Bogie numerical control machining error compensation method and system based on digital twinning

Abstract

The invention relates to a digital twinning-based bogie numerical control machining error compensation method and system, and the method comprises the steps: constructing an initial digital twinning model comprising material characteristics and geometric constraints through fusing the historical machining data of a bogie and the structural characteristics of a workpiece; according to the real-time cutting force and temperature data dynamic correction model, time-varying stiffness parameters are updated, and a deformation prediction result is generated; identifying an error exceeding area based on the precision requirement of the bearing seat assembly surface; and the feed rate compensation amount is calculated according to the thin-wall area deformation gradient, and a pre-compensation parameter scheme used for adjusting the cutting path is generated. The method overcomes the industrial problems that a traditional compensation technology cannot dynamically suppress instantaneous deformation, multi-source error coupling is difficult to decouple, and parameter adjustment depends on trial cutting verification, and the machining precision and efficiency of the bogie are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of CNC error compensation, and in particular to a bogie CNC machining error compensation method and system based on digital twins. Background Art

[0002] As the core load-bearing component of high-speed rail vehicles, the machining accuracy of key parts such as the bearing seats and side beams directly impacts the safety and stability of train operation. During CNC machining, machining errors are inevitable due to factors such as fluctuations in cutting forces, thermal deformation, and changes in workpiece rigidity caused by material removal. Traditional bogie machining error compensation methods rely primarily on post-machining three-dimensional coordinate measurement data, analyzing the measurement results to adjust subsequent machining parameters. This method establishes a statistical relationship between cutting parameters and errors, generates compensation instructions, and corrects the tool path, thereby reducing the impact of systematic errors.

[0003] However, existing technologies have significant limitations. First, compensation methods based on post-process measurement cannot respond to dynamic disturbances during machining. When cutting forces suddenly change or temperatures rise sharply, transient workpiece deformation cannot be suppressed in a timely manner. Second, due to the coupling of multiple sources of error, such as cutting force, thermal deformation, and structural vibration, it is difficult to accurately decouple the contribution of each factor to the total error, resulting in a lack of targeted compensation strategies. Finally, parameter adjustments require repeated trial cuts and verification, with a single adjustment cycle lasting several hours, seriously affecting production efficiency. These issues lead to defects such as excessive deformation in thin-walled areas and insufficient assembly surface contour accuracy, making it difficult to meet the stringent ±0.015mm precision requirements of high-speed rail bogies. Summary of the Invention

[0004] Based on this, the purpose of the present invention is to provide a bogie CNC machining error compensation method and system based on digital twins that can realize dynamic error prediction and real-time compensation.

[0005] The purpose of the present invention is achieved by the following scheme:

[0006] In a first aspect, the present invention provides a bogie CNC machining error compensation method based on digital twin, comprising the following steps:

[0007] S1: Process the acquired historical machining data and workpiece structural features of the bogie, build an initial virtual machining environment by integrating material property data with geometric structural features, and generate an initial digital twin model;

[0008] S2: Obtain real-time monitoring of cutting force and temperature data, combine it with the initial digital twin model for dynamic correction, update the model's time-varying stiffness parameters, and generate corrected deformation prediction results;

[0009] S3: Perform error risk assessment on the deformation prediction results, identify the risk areas of exceeding standards based on the machining accuracy requirements of the bearing seat assembly surface, and generate the coordinates of the error exceeding standards areas;

[0010] S4: Perform reverse engineering on the cutting parameters of the coordinates of the area where the error exceeds the standard, calculate the feed rate compensation amount based on the deformation gradient of the thin-walled area, and generate a pre-compensation parameter solution. The pre-compensation parameter solution is used to instruct the CNC system to adjust the cutting path.

[0011] In one embodiment, the present invention provides a bogie CNC machining error compensation method based on digital twin, S1 specifically comprising the following steps:

[0012] S11: extracting features from the cutting parameter sequence in the acquired historical processing data of the bogie, calling a time series analysis method to separate the process parameters from noise interference, and generating a historical process feature set;

[0013] S12: topological mapping is performed on the bearing seat surface geometry in the acquired workpiece structural features, a corresponding relationship between the three-dimensional point cloud and the finite element mesh is established, and a geometric constraint matrix is ​​generated;

[0014] S13: Physical property fusion is performed on the historical process feature set and the geometric constraint matrix, and the material stiffness attenuation model is coupled with the thermal expansion coefficient to generate an initial digital twin model. The initial digital twin model is used to indicate the dynamic deformation behavior during the bogie processing.

[0015] In one embodiment, the present invention provides a bogie CNC machining error compensation method based on digital twin S2, which specifically includes the following steps:

[0016] S21: Acquire real-time monitored cutting force data and perform dynamic load analysis, invert the instantaneous cutting force distribution through the spindle motor current signal, and generate a real-time cutting force vector field;

[0017] S22: Acquire real-time monitored temperature data, combine it with the real-time cutting force vector field to perform multi-physics field coupling, call the initial digital twin model to calculate the combined force-heat effect, and generate a dynamic load matrix;

[0018] S23: Perform time-varying stiffness update processing on the dynamic load matrix, call the recursive least squares algorithm to correct the stiffness attenuation coefficient of the material removal area, and generate a corrected deformation prediction result.

[0019] In one embodiment, S3 of a bogie CNC machining error compensation method based on digital twin provided by the present invention specifically includes the following steps:

[0020] S31: Extract the assembly surface area based on the deformation prediction results, lock the preset buffer range around the bearing seat installation hole, and generate the key analysis area;

[0021] S32: Compare tolerance zones for key analysis areas, calculate the excess profile deviation according to ISO 1101, and generate a heat map of excess risk.

