coupling relationship slope of each node are extracted to form a structural feature vector, which is used for subsequent vibration prediction model training. The feature data of the nodes are summarized in a unified format to form a complete cutting-vibration coupling simulation result dataset.

[0087] like Figure 4As shown, step S2 specifically includes the following steps:

[0088] S201: Based on the workpiece mesh cell information in the cutting-vibration coupling simulation results, the change in material removal volume in each mesh cell is obtained, and the rate of change of the cutting force vector in adjacent time steps is calculated by differential operation to obtain the rate of change of the cutting force vector.

[0089] Among them, the rate of change of the cutting force vector refers to the rate of change of the node cutting force vector per unit time, reflecting the dynamic evolution trend of the force response during the material removal process;

[0090] Specifically, based on the workpiece mesh element information from the cutting-vibration coupling simulation results, the initial material volume of each mesh element throughout the entire simulation cycle is first extracted. Then extract the remaining material volume of the same grid cell at adjacent time steps. By the difference between the two:

[0091] ;

[0092] The change in the volume of material removed from the unit during this time period is obtained, assuming an initial volume of 15.2 mm. 3 The current time step volume is 14.8 mm. 3 ,but The time step is 0.001 seconds. Differential operations are performed on consecutive time steps to record the material removal rate at each moment. Then, the cutting force vector value of the cutting node corresponding to the mesh cell is extracted. Then, compare it with the cutting force vector value from the previous time step. Subtraction yields the change. and divide by the time interval To obtain the rate of change of the cutting force vector, let:

[0093] , ,but:

[0094] ;

[0095] Based on the above results, the change in volume removed was aligned and matched with the rate of change in cutting force. The force nodes corresponding to the material removal source were identified through node position indexing. Taking the thin-walled region of aluminum alloy in the real-time machining path as an example, the force value in the Z-axis direction increased by more than 2000 N / s, and the material removal volume reached 0.4 mm. 3 When a node has a non-zero value in three consecutive time steps, it is recorded as a complete removal event and included in the statistics. and If a node is identified as a high-frequency cutting point, its proportion among nodes is statistically analyzed. Assuming a total of 200 nodes, 18 nodes meet the criteria, representing 9% of the total. These 9% serve as the basis for extracting load reference points for subsequent stiffness recursion. Table 2 lists the material volume changes and cutting force rates of change for some nodes.

[0096] Table 2: Rate of Change in Material Volume and Cutting Force at Cutting Nodes

[0097]

[0098] As shown in Table 2, nodes C07 and E18 are typical high-frequency cutting points, with cutting force vector change rates exceeding 2000 N / s in all three directions and material removal volumes exceeding 0.5 mm. 3 This can be used as the basis for the selection of loading points in subsequent stiffness updates, and finally the set of cutting force vector change rate arrays is obtained. The arrays are structured and organized into three-dimensional change rate values ​​corresponding to each node and each time step, which are used as inputs for the next stiffness iteration.

[0099] S202: Based on the rate of change of the cutting force vector and the change in the material removal volume of each node, the finite element stiffness is recursively calculated to obtain the updated node stiffness coefficients.

[0100] Among them, the finite element stiffness recursive calculation is based on the material removal and cutting force changes of each node during the cutting-vibration process, and the node stiffness coefficient is updated in real time to dynamically reflect the stiffness evolution of thin-walled workpieces.

[0101] Specifically, based on the cutting force vector change rate, first read the material removal volume data and cutting force change rate data set of the nodes marked in the previous segment, and then calculate the material volume change for each node. Its corresponding rate of change of cutting force Pairing is performed to obtain reference values ​​for nodal stress changes, and then the initial elastic modulus preset in the material parameter library for each node is consulted. Poisson's ratio The influence factor of volume change is introduced, and the node material is assumed to be aluminum alloy 7075. , For each node, the stiffness reduction rate is calculated based on the percentage change in its volume. For example, if Let the reduction factor proportionality constant be 3.5, then we have Then, this reduction rate is applied to the coefficient terms of the original stiffness matrix to obtain the updated stiffness coefficients. ,in Let the cross-sectional area of ​​the node be... Node spacing ,but:

[0102] ;

[0103] Substituting the reduction rate, we get:

[0104] ;

[0105] For each of these nodes, a new node stiffness library is constructed after the stiffness update. This library is then compiled into updated result data, containing each node's number, initial stiffness, volume change, reduction rate, and updated stiffness. This data is recorded in the node stiffness update mapping table. If a certain node... If the volume loss exceeds 20% of the original volume, then the node is marked as having severely degraded stiffness. Let the initial volume of node F05 be 12.5 mm. 3 The current volume is 9.8mm. 3 The loss would then be 2.7mm. 3 The proportion was 21.6%, which was classified as a severely degraded area, and these nodes were modeled separately in the subsequent multi-scale analysis.

[0106] S203: Obtain multi-scale stiffness degradation evolution data of thin-walled parts based on nodal stiffness coefficients, perform multi-scale coupled calculations based on multi-scale stiffness degradation evolution data, analyze the stiffness degradation parameters of each level, and obtain multi-scale nodal stiffness degradation data.

[0107] Specifically, to retrieve the node stiffness coefficient update results, the nodes are first divided into three levels according to their geometric regions: micro-scale (e.g., the tool contact area within 2mm), meso-scale (e.g., the transition area of ​​thin-walled structures within 2-10mm), and macro-scale (the entire part within 10mm or more). The node set within each scale is then extracted one by one, and its stiffness update data is read. Then, based on the average stiffness of nodes within the region, a representative value for each scale is calculated. Let the average stiffness of the micro-scale region be 26.5 kN / mm, the meso-scale be 28.7 kN / mm, and the macro-scale be 29.1 kN / mm. The stiffness gradient between adjacent scales is calculated as follows:

[0108] ;

[0109] The gradient value is used to determine the evolution trend of stiffness degradation across multiple scales, and then the dominant frequency of nodal vibration within each scale is determined. , maximum amplitude Statistical analysis shows that if more than 30% of the node frequencies are concentrated in the sensitive region [180, 220] Hz at a certain scale, and their stiffness reduction rate is high... If the scale is defined as a resonance-sensitive region, such as a mesoscale region with 40 nodes, 15 of which meet the above conditions, accounting for 37.5%, this scale is marked as a flutter high-risk region. Finally, the system outputs characteristic parameters such as the average stiffness degradation at multiple scales, the number of the node with the largest degradation, the proportion of the dominant frequency concentration, the total number of nodes, the reduction rate threshold, and the frequency concentration interval, forming a standardized multi-scale stiffness degradation data structure, which provides basic data input for the next step of constructing multi-layer feature vectors.

[0110] like Figure 5 As shown, step S3 specifically includes the following steps:

[0111] S301: Obtain the stiffness degradation value of the node based on multi-scale node stiffness degradation data, obtain the node response coordinate of the workpiece surface position based on the stiffness degradation value and the workpiece displacement record at the current processing time, and obtain the geometric position data of the workpiece surface based on the node response coordinate. Specifically, integrate the coordinate data and node corresponding information at multiple times according to the time step to obtain the geometric position data of the workpiece surface.

[0112] Specifically, to retrieve the stiffness degradation values ​​of nodes from multi-scale node stiffness degradation data, it is first necessary to establish the correspondence between nodes and their initial stiffness reference values. This involves numbering the nodes in the initial machining state using a 3D workpiece model, comparing the changes in node stiffness over time steps during machining with their initial values, and calculating the normalized stiffness degradation values. Taking the machining of a thin-walled structural component as an example, the stiffness values ​​for its 300 nodes in the initial state are set as follows: Taking node 1 as an example, its initial stiffness is 1350 N / mm, and the stiffness is measured to be 1080 N / mm after 20 seconds of machining. Therefore, the normalized value of the stiffness degradation of this node is... This operation needs to be performed on all nodes. Combining the workpiece displacement record at the current processing moment, the real-time spatial position of the nodes can be obtained using displacement sensors or a 3D scanner. The displacement data of the nodes from t=0s to t=30s is recorded with a time step resolution of 1s. The acquisition method involves calculating the position offset between the theoretical spatial coordinates of the node and the currently detected real-time coordinates, thereby obtaining the response coordinate information at the current processing moment. For example, if node 1's position at t=0s is (10.0, 15.2, 3.5)mm, and its position at t=20s is (10.5, 15.8, 3.3)mm, then its displacement is (0.5, 0.6, -0.2)mm. Subsequently, the response coordinates of the nodes within each time step are integrated over time to construct a set of 3D matrix data. Each layer corresponds to the spatial response position of the node at one time step. The node number is matched with the coordinate value to reconstruct the correspondence between coordinates and nodes. This method generates the geometric position data of the workpiece surface, ultimately forming a 3D array of node number, time step, and 3D coordinates. For example, the 30s response path of node 1 is represented as follows: The total number of nodes is N=300, and each node records coordinate information once per second, for a total of 9000 sets of data.

[0113] S302: Based on geometric position data and multi-scale node stiffness degradation data, obtain the real-time displacement vector and stiffness change of the node at the processing time, and perform proportional weighted summation calculation based on the real-time displacement vector and stiffness degradation data to obtain the structural modal frequency weight coefficient.