[0022] S33: Perform continuous region identification processing on the risk heat map of exceeding the standard, extract the connected areas whose deviation values ​​exceed the preset exceeding standard threshold, and generate the coordinates of the error exceeding standard area.

[0023] In one embodiment, S4 of a bogie CNC machining error compensation method based on digital twin provided by the present invention specifically includes the following steps:

[0024] S41: Perform deformation gradient analysis on the thin-walled structure in the coordinates of the error-exceeding region, calculate the sensitivity coefficients of the curvature radius and feed speed of the thin-walled region, and generate a dynamic compensation coefficient matrix;

[0025] S42: performing parameter inverse solution on the dynamic compensation coefficient matrix, calculating the optimal feed rate adjustment amount of the compensation position point through the Jacobian matrix, and generating a theoretical compensation parameter set;

[0026] S43: Perform machine tool dynamic performance verification on the theoretical compensation parameter set, verify the acceleration limit and servo tracking capability of the CNC system's axial motion, and generate a pre-compensation parameter solution.

[0027] In a second aspect, the present invention provides a bogie CNC machining error compensation system based on digital twinning, which is configured with the following modules:

[0028] The initial twin construction module is used to process the acquired historical processing data and workpiece structural features of the bogie, build an initial virtual processing environment by fusing material property data with geometric structural features, and generate an initial digital twin model;

[0029] The model dynamic correction module is used to obtain real-time monitoring cutting force and temperature data, combine it with the initial digital twin model for dynamic correction, update the model's time-varying stiffness parameters, and generate corrected deformation prediction results;

[0030] The error risk assessment module is used to perform error risk assessment on the deformation prediction results, identify the risk areas of exceeding the standard based on the machining accuracy requirements of the bearing seat assembly surface, and generate the coordinates of the error exceeding the standard area;

[0031] The cutting parameter compensation module is used to reversely solve the cutting parameters of the coordinates of the area where the error exceeds the standard, calculate the feed rate compensation amount according to the deformation gradient of the thin-walled area, and generate a pre-compensation parameter scheme. The pre-compensation parameter scheme is used to instruct the CNC system to adjust the cutting path.

[0032] In a third aspect, the present application provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, it implements any of the above-mentioned digital twin-based bogie CNC machining error compensation methods.

[0033] In a fourth aspect, the present application provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, it implements any of the above-mentioned digital twin-based bogie CNC machining error compensation methods.

[0034] In summary, the digital twin-based bogie CNC machining error compensation method provided in this application can achieve high-precision virtual mapping of the machining environment by constructing a digital twin model that integrates material properties and geometric constraints, laying a physical foundation for error prediction; dynamically correcting time-varying stiffness parameters based on real-time cutting force and temperature data can achieve the effect of instantaneous response to cutting disturbances and thermal deformation, and effectively suppress sudden flexural deformation in thin-walled areas; combining the bearing seat assembly surface accuracy standard to identify the risk area of ​​exceeding the standard, it can decouple the multi-source error coupling effect and accurately locate the deformation out-of-tolerance area caused by sudden change in structural stiffness or concentrated thermal load; by inversely solving the feed rate compensation amount for the deformation gradient of the thin-walled area, a cutting path adjustment plan that directly interacts with the CNC system can be generated to achieve rapid deployment of the compensation strategy without trial cutting.

[0035] This method fundamentally solves the three major bottlenecks in traditional technologies: dynamic response lag, difficult error coupling analysis, and lengthy compensation cycles. It significantly improves the contour accuracy stability and processing efficiency of complex structural parts, and provides technical support for the high-precision processing needs in the field of rail transit equipment manufacturing.

[0036] For better understanding and implementation, the present invention is described in detail below with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 A schematic flow chart of a bogie CNC machining error compensation method based on digital twins provided in an embodiment of the present application;

[0038] Figure 2 A schematic diagram of a process for generating a pre-compensation parameter solution according to an embodiment of the present application;

[0039] Figure 3 A structural schematic diagram of a bogie CNC machining error compensation system based on digital twin is provided in another embodiment of the present application. DETAILED DESCRIPTION

[0040] To facilitate understanding of the present invention, the present invention will be described more fully below with reference to the accompanying drawings. The drawings illustrate preferred embodiments of the present invention. However, the present invention may be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided to provide a more thorough and comprehensive understanding of the disclosure.

[0041] Unless otherwise defined, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which this invention pertains. The terms used in this specification are for the purpose of describing specific embodiments only and are not intended to limit the invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0042] In one embodiment, Figure 1 As shown, a bogie CNC machining error compensation method based on digital twin is provided. This embodiment uses the method applied to a terminal as an example for illustration. It is understandable that the method can also be applied to a server, and can also be applied to a system including a terminal and a server, and implemented through the interaction between the terminal and the server. In this embodiment, the method includes the following steps:

[0043] S1: Process the historical processing data and workpiece structural features of the acquired bogie, build the initial virtual processing environment by integrating the material property data with the geometric structure features, and generate the initial digital twin model.

[0044] Specifically, historical machining data encompasses cutting parameters, error measurement records, tool status data, and machine tool operating parameters for different batches of bogies. The system converts this data into a unified data storage structure. Preferably, interpolation algorithms can be used to supplement missing values ​​to ensure data integrity. Workpiece structural features are extracted from the 3D design model, including the geometric dimensions, structural morphology, and material distribution of each component. The system converts this information into structured data to establish a geometric feature database.