[0114] Specifically, based on the synchronous matching results of workpiece surface geometric position data and multi-scale nodal stiffness degradation data, the first step is to match the response coordinate values ​​of the nodes with their corresponding stiffness degradation values ​​at the machining time. That is, at second t, the nodes are matched... Call its coordinate values Normalized value of stiffness degradation Create an associative array Then, the real-time displacement direction of the node is extracted. By calculating the difference vector between its coordinates and the initial position coordinates, the spatial motion direction of the node is determined. For example, if the initial position of node 5 is (20.0, 30.0, 10.0) mm, and its position at machining t=10s is (21.2, 31.5, 9.7) mm, then its displacement vector is (1.2, 1.5, -0.3), and the direction is the normalized result of this vector. Subsequently, the spatial coordinate values ​​of the node and the stiffness degradation values ​​are proportionally weighted and summed to calculate the structural modal frequency weighting coefficient, using the formula:

[0115] ;

[0116] in, Represents the structural modal frequency weighting coefficients;

[0117] This represents the total number of nodes used for frequency weight calculation at the current processing time.

[0118] Representing the The real-time node response coordinates of each node at the current processing moment;

[0119] Representing the Normalized values ​​of node stiffness degradation;

[0120] This represents the average value of the response coordinates of all nodes at the current processing moment;

[0121] It represents a dimensionless, extremely small positive number with a denominator of zero.

[0122] in, This represents the modulus of a node, i.e.:

[0123] ;

[0124] The average coordinates of the nodes at the current time. Set as To avoid division by zero errors, the number of nodes The following are some node parameter values:

[0125] Table 3: Machining Displacement and Stiffness Degradation of Structural Nodes

[0126]

[0127] As shown in Table 3, during the calculation process using the formula, taking node 1 as an example, its modulus... mm, assuming an average difference of 2.68 mm and a stiffness degradation value of 0.2, substituting these values ​​into the calculation yields:

[0128] Node 1 ;

[0129] Similarly, the remaining nodes can be calculated, summed, and then divided by N to obtain the final structural modal frequency weighting coefficients. Assuming the sum of all nodes is 125.85, substituting it in, we get:

[0130] ;

[0131] The result shows that the structural modal frequency weighting coefficient is 0.4195. The advantage of this formula lies in its ability to comprehensively reflect the dominance of nodes in the overall modal response by introducing the ratio of the product of the nodal response offset and stiffness degradation parameters to the deviation from the mean.

[0132] S303: Based on the structural modal frequency weighting coefficient, obtain the dynamic response path of the workpiece at the node during the processing; based on the dynamic response path, obtain the distribution function curve of the response amplitude and frequency; based on the distribution function curve, obtain the flutter sensitive frequency range; and based on the structural modal frequency weighting coefficient and the flutter sensitive frequency range, perform fusion calculation to obtain the flutter critical frequency simulation data.

[0133] Specifically, based on the structural modal frequency weighting coefficients, the dynamic response paths of workpiece nodes during the machining process are extracted. First, the node response paths are extracted based on the time series formed by the changes in the three-dimensional coordinates of the nodes over time, that is, for each node... Construct its displacement time series from 0s to 30s. The coordinate differences between time steps are processed using first-order difference to obtain the dynamic amplitude change sequence of the node response. The rate of change in the differential frequency bands is calculated. The node response signal is then subjected to Fast Fourier Transform (FFT) to obtain the frequency domain response spectrum. The frequency-amplitude distribution curve is extracted. Furthermore, based on the amplitude growth rate within different frequency bands (e.g., 5Hz~15Hz, 15Hz~30Hz, 30Hz~60Hz), the frequency band with the largest amplitude change rate is determined, thus defining the critical frequency range for flutter occurrence. For example, in the frequency distribution of node 1 response, the amplitude is 0.8mm at 15Hz and 2.6mm at 30Hz, then the growth rate is... If the frequency band is set to a threshold of 0.1 mm / Hz, then the frequency band is considered to have a potential flutter risk. Finally, the frequency band growth rate calculated for each node is fused with the overall structural modal frequency weighting coefficient to calculate the overall flutter critical frequency in a weighted manner. Assuming the weighting coefficient is 0.4195, and the corresponding frequency band median is 22.5 Hz, the simulated critical frequency data are as follows:

[0134] ;

[0135] This indicates that the critical frequency of the workpiece under this processing state is 31.99 Hz. This result shows that this frequency enters the critical frequency range of unstable modal response.

[0136] like Figure 6 As shown, step S4 specifically includes the following steps:

[0137] S401: Obtain the critical frequency value based on the chatter critical frequency simulation data, obtain the spindle frequency and tool feed trajectory coordinates, where the spindle frequency includes the spindle speed and feed rate. Based on the critical frequency value, spindle frequency and feed trajectory coordinates, the arrangement of discrete points of the monitoring path in the time dimension and the correlation of frequency and working condition parameters, obtain the path segment frequency matching data.

[0138] Specifically, when obtaining the critical frequency value from the chatter critical frequency simulation data, the data generated by the tool-process system dynamics simulation module is first imported into the simulation database. This data includes multiple sets of critical frequency information. Each set of frequency values ​​is determined by modeling under a combination of differentiated structural stiffness and tool overhang length. For example, when the tool diameter is 10mm, the overhang length is 60mm, and the workpiece stiffness is 3.2×10⁻⁶... 6At N / m, the corresponding critical frequency is 920Hz. When acquiring the coordinates of the tool feed path points, the coordinates of the midpoints of the G1 straight line segment and the G2 circular arc segment are parsed from the machining program G code. A set of coordinates is sampled every 1ms, and the current spindle speed and feed rate are recorded simultaneously. For example, if the spindle speed is set to 8000rpm and the feed rate is set to 600mm / min, the path points include coordinates (20.0, 10.0), (20.5, 10.0), (21.0, 10.0), etc. When matching the path coordinates with the frequency data, the timestamp of each path point is used. Calculate the current spindle angular frequency , such as when At rpm, The frequency value (Hz) is mapped to the critical frequency value in the simulation database along a time axis. During the monitoring of the discrete points along the path in the time dimension, the corresponding frequency values ​​are recorded according to the order of the path points and their time distribution. A frequency information trajectory is generated every 10ms, forming a time series array. When associating the frequency with operating parameters, the corresponding point's feed rate, spindle frequency, tool number, and other machining parameters are bound to the path point structure, constructing the following structure array:

[0139] ;

[0140] For example, the first point is: When generating path segment frequency matching data, 10 consecutive points are taken as a path segment. The average frequency and maximum frequency offset of the segment are calculated, and the segment start and end index, average frequency, and corresponding tool number are recorded. The final output matching data is shown in Table 4.

[0141] Table 4: Path Segment Frequency Matching Data Table

[0142]

[0143] As shown in Table 4, the start and end indices of path segment number 1 are 0 to 9, corresponding to a spindle frequency of 133.33Hz, a critical matching frequency of 920Hz, tool number T1, and a feed rate of 600mm / min. This data table is used to establish the correspondence between frequency and path.

[0144] S402: Based on the path segment frequency matching data, the frequency matching algorithm is used to obtain the difference between the main axis frequency and the critical frequency value. Based on the comparison between the difference and the set risk threshold, the path segments with frequency deviations exceeding the threshold are screened, the path index and risk coefficient are extracted, and the resonance risk path segment data are obtained.

[0145] Specifically, the risk threshold refers to the minimum safe interval between the spindle frequency and the chatter critical frequency, which is determined based on the dynamic characteristics of the tool-process system, and is taken as 2% to 5% of the critical frequency as a reference.

[0146] Based on path segment frequency matching data, when analyzing the relative deviation between the principal axis frequency and the critical frequency of a path segment, the principal axis frequency corresponding to each group of path segment numbers is... Critical frequency corresponding to this segment The difference operation is performed using the following formula: For example, in path segment 1 Hz, Hz, get Hz; When setting the risk threshold, 2% to 5% of the critical frequency is used as the frequency safety interval, and the median value of 3.5% is selected as the benchmark, that is, the threshold is set as: For path segment 1, the threshold is Hz; the screening process is as follows: if If so, it is determined that the current path segment has a resonance risk, for example, in path segment 2. Hz, Hz, get Hz, and Hz, meeting the condition; when extracting the path index, output the start and end point numbers of the segment, such as segment 2 as (10, 19); the risk coefficient is set as the ratio of the reference difference to the critical frequency, defined as:

[0147] ;

[0148] As in paragraph 2:

[0149] ;

[0150] The output result of obtaining the resonance risk path segment data is: (path segment number 2, index 10~19, risk coefficient 0.55), only retaining the data. The segment is designated as a resonance risk segment and will be further processed.

[0151] S403: Based on the resonance risk path segment data, the spacing between path points is adjusted while maintaining the sequence duration, the trajectory coordinate distribution is optimized, and the frequency change amplitude of the path segment is recalculated to obtain a flutter suppression trajectory simulation scheme.

[0152] Specifically, based on the resonance risk path segment data, the identified resonance risk path segments are processed segment by segment. When adjusting the spacing between path points, the overall length of the path segment remains unchanged. Point density is changed using point interpolation or elimination methods to uniformly redistribute the time intervals between points, keeping the total sequence duration constant. For example, path segment 2 originally had a point spacing of 0.5mm, a length of 5mm, and 10 points. After adjustment, the spacing is set to 0.25mm, with 20 points, and the time interval between each point is adjusted from 10ms to 5ms. When optimizing the trajectory coordinate distribution, the newly distributed coordinate points are called to regenerate the path segment coordinate sequence, maintaining the spatial continuity of each point. A smooth trajectory is constructed using a quadratic interpolation method. After reconstruction, the path segment becomes (20.0, 10.0), (20.25, 10.0), (20.5, 10.0)...(22.5, 10.0). When recalculating the frequency change amplitude of the path segment, the corresponding angular frequency of the main shaft at each point is calculated based on the new point interval and the main shaft rotation speed. And obtain the frequency distribution sequence of multiple points within this segment, and count the maximum frequency fluctuation amplitude. For example, if the frequency distribution is [133.30, 133.32, ..., 133.40], then... Hz, finally generating a flutter suppression trajectory simulation scheme, which includes path segment number, point coordinate sequence, frequency change amplitude and comparison data before and after optimization, and outputs a new processing path simulation scheme.