[0045] Specifically, the system integrates material property data with geometric structural features and, through a data association algorithm, binds this data to the corresponding geometric structural components to form a geometric model that incorporates physical properties. Material property data includes elastic modulus, Poisson's ratio, thermal expansion coefficient, and more. Based on this data, the system constructs an initial virtual machining environment, which includes virtual mappings of the machine tool, cutting tool, and workpiece, as well as the physical field simulation parameters that may be involved in the machining process. Using multi-body dynamics modeling, the models of each component are assembled to generate an initial digital twin model that reflects the fundamental physical laws of the bogie machining process.

[0046] S2: Obtain real-time monitoring cutting force and temperature data, combine them with the initial digital twin model for dynamic correction, update the model's time-varying stiffness parameters, and generate corrected deformation prediction results.

[0047] Specifically, the system acquires real-time cutting force and temperature data from sensors deployed at the machining site. Cutting force data is acquired via high-precision cutting force sensors installed at key locations on the machine tool, enabling real-time sensing of changes in cutting force components in three directions. Temperature data is obtained by high-frequency sampling of the machining area by temperature sensors, enabling precise understanding of the temperature field distribution within the machining area. The data collected by the sensors is sent in real time to the system's data receiving module via a data transmission protocol. The system verifies the received data, removes abnormal data, decomposes the cutting force data to obtain force components in different directions, and reconstructs the spatial distribution of the temperature data to form temperature field data.

[0048] Specifically, the system couples real-time data with the initial digital twin model for calculations, invoking a model correction algorithm to adjust the cutting force parameters in the model based on the real-time cutting force data, ensuring that the cutting forces in the virtual model are consistent with the actual cutting forces. Simultaneously, the system corrects the heat conduction parameters in the model based on the temperature data and updates the model's temperature field distribution. During this adjustment process, the system calculates the model's time-varying stiffness parameters in real time, continuously optimizing the stiffness parameter values ​​by comparing the model's predicted deformation results with deformation trends inverted based on real-time data. After dynamic correction, the system generates updated deformation predictions that more accurately reflect the actual deformation during machining.

[0049] S3: Perform error risk assessment on the deformation prediction results, identify the risk area of ​​exceeding the standard based on the machining accuracy requirements of the bearing seat assembly surface, and generate the coordinates of the area of ​​exceeding the standard.

[0050] Specifically, machining accuracy requirements involve multiple key indicators such as flatness, perpendicularity, and position. Based on these indicators and the bogie's machining accuracy requirements, the permissible error range for each key component is determined. The predicted deformation results are then compared with the permissible range to calculate the error value for each component. Preferably, the system can employ a multi-factor analysis approach, comprehensively considering the impact of factors such as cutting force, temperature, and material properties on error, to assess the risk level of error.

[0051] For example, the system analyzes the error risk assessment results, identifies risk areas that may exceed the standard, and uses a spatial analysis algorithm to map the deformation prediction results to the three-dimensional coordinate system of the workpiece to determine the deformation position and range of each part. Among them, for areas where the error value exceeds the allowable range, the system can perform cluster analysis and merge adjacent exceeding-standard areas into continuous exceeding-standard area units. At the same time, the system associates the accuracy requirements of key parts such as the bearing seat assembly surface, classifies and marks the exceeding-standard areas, clarifies the degree of influence of different exceeding-standard areas on the processing accuracy, and generates coordinate data for the error exceeding-standard areas.

[0052] S4: Perform reverse engineering on the cutting parameters of the coordinates of the area where the error exceeds the standard, calculate the feed rate compensation amount based on the deformation gradient of the thin-walled area, and generate a pre-compensation parameter solution. The pre-compensation parameter solution is used to instruct the CNC system to adjust the cutting path.

[0053] Specifically, the system uses a parameter-solving algorithm, using the deformation in the error-exceeding area as a constraint and, combined with the physical model of the cutting process, reversely calculates the required cutting parameter adjustments. For thin-walled areas, the system establishes a mathematical relationship between the deformation gradient and the feed rate compensation based on the deformation gradient data, calculating the specific compensation value for the feed rate. Based on the calculated cutting parameter adjustments, the system generates a pre-compensation parameter solution containing parameters such as the adjusted feed rate, cutting speed, and tool path. The system then converts these parameters into an instruction format recognizable by the CNC system.

[0054] When generating instructions, the system considers the tool's motion constraints to ensure that the adjusted tool path does not interfere with non-machining areas of the workpiece. Once the pre-compensation parameter plan is generated, the system sends it to the CNC system via a data transmission interface. The CNC system adjusts the cutting path accordingly, achieving real-time compensation for machining errors and reducing the occurrence of errors exceeding the specified limit.

[0055] In summary, the digital twin-based bogie CNC machining error compensation method provided in this application can achieve high-precision virtual mapping of the machining environment by constructing a digital twin model that integrates material properties and geometric constraints, laying a physical foundation for error prediction; dynamically correcting time-varying stiffness parameters based on real-time cutting force and temperature data can achieve the effect of instantaneous response to cutting disturbances and thermal deformation, and effectively suppress sudden flexural deformation in thin-walled areas; combining the bearing seat assembly surface accuracy standard to identify the risk area of ​​exceeding the standard, it can decouple the multi-source error coupling effect and accurately locate the deformation out-of-tolerance area caused by sudden change in structural stiffness or concentrated thermal load; by inversely solving the feed rate compensation amount for the deformation gradient of the thin-walled area, a cutting path adjustment plan that directly interacts with the CNC system can be generated to achieve rapid deployment of the compensation strategy without trial cutting.