[0153] like Figure 7 As shown, step S5 specifically includes the following steps:

[0154] S501: Based on the chatter suppression trajectory simulation scheme, the path point position is obtained, the feed direction and cutting parameters are obtained, the path spacing and direction offset are obtained based on the path point position and feed direction, the coordinate gradient change is obtained based on the path spacing and direction offset, and the initial cutting parameter coefficient set of the path point is obtained based on the coordinate gradient change and cutting parameters. Among them, the cutting parameters include spindle speed, feed rate and depth of cut.

[0155] Specifically, firstly, regarding the coordinate information of the path points, it is necessary to extract the specific position coordinates of each path point from the path simulation data, and analyze them in conjunction with the feed direction to calculate the path spacing and directional offset. The path point positions and feed directions can be obtained from the path planning parameters of the CNC system. For example, the path point positions... arrive The position data is used, while the feed direction is matched using the feed speed direction parameters of the machining equipment; based on this data, the path spacing can be calculated. and direction offset For example, if path point 1 is... Path point 2 is Then the path spacing is:

[0156] ;

[0157] Directional offset The angle change between paths is calculated based on the dot product formula of vectors; the gradient change of the path needs to be calculated, which will help identify factors such as sharp turns in the path, which is particularly important for vibration suppression during the machining process; finally, by comprehensively considering the gradient change of the path points, the rotational speed, the feed rate, and the depth of cut, a preliminary set of cutting parameter coefficients is obtained; the depth of cut and feed rate are dynamically adjusted according to the material properties of the workpiece and the real-time machining requirements. For example, in the case of a certain metal with a cutting depth of 2mm and a feed rate set to 120mm / min, the initial cutting parameter coefficients can be set through experimental data or literature.

[0158] S502: Obtain workpiece roughness and machining efficiency; obtain roughness fluctuation rate and efficiency response time based on workpiece roughness and machining efficiency; obtain a multi-objective optimization function based on roughness fluctuation rate and efficiency response time; obtain the multi-objective function input parameter set based on the multi-objective optimization function and the initial cutting parameter coefficient set of the path point; wherein, the multi-objective optimization function is a roughness and efficiency dual-objective optimization function, which jointly models the path point offset and cutting parameters based on the two objectives of workpiece surface roughness and machining time;

[0159] Specifically, the initial cutting parameter coefficient set of the path points is called to further detect the workpiece roughness and machining time during the machining process. During the detection process, the roughness changes need to be obtained through real-time measurement, detected using a roughness instrument, and converted into roughness fluctuation rate. Assume the roughness data during the machining process is as follows: , (representing the roughness of the workpiece before and after machining, respectively), then the roughness fluctuation rate can be calculated as: Next, the efficiency response time is calculated, which is the processing parameter that can be responded to and adjusted within a specific time period. Assuming the roughness change response time is 15 seconds during processing, the efficiency response time is the value within 15 seconds. Combining the roughness fluctuation rate with the efficiency response time, a bi-objective optimization function is constructed. In this process, the optimization function design considers two objectives: minimizing workpiece roughness fluctuation and maximizing processing efficiency. At this stage, not only the surface quality of the workpiece is considered, but also the processing efficiency is balanced to ensure an ideal balance between production time and processing accuracy.

[0160] For example, consider the path point offset as The cutting parameters are a feed rate of 120 mm / min, which will be used for dynamic correction after path point optimization. When obtaining the input parameter set of the dual objective function, multiple factors such as the material properties of the workpiece, the wear of the cutting tool, the feed rate, and the spindle speed need to be considered comprehensively so that the selection of the optimization objective can cover the entire machining process.

[0161] S503: The NSGA-II algorithm is based on the input parameter set of a multi-objective function. Specifically, the improved NSGA-II algorithm is used for initialization and crossover mutation to obtain multiple solutions. The congestion and dominance relationships of the solutions are calculated based on the multiple solutions. The non-dominant solution path point offset and cutting parameter combination are screened to obtain the co-simulation optimization parameter set.

[0162] Among them, the improved NSGA-II algorithm adjusts crossover and mutation probabilities through self-matching, optimizes population initialization strategy, and introduces a local search mechanism to optimize multi-objective parameters.

[0163] Specifically, after calling the dual-objective function and inputting the parameter set, the improved NSGA-II algorithm is used to optimize the path points. Specifically, the NSGA-II algorithm first performs an initialization operation based on the path point offsets and cutting parameter set in the initial population, generating multiple solutions. At this point, the crowding degree of the solution refers to the density of the solution in the target space, representing the distribution of the solution in the target space. By calculating the crowding degree of the solution, it can be determined whether a path point is in a region of excellent solutions. New path point parameters are generated through crossover and mutation operations, and multiple optimization iterations are performed. Non-dominated solutions are then selected based on the crowding degree and dominance relationships. The core objective of this process is to improve machining efficiency without compromising machining accuracy. Specifically, it involves calculating the dominance relationship of the solution in the objective function space, selecting the optimal solution path point, and further verifying the performance of the path point in real-time machining through simulation. Finally, the co-simulation optimization parameter set can provide specific combinations of machining parameters, such as path point offset, depth of cut, spindle speed, and feed rate, ensuring high efficiency and stability in the machining process. In some cases, the improved NSGA-II algorithm also corrects path points exceeding thresholds, ensuring that parameter settings do not exceed safe limits during optimization.

[0164] Table 5: Examples of Path Point Optimization Parameters

[0165]

[0166] As shown in Table 5, the optimized set of parameters such as path point offset, feed rate, spindle speed, and depth of cut helps to improve the efficiency, response time, and surface roughness control of the machining process.

[0167] In summary, the method of this invention, based on the joint acquisition of real-time cutting force vector and vibration data, combined with dynamic nodal stiffness recursion and synchronous analysis of workpiece geometric information, and fused with modal frequency self-matching identification, realizes the dynamic correlation between frequency and machining motion parameters, dynamically tracks structural state changes at each machining moment, avoids resonance risks through frequency matching and real-time trajectory adjustment, strengthens the fusion of information between multiple physical fields during the cutting process, improves the analytical depth of chatter formation mechanism and the timeliness of control measures, makes the overall simulation prediction more compatible with real-time machining conditions, significantly improves the accuracy of chatter risk identification and control feedback speed, and optimizes the ability to ensure the stability of complex thin-walled parts during process development.

[0168] The present invention also provides a system. The embodiments of the present invention also propose a multi-scale co-simulation system for thin-walled part cutting oriented towards chatter suppression. Its principle is similar to the multi-scale co-simulation method for thin-walled part cutting oriented towards chatter suppression, and will not be described in detail here.

[0169] Figure 8 This is a schematic diagram of a multi-scale co-simulation system for thin-walled part cutting aimed at chatter suppression. The system 100 includes:

[0170] The first module 110 is used to acquire vector data of cutting force and machining vibration signal, and to perform coupled calculation of multibody dynamics based on vector data and machining vibration signal to obtain cutting-vibration coupled simulation results;

[0171] The second module 120 is used to obtain the change in material removal volume and the rate of change of cutting force vector at the node based on the cutting-vibration coupling simulation results, and to perform finite element stiffness coefficient update calculation based on the change and the rate of change to obtain multi-scale node stiffness degradation data.

[0172] The third module 130 obtains the geometric position data of the workpiece at the processing time based on the multi-scale node stiffness degradation data, performs a proportional weighted summation calculation based on the multi-scale node stiffness degradation data and the geometric position data to obtain the structural modal frequency weighting coefficient of the workpiece, obtains the flutter sensitive frequency range based on the structural modal frequency weighting coefficient, and performs a fusion calculation based on the structural modal frequency weighting coefficient and the flutter sensitive frequency range to obtain flutter critical frequency simulation data.

[0173] The fourth module 140 is used to obtain the critical frequency value based on the flutter critical frequency simulation data, obtain the spindle frequency, and compare the critical frequency value and the spindle frequency using a frequency matching algorithm to obtain a flutter suppression trajectory simulation scheme.

[0174] Figure 9This is another schematic diagram of a multi-scale co-simulation system for thin-walled part cutting aimed at chatter suppression. This system is largely the same as the one described above, except that it also includes:

[0175] The fifth module 150 is used to obtain the cutting parameters of the path points based on the chatter suppression trajectory simulation scheme, and to obtain the workpiece roughness and machining efficiency. Based on the cutting parameters, workpiece roughness and machining efficiency, a multi-objective optimization function is obtained. Based on the multi-objective optimization function, the NSGA-II algorithm is performed to obtain a set of co-simulation optimization parameters.

[0176] This invention also provides a computer device. Figure 10 This is a schematic diagram of a computer device in an embodiment of the present invention. The computer device 500 includes a memory 510, a processor 520, and a computer program 530 stored in the memory 510 and executable on the processor 520. When the processor 520 executes the computer program 530, it implements the above-mentioned multi-scale co-simulation method for thin-walled part cutting oriented towards chatter suppression.