[0056] This method fundamentally solves the three major bottlenecks in traditional technologies: dynamic response lag, difficult error coupling analysis, and lengthy compensation cycles. It significantly improves the contour accuracy stability and processing efficiency of complex structural parts, and provides technical support for the high-precision processing needs in the field of rail transit equipment manufacturing.

[0057] In one embodiment, the present invention provides a bogie CNC machining error compensation method based on digital twin, S1 specifically comprising the following steps:

[0058] S11: Extract features from the cutting parameter sequence in the acquired historical processing data of the bogie, call the time series analysis method to separate the process parameters and noise interference, and generate a historical process feature set.

[0059] Specifically, the system receives a sequence of cutting parameters from the bogie's historical machining data, and the sequence includes continuous values ​​of cutting speed, feed rate, and cutting depth over machining time. Preferably, the system preprocesses the cutting parameter sequence, identifies abnormal values ​​that exceed a preset range through a sliding window algorithm, and the length of the sliding window is determined according to the average duration of the machining process. The parameter mean and standard deviation σ are calculated within the window. Values ​​outside the 3σ range are marked as abnormal values, and the system calls a linear interpolation function to calculate a replacement value using a fitted straight line of three valid data points before and after the abnormal value. Preferably, the system can call a discrete Fourier transform to decompose the preprocessed sequence, and the transformation formula is:

[0060]

[0061] Wherein, x(n) is the discrete time series value, N is the sequence length, and k is the frequency component index. The frequency threshold f0 is set according to the natural frequency range of the processing equipment, and the components higher than f0 in the frequency domain are filtered out. The retained low-frequency components are converted back to the time domain by inverse Fourier transform to obtain a process parameter sequence with noise interference removed. The system divides the process parameter sequence into work steps, calculates the mean, variance and maximum change rate within each work step, and extracts the starting value and ending value of the parameter sequence. These feature quantities are associated with the work step number and the processing part identification, and are stored in the form of a structured data table to form a historical process feature set. Each record in the data table corresponds to the complete process parameter characteristics of a work step. Preferably, the calculation formulas for the mean, variance and maximum change rate are:

[0062]

[0063] Among them, μ is the mean, σ 2 is the variance, N-1 is the degree of freedom, r max is the maximum rate of change, x(t i+1 )、x(t i ) is the adjacent time t i+1 , t iParameter value.

[0064] S12: Perform topological mapping on the bearing seat surface geometry in the acquired workpiece structural features, establish a corresponding relationship between the three-dimensional point cloud and the finite element mesh, and generate a geometric constraint matrix.

[0065] Specifically, the system imports the 3D design model of the bearing seat surface and obtains the surface point cloud data through the 3D scanning protocol. The point cloud density is dynamically adjusted according to the surface curvature. The point cloud density increases accordingly in areas with large curvature. The expression of the surface point cloud data is:

[0066] P={p m |p m =(x m ,y m ,z m ),m=1,2,...,M}

[0067] Among them, P is the surface point cloud data, p m is the mth point in the point cloud, (x m ,y m ,z m ) is p m The system performs voxel filtering on P, setting the side length of the cube voxel to one tenth of the surface accuracy requirement. It then traverses all point clouds, retaining the point closest to the voxel center within each voxel and deleting the remaining points, thus reducing the amount of data while ensuring the integrity of the surface features. The system loads the preset finite element mesh model G:

[0068] G={g n |g n =(x n ,y n ,z n ),n=1,2,...,N}

[0069] Among them, g n is the nth node in the grid, (x n ,y n ,z n ) is g n The three-dimensional coordinates of the grid unit type are selected according to the surface structure. Quadrilateral units are used in the smooth area of ​​the surface, and triangular units are used in the area with sudden curvature. Preferably, the system can call the K nearest neighbor algorithm to calculate the distance between each point cloud point p m Calculate the g of all grid nodes n The Euclidean distance d mn :

[0070]

[0071] Select the k nodes with the smallest distance as the neighboring nodes, and the k value is set according to the grid density. The influence weight w of the point cloud point on each neighboring node is calculated by the inverse distance weighted method. mn :

[0072]

[0073] The weight of non-neighborhood nodes is set to 0. The system generates an M×N order geometric constraint matrix C, where the element in the mth row and nth column of the matrix is ​​w mn , the matrix is ​​stored in sparse matrix format, retaining only non-zero elements and their indices to reduce storage space.

[0074] S13: Physical property fusion is performed on the historical process feature set and the geometric constraint matrix, and the material stiffness attenuation model is coupled with the thermal expansion coefficient to generate an initial digital twin model. The initial digital twin model is used to indicate the dynamic deformation behavior during the bogie processing.