[0177] This invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the aforementioned multi-scale co-simulation method for thin-walled part cutting aimed at chatter suppression.

[0178] This invention also provides a computer program product, which includes a computer program that, when executed by a processor, implements the above-described multi-scale co-simulation method for thin-walled part cutting aimed at chatter suppression.

[0179] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0180] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0181] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0182] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0183] The above description is merely a specific embodiment of the present invention and should not be construed as limiting the scope of the invention. Therefore, substitutions of equivalent components, or equivalent changes and modifications made within the scope of protection of this application, should still fall within the scope of this application. Furthermore, the technical features, technical features and technical solutions, and technical solutions in this invention can be freely combined and used.

Claims

1. A multi-scale co-simulation method for machining thin-walled parts to suppress chatter, characterized in that, The method includes the following steps: S1: Obtain vector data of cutting force and machining vibration signal, and perform multibody dynamics coupling calculation based on the vector data and machining vibration signal to obtain cutting-vibration coupling simulation results; S2: Based on the cutting-vibration coupling simulation results, obtain the change in material removal volume and the rate of change of cutting force vector at the node, and perform finite element stiffness coefficient update calculation based on the change and the rate of change to obtain multi-scale node stiffness degradation data. S3: Obtain the geometric position data of the workpiece at the processing time based on the multi-scale node stiffness degradation data, perform a proportional weighted summation calculation based on the multi-scale node stiffness degradation data and the geometric position data to obtain the structural modal frequency weighting coefficient of the workpiece, obtain the flutter sensitive frequency range based on the structural modal frequency weighting coefficient, and perform a fusion calculation based on the structural modal frequency weighting coefficient and the flutter sensitive frequency range to obtain the flutter critical frequency simulation data; S4: Based on the chatter critical frequency simulation data, obtain the critical frequency value and the spindle frequency. Compare the critical frequency value and the spindle frequency using a frequency matching algorithm to obtain a chatter suppression trajectory simulation scheme. Step S4 specifically includes the following steps: S401: Based on the chatter critical frequency simulation data, obtain the critical frequency value, the spindle frequency, and the tool feed trajectory coordinates. Obtain path segment frequency matching data based on the critical frequency value, the spindle frequency, and the feed trajectory coordinates. S402: Based on the path segment frequency matching data, use a frequency matching algorithm to obtain the difference between the spindle frequency and the critical frequency value. Compare the difference with a set risk threshold to obtain resonance risk path segment data. S403: Based on the resonance risk path segment data, obtain the chatter suppression trajectory simulation scheme.

2. The multi-scale co-simulation method for thin-walled part cutting oriented towards chatter suppression according to claim 1, characterized in that, Step S1 specifically includes the following steps: S101: Acquire vector data of cutting force collected by a three-dimensional force sensor and machining vibration signals of workpiece surface nodes collected by a vibration sensor, wherein the machining vibration signals include stress data and displacement data. Based on the vector data, the stress data and the displacement data, an alignment method is used to obtain a surface node stress-displacement dataset. S102: Perform multibody dynamics coupling calculations based on the surface node stress-displacement dataset to obtain cutting force vibration coupling response data; S103: Based on the cutting force vibration coupling response data, obtain the vibration amplitude and frequency of the nodes during the simulation period. Based on the vibration amplitude and frequency, the stress data and the displacement data, obtain the cutting-vibration coupling simulation results.

3. The multi-scale co-simulation method for chatter suppression in thin-walled part cutting according to claim 1, characterized in that, Step S2 specifically includes the following steps: S201: Based on the workpiece mesh cell information in the cutting-vibration coupling simulation results, obtain the change in the material removal volume in each mesh cell, and calculate the rate of change of the cutting force vector within adjacent time steps through differential operation to obtain the rate of change of the cutting force vector. S202: Based on the rate of change of the cutting force vector and the change in the material removal volume of each node, perform finite element stiffness recursive calculation to obtain the updated node stiffness coefficients; S203: Obtain multi-scale stiffness degradation evolution data of thin-walled parts based on the nodal stiffness coefficients, and perform multi-scale coupling calculations based on the multi-scale stiffness degradation evolution data to obtain the multi-scale nodal stiffness degradation data.