[0075] Specifically, the system reads the historical process feature set and the geometric constraint matrix C, and assigns the process parameter characteristics of each step to the corresponding grid unit through the mapping relationship between the step number and the grid unit coordinates. The system loads the material stiffness attenuation model, and the model parameters are initialized through the material performance records in the historical processing data, where the initial stiffness K0 is the product of the elastic modulus of the material and the cross-sectional area of ​​the grid unit, and the attenuation coefficient α is retrieved from the material database according to the material type. The material removal amount V is calculated by the cutting parameters in the historical process feature set, and the cutting depth a of each step is calculated by the cutting depth a of each step. p , feed rate f, cutting speed v c Integrating on the time axis, the formula is:

[0076]

[0077] Where t1 and t2 are the start and end times of the work step, respectively. Substituting them into the stiffness calculation formula, we can obtain the real-time stiffness K(V) of each grid element:

[0078] K(V)=K0·exp(-αV)

[0079] The system retrieves the thermal expansion coefficient α from the material database T , combined with the temperature records in the historical process feature set, the linear expansion of each grid unit at the processing temperature T is calculated:

[0080] ΔL=L0·α T (T-T0)

[0081] Where L0 is the initial edge length of the mesh element, and T0 is the ambient temperature. The system substitutes the mesh element stiffness K(V) and linear expansion ΔL into the geometric constraint matrix C to construct an overall stiffness matrix. Combined with the cutting force load vector F obtained from the cutting parameter mapping, the finite element solver solves the equilibrium equation K·U=F to obtain the displacement vector U for each node.

[0082] The system integrates the grid model, overall stiffness matrix, load vector, displacement solution interface and parameter input interface to generate an initial digital twin model. The model can receive new process parameter inputs, calculate and output the corresponding node displacement distribution data, i.e., deformation field distribution, by calling the internal solver.

[0083] In one embodiment, the present invention provides a bogie CNC machining error compensation method based on digital twin S2, which specifically includes the following steps:

[0084] S21: Acquire the real-time monitored cutting force data and perform dynamic load analysis, invert the instantaneous cutting force distribution through the spindle motor current signal, and generate a real-time cutting force vector field.

[0085] Specifically, the system obtains real-time monitoring cutting force data, which is collected by sensors installed on the spindle and tool holder of the machine tool, and covers the force value information of the contact area between the tool and the workpiece during the processing. The system performs dynamic load analysis on the collected cutting force data, removes high-frequency vibration interference in the data through filtering, and retains the effective signal reflecting the actual cutting state. Preferably, the system can call the spindle motor current signal, establish a mapping relationship between the current signal and the cutting force, and invert the magnitude and direction of the instantaneous cutting force by analyzing the change law of the current signal. Based on the instantaneous cutting force data obtained by inversion, combined with the relative position relationship between the tool and the workpiece, a real-time cutting force vector field is constructed in three-dimensional space. The vector parameters of each point in the vector field include the magnitude, direction and coordinates of the cutting force, which can fully present the distribution of the cutting force in the processing area.

[0086] S22: Acquire real-time monitored temperature data, combine it with the real-time cutting force vector field for multi-physics field coupling, call the initial digital twin model to calculate the force-heat joint effect, and generate a dynamic load matrix.

[0087] Specifically, the system receives real-time temperature data collected by temperature sensors distributed on the surface and inside the workpiece. The data contains instantaneous temperature values ​​at multiple measurement points. Preferably, the system performs spatial interpolation on the temperature data, using an interpolation algorithm to expand the discrete temperature measurements to the entire three-dimensional space of the workpiece, forming continuous temperature field distribution data. The system then calls the multi-physics field coupling module in the initial digital twin model and inputs the real-time cutting force vector field and temperature field data into the module. The module calculates the heat generated during the cutting process based on the work done by the cutting force and, combined with the thermal conductivity characteristics of the material, calculates the transfer and distribution of temperature within the workpiece.

[0088] At the same time, the system considers the impact of temperature-induced material property changes on the effects of cutting forces, as well as the reaction of workpiece deformation caused by cutting forces on temperature distribution, achieving bidirectional force-heat coupling calculations. Through calculations, the magnitude and direction of the load borne by each grid cell under the combined effects of force and heat are determined. The system arranges this load data by grid cell number and time step to generate a dynamic load matrix. Each element in the matrix corresponds to the load information for a specific grid cell at a specific time step.

[0089] S23: Perform time-varying stiffness update processing on the dynamic load matrix, call the recursive least squares algorithm to correct the stiffness attenuation coefficient of the material removal area, and generate a corrected deformation prediction result.

[0090] Specifically, the system reads the dynamic load matrix and the stiffness parameters from the initial digital twin model, and determines the areas requiring stiffness updates based on the material removal during machining. Preferably, the system uses a recursive least squares algorithm, based on historical data on material removal and stiffness changes in the corresponding areas. Initial weights and forgetting factors are set for the algorithm, and the coordinates of the material removal areas, collected in real time, and the load values ​​from the dynamic load matrix, are fed into the algorithm. Through continuous iteration, the algorithm corrects the stiffness attenuation coefficients for the material removal areas.

[0091] During the correction process, the algorithm compares the stiffness values ​​measured during actual machining with the stiffness values ​​predicted by the model. Based on the deviation, the attenuation coefficient is adjusted to ensure that the predicted stiffness values ​​are consistent with the actual measured values. The system applies the corrected stiffness attenuation coefficient to the corresponding mesh elements, updating the stiffness parameters of these elements. Using the updated stiffness parameters and the dynamic load matrix, the finite element solver calculates the displacement of each mesh node of the workpiece. This displacement data is converted into deformation values ​​for each part of the workpiece, generating a corrected deformation prediction result. The result is stored as a three-dimensional deformation distribution diagram containing deformation information for each position of the workpiece.

[0092] In one embodiment, S3 of a bogie CNC machining error compensation method based on digital twin provided by the present invention specifically includes the following steps:

[0093] S31: Extract the assembly surface area based on the deformation prediction results, lock the preset buffer range around the bearing seat installation hole, and generate the key analysis area.

[0094] Specifically, the system retrieves the design coordinates of the bearing seat mounting hole from the database of the bogie three-dimensional design model. The coordinate information includes the three-dimensional coordinates of the hole center and the axial direction vector of the hole. The design coordinates are converted into a coordinate system consistent with the deformation prediction results, and the corresponding grid node set of the mounting hole in the deformation prediction results is located through coordinate matching.

[0095] Preferably, the system sets calculation rules for the buffer zone based on the functional requirements of the mounting hole. The rules are determined based on the hole diameter, the clearance during assembly, and the structural dimensions of the surrounding associated parts. The buffer zone is a spherical area centered on the hole center. The system traverses all mesh nodes in the deformation prediction results, calculates the spatial distance between each node and the hole center, and filters out nodes whose distance is less than or equal to the buffer zone radius. The filtered node data includes the node number, three-dimensional coordinates, and the corresponding displacement values ​​in three directions. These data are sorted by spatial position to form a continuous regional data set, namely the key analysis area. This regional data serves as the basic data for subsequent tolerance analysis.

[0096] S32: Compare tolerance zones for key analysis areas, calculate the contour deviation exceeding the standard according to ISO 1101, and generate a heat map of the exceeding standard risk.

[0097] Specifically, the system reads all node data in the key analysis area and retrieves the definition and calculation details of the profile deviation in the ISO1101 standard from the standard database. Specifically, based on the three-dimensional design model of the bogie, the system extracts the theoretical profile data of the key analysis area. The theoretical profile data consists of a series of continuous three-dimensional coordinate points that describe the expected geometric shape of the area. By adding the design coordinates of each node in the key analysis area to the displacement component in the deformation prediction result, the actual position coordinates of the node are obtained, and the shortest distance between the actual position coordinates and the theoretical profile is calculated. This distance is the profile deviation value of the node. During the calculation process, a distance algorithm from a spatial point to a curve can be used to traverse all points on the theoretical profile and take the minimum distance as the result.

[0098] Preferably, the system retrieves the tolerance band data for the critical analysis area from the process file database. The tolerance band data includes upper and lower deviation values. The system compares the contour deviation value of each node with the tolerance band data to determine whether the deviation value exceeds the tolerance band range. The value of the excess portion is the overscalar. The spatial coordinates of the critical analysis area are used as the coordinate axis, and the overscalar of the node is used as the variable. The spatial area between the nodes is filled with an interpolation algorithm to form continuous overscalar distribution data. The system converts the overscalar distribution data into image data. The color of each pixel is determined by the overscalar at the corresponding position, and a heat map of overscalar risk is generated. The coordinate range of the heat map is consistent with the spatial range of the critical analysis area.

[0099] S33: Perform continuous region identification processing on the risk heat map of exceeding the standard, extract the connected areas whose deviation values ​​exceed the preset exceeding standard threshold, and generate the coordinates of the error exceeding standard area.

[0100] Specifically, the system retrieves a preset threshold for exceeding the standard from the process requirement database. This threshold is determined based on the bogie's assembly accuracy requirements and the functional performance indicators of the bearing seat. Specifically, the system scans the underlying data of the exceeding-standard risk heat map point by point, determines whether the deviation exceeding the standard value of each sampling point is greater than the preset exceeding-standard threshold, and records the coordinates of all sampling points that meet the conditions. The recorded sampling point coordinates are then analyzed for spatial continuity using a neighborhood judgment method. Starting from a sampling point, the system checks the adjacent sampling points in the six directions around it to see if they are exceeding the standard. If so, the points are grouped into the same area, and the inspection range is expanded starting from the newly included points until no new exceeding-standard points can be included.

[0101] For each continuous region, the system calculates the coordinate extremes of all sampling points within the region. These extremes include the maximum and minimum values ​​along the three coordinate axes, and these extremes constitute the region's boundary coordinates. The system organizes the boundary coordinates of all continuous regions by region number, and each region's boundary coordinates are represented by six numerical values, forming a coordinate set for the region where the error exceeds the standard.

[0102] In one embodiment, Figure 2 As shown, S4 of the bogie CNC machining error compensation method based on digital twin provided by the present invention specifically includes the following steps:

[0103] S41: Perform deformation gradient analysis on the thin-walled structure in the coordinates of the error-exceeding area, calculate the sensitivity coefficient of the curvature radius and feed speed of the thin-walled area, and generate a dynamic compensation coefficient matrix.

[0104] Specifically, the system performs deformation gradient analysis on the thin-walled structure in the coordinates of the error-exceeding area. By extracting the deformation data of discrete points on the surface of the thin-walled structure, the deformation difference between two adjacent points is calculated, and then the difference is divided by the spatial distance between the two points to obtain the rate of change of deformation along the thickness direction and extension direction of the thin wall. The rate of change reflects the distribution gradient of the deformation on the thin-walled structure.

[0105] At the same time, the system extracts the radius of curvature of the thin-walled area and establishes a surface equation by curve fitting discrete points on the thin-walled surface. Based on this surface equation, the radius of curvature of each point is calculated. The radius of curvature reflects the degree of curvature of the thin-walled structure. The computer system establishes a correlation model between the radius of curvature of the thin-walled area and the feed speed. By inputting different feed speed parameters, the corresponding changes in the radius of curvature are simulated and calculated. The numerical relationship between the two is analyzed and the sensitivity coefficient is calculated. The sensitivity coefficient is used to quantify the impact of feed speed changes on the radius of curvature.