4. The multi-scale co-simulation method for chatter suppression in thin-walled part cutting according to claim 1, characterized in that, Step S3 specifically includes the following steps: S301: Obtain the stiffness degradation value of the node based on the multi-scale node stiffness degradation data, obtain the node response coordinates of the workpiece surface position based on the stiffness degradation value and the workpiece displacement record at the current processing time, and obtain the geometric position data of the workpiece surface based on the node response coordinates. S302: Based on the geometric position data and the multi-scale node stiffness degradation data, obtain the real-time displacement vector and stiffness change of the node at the processing time, and perform proportional weighted summation calculation based on the real-time displacement vector and the stiffness degradation data to obtain the structural modal frequency weighting coefficient. S303: Based on the structural modal frequency weighting coefficient, obtain the dynamic response path of the workpiece at the node during the processing; based on the dynamic response path, obtain the distribution function curve of the response amplitude and frequency; based on the distribution function curve, obtain the flutter sensitive frequency range; and based on the structural modal frequency weighting coefficient and the flutter sensitive frequency range, perform fusion calculation to obtain the flutter critical frequency simulation data.

5. The multi-scale co-simulation method for chatter suppression in thin-walled part cutting according to claim 1, characterized in that, The method further includes: S5: Based on the chatter suppression trajectory simulation scheme, obtain the cutting parameters of the path points, and obtain the workpiece roughness and machining efficiency. Based on the cutting parameters, the workpiece roughness and the machining efficiency, obtain a multi-objective optimization function. Based on the multi-objective optimization function, perform the NSGA-II algorithm to obtain a set of co-simulation optimization parameters.

6. The multi-scale co-simulation method for chatter suppression in thin-walled part cutting according to claim 5, characterized in that, Step S5 specifically includes the following steps: S501: Based on the chatter suppression trajectory simulation scheme, obtain the path point position, obtain the feed direction and cutting parameters, obtain the path spacing and direction offset based on the path point position and the feed direction, obtain the coordinate gradient change based on the path spacing and the direction offset, and obtain the initial cutting parameter coefficient set of the path point based on the coordinate gradient change and the cutting parameters, wherein the cutting parameters include spindle speed, feed rate and depth of cut; S502: Obtain workpiece roughness and machining efficiency; obtain roughness fluctuation rate and efficiency response time based on the workpiece roughness and machining efficiency; obtain a multi-objective optimization function based on the roughness fluctuation rate and efficiency response time; obtain a multi-objective function input parameter set based on the multi-objective optimization function and the initial cutting parameter coefficient set of the path point. S503: Based on the input parameter set of the multi-objective function, perform the NSGA-II algorithm to obtain multiple solutions. Based on the multiple solutions, calculate the congestion degree and dominance relationship of the solutions to obtain a set of co-simulation optimization parameters.

7. A multi-scale co-simulation system for cutting thin-walled parts to suppress chatter, characterized in that, The collaborative simulation system includes: The first module is used to acquire vector data of cutting force and machining vibration signal, and to perform multibody dynamics coupling calculation based on the vector data and machining vibration signal to obtain cutting-vibration coupling simulation results; The second module is used to obtain the change in material removal volume and the rate of change of cutting force vector at the node based on the cutting-vibration coupling simulation results, and to perform finite element stiffness coefficient update calculation based on the change and the rate of change to obtain multi-scale node stiffness degradation data. The third module obtains the geometric position data of the workpiece at the processing time based on the multi-scale node stiffness degradation data, performs a proportional weighted summation calculation based on the multi-scale node stiffness degradation data and the geometric position data to obtain the structural modal frequency weighting coefficient of the workpiece, obtains the flutter sensitive frequency range based on the structural modal frequency weighting coefficient, and performs a fusion calculation based on the structural modal frequency weighting coefficient and the flutter sensitive frequency range to obtain the flutter critical frequency simulation data. The fourth module is used to obtain a critical frequency value based on the chatter critical frequency simulation data, obtain the spindle frequency, compare the critical frequency value and the spindle frequency using a frequency matching algorithm to obtain a chatter suppression trajectory simulation scheme, obtain a critical frequency value based on the chatter critical frequency simulation data, obtain the spindle frequency and tool feed trajectory coordinates, obtain path segment frequency matching data based on the critical frequency value, the spindle frequency and the feed trajectory coordinates; obtain the difference between the spindle frequency and the critical frequency value using a frequency matching algorithm based on the path segment frequency matching data, obtain resonance risk path segment data based on the comparison of the difference with a set risk threshold; and obtain the chatter suppression trajectory simulation scheme based on the resonance risk path segment data.

8. The multi-scale co-simulation system for thin-walled part cutting for chatter suppression according to claim 7, characterized in that, The system also includes: The fifth module is used to obtain the cutting parameters of the path points based on the chatter suppression trajectory simulation scheme, and to obtain the workpiece roughness and machining efficiency. Based on the cutting parameters, the workpiece roughness and the machining efficiency, a multi-objective optimization function is obtained. Based on the multi-objective optimization function, the NSGA-II algorithm is performed to obtain a set of co-simulation optimization parameters.

9. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method of any one of claims 1 to 6.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the method of any one of claims 1 to 6.

11. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by a processor, implements the method of any one of claims 1 to 6.

Images