[0106] Finally, the system matches the calculated sensitivity coefficients with the coordinates of each position in the thin-walled structure one by one, and arranges them in coordinate order to form a dynamic compensation coefficient matrix. The value of each element in the matrix corresponds to the compensation coefficient of a specific position in the thin-walled area. These coefficients reflect the sensitivity of different positions to feed speed adjustment.

[0107] S42: Perform parameter inverse solution on the dynamic compensation coefficient matrix, calculate the optimal feed rate adjustment amount of the compensation position point through the Jacobian matrix, and generate a theoretical compensation parameter set.

[0108] Specifically, the system performs inverse parameter solution on the dynamic compensation coefficient matrix, takes the deformation amount that needs to be eliminated in the error-exceeding area as the target value, sets the feed rate adjustment amount as the variable to be solved, and establishes a mathematical function relationship between the deformation amount and the feed rate adjustment amount, and uses the dynamic compensation coefficient matrix as the coefficient matrix of the function.

[0109] The system calls the Jacobian matrix, which is constructed by taking partial derivatives of the above functional relationship. The elements in the matrix reflect the local rate of change of the deformation to the feed rate adjustment. The system uses the Jacobian matrix for iterative calculation, initially sets the initial value of the feed rate adjustment, substitutes it into the functional relationship to calculate the corresponding deformation, compares the calculated result with the target value, adjusts the feed rate adjustment according to the difference, and repeats the iterative process until the difference between the calculated deformation and the target value is within the allowable range. Preferably, the calculation formula for the rate adjustment is:

[0110]

[0111] Where Δv f is the feed rate adjustment, J -1 is the inverse matrix of the sensitivity matrix, is the maximum deformation gradient, R c is the local curvature radius.

[0112] The system associates the final feed rate adjustment amount with the corresponding compensation position point coordinates, and organizes them in position order to form a theoretical compensation parameter set. Each entry in the theoretical compensation parameter set contains the three-dimensional coordinates of the compensation position point and the corresponding optimal feed rate adjustment amount.

[0113] S43: Perform machine tool dynamic performance verification on the theoretical compensation parameter set, verify the acceleration limit and servo tracking capability of the CNC system's axial motion, and generate a pre-compensation parameter solution.

[0114] Specifically, the system verifies the dynamic performance of the machine tool against the theoretical compensation parameter set. By accessing the machine tool's control system interface and calling the machine tool's parameter database, the system extracts the acceleration limit for the CNC system's axial motion. This acceleration limit is determined by the machine tool's mechanical structure and drive system performance. Based on the feed rate adjustment in the theoretical compensation parameter set and the motion trajectory of each axis, the system calculates the acceleration change curve for each axis during the execution of the compensation instruction. The values ​​on the curve are compared with the acceleration limit to determine whether the acceleration is within the machine tool's tolerance range. Simultaneously, the system verifies the CNC system's servo tracking capability by simulating the feed rate change curve in the theoretical compensation parameter set. This curve is then input into the servo system model as a command and the actual output feed rate of the servo system is calculated. The difference between the two is the servo tracking error.

[0115] Based on the acceleration limit verification results and the servo tracking capability verification results, the system corrects the feed rate adjustment in the theoretical compensation parameter set to ensure that the corrected parameters will not cause the machine tool to exceed its dynamic performance range. The corrected parameters are organized into an ordered set according to the sequence of compensation positions to generate a pre-compensation parameter scheme. The pre-compensation parameter scheme contains the coordinates of each compensation position, the corresponding feed rate adjustment amount and the execution order of the compensation instructions. It can be directly read by the CNC system and converted into a control signal for adjusting the cutting path.

[0116] In summary, the digital twin-based bogie CNC machining error compensation method provided in this application can achieve precise decoupling of the deformation mechanism of thin-walled areas through deformation gradient analysis, quantify the sensitive relationship between the curvature radius and the feed speed, and generate a dynamic compensation coefficient matrix reflecting the structural weaknesses; inverse parameter solution based on the matrix can achieve the effect of generating optimal compensation parameters without trial cutting, and establish a quantitative mapping of deformation gradient and feed rate through the Jacobian matrix to achieve precise rate adjustment of the compensation position point; dynamic machine tool calibration of theoretical parameters can ensure the matching of the compensation scheme with the equipment performance, and verify the acceleration limit and servo tracking capability to avoid the risk of over-compensation or under-compensation.

[0117] It should be understood that, although the various steps in the flowcharts involved in the various embodiments described above are displayed in sequence according to the instructions of the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and these steps can be executed in other orders. Moreover, at least a portion of the steps in the flowcharts involved in the various embodiments described above can include multiple steps or multiple stages, and these steps or stages are not necessarily executed and completed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a portion of steps or stages in other steps.

[0118] Based on the same inventive concept, the embodiments of the present application also provide a bogie CNC machining error compensation system based on digital twins for implementing the above-mentioned bogie CNC machining error compensation method based on digital twins. The implementation solution provided by this system is similar to the implementation solution described in the above-mentioned method. Therefore, the specific limitations of one or more embodiments of the bogie CNC machining error compensation system based on digital twins provided below can be referred to the limitations of the bogie CNC machining error compensation method based on digital twins above, and will not be repeated here.

[0119] Preferably, if Figure 3 As shown, the present invention provides a bogie CNC machining error compensation system 500 based on digital twin, which is configured with the following modules:

[0120] The initial twin construction module 510 is used to process the acquired historical machining data and workpiece structural features of the bogie, construct an initial virtual machining environment by fusing material property data with geometric structural features, and generate an initial digital twin model;

[0121] The model dynamic correction module 520 is used to obtain the real-time monitoring cutting force and temperature data, perform dynamic correction in combination with the initial digital twin model, update the time-varying stiffness parameters of the model, and generate the corrected deformation prediction results;

[0122] The error risk assessment module 530 is used to perform error risk assessment on the deformation prediction results, identify the risk areas of exceeding the standard in combination with the machining accuracy requirements of the bearing seat assembly surface, and generate the coordinates of the error exceeding the standard area;

[0123] The cutting parameter compensation module 540 is used to reversely solve the cutting parameters of the coordinates of the error-exceeding area, calculate the feed rate compensation amount according to the deformation gradient of the thin-walled area, and generate a pre-compensation parameter solution. The pre-compensation parameter solution is used to instruct the CNC system to adjust the cutting path.

[0124] Preferably, the initial twin building module 510 provided in this application is configured with the following units:

[0125] A process feature extraction unit is used to extract features from the cutting parameter sequence in the acquired historical processing data of the bogie, call a time series analysis method to separate the process parameters from noise interference, and generate a historical process feature set;

[0126] The geometric topology mapping unit is used to perform topological mapping on the bearing seat surface geometry in the acquired workpiece structural features, establish the corresponding relationship between the three-dimensional point cloud and the finite element mesh, and generate the geometric constraint matrix;

[0127] The physical property fusion unit is used to fuse the physical properties of the historical process feature set and the geometric constraint matrix, couple the material stiffness attenuation model with the thermal expansion coefficient, and generate an initial digital twin model. The initial digital twin model is used to indicate the dynamic deformation behavior during the bogie processing.

[0128] Preferably, the model dynamic correction module 520 provided in this application is configured with the following units:

[0129] The cutting force analysis unit is used to obtain real-time monitoring cutting force data and perform dynamic load analysis. It inverts the instantaneous cutting force distribution through the spindle motor current signal to generate a real-time cutting force vector field.

[0130] The multi-field coupling calculation unit is used to obtain real-time monitored temperature data, combine it with the real-time cutting force vector field to perform multi-physics field coupling, call the initial digital twin model to calculate the combined force-heat effect, and generate a dynamic load matrix;

[0131] The stiffness correction unit is used to perform time-varying stiffness update processing on the dynamic load matrix, call the recursive least squares algorithm to correct the stiffness attenuation coefficient of the material removal area, and generate the corrected deformation prediction result.

[0132] Preferably, the error risk assessment module 530 provided in this application is configured with the following units:

[0133] The assembly surface area locking unit is used to extract the assembly surface area based on the deformation prediction results, lock the preset buffer range around the bearing seat installation hole, and generate the key analysis area;

[0134] Tolerance deviation calculation unit, used to compare tolerance bands in key analysis areas, calculate the contour deviation exceeding the standard according to ISO 1101 standard, and generate a heat map of exceeding standard risk;

[0135] The exceeding standard area identification unit is used to perform continuous area identification processing on the exceeding standard risk heat map, extract the connected areas whose deviation values ​​exceed the preset exceeding standard threshold, and generate the coordinates of the error exceeding standard area.

[0136] Preferably, the cutting parameter compensation module 540 provided in this application is configured with the following units:

[0137] The deformation gradient analysis unit is used to perform deformation gradient analysis on the thin-walled structure in the coordinates of the error-exceeding area, calculate the sensitivity coefficient of the curvature radius and feed speed of the thin-walled area, and generate a dynamic compensation coefficient matrix;

[0138] The compensation parameter solving unit is used to perform parameter inverse solution on the dynamic compensation coefficient matrix, calculate the optimal feed rate adjustment amount of the compensation position point through the Jacobian matrix, and generate a theoretical compensation parameter set;

[0139] The performance verification and adjustment unit is used to verify the dynamic performance of the machine tool based on the theoretical compensation parameter set, verify the acceleration limit and servo tracking capability of the axial motion of the CNC system, and generate a pre-compensation parameter solution.

[0140] In one embodiment, the present application also provides a computer device including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the above-mentioned digital twin-based bogie CNC machining error compensation method is implemented.

[0141] In one embodiment, the present application also provides a computer-readable storage medium having a computer program stored thereon, which implements the above-mentioned digital twin-based bogie CNC machining error compensation method when the computer program is executed by a processor.

[0142] In the description of this specification, the reference terms "one embodiment," "some embodiments," "example," "specific example," or "some examples" mean that the specific features, structures, materials, or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. Moreover, the specific features, structures, materials, or characteristics described may be combined in any appropriate manner in any one or more embodiments or examples. In addition, those skilled in the art may combine and integrate different embodiments or examples described in this specification, as well as features of different embodiments or examples, unless they are mutually inconsistent.

[0143] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to the partial description of the method embodiments. The device embodiments described above are merely illustrative, wherein the components described as separate parts may or may not be physically separated, and the parts displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the disclosed solution. A person of ordinary skill in the art can understand and implement it without expending creative work.

[0144] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any person skilled in the art can easily conceive of various modifications or substitutions within the technical scope disclosed in this application, and such modifications or substitutions should be included within the scope of protection of the present application. Therefore, the scope of protection of the present application should be based on the scope of protection of the claims.

Claims

1. A bogie CNC machining error compensation method based on digital twin, characterized in that: The following steps are involved: S1: Process the acquired historical machining data and workpiece structural features of the bogie, build an initial virtual machining environment by integrating material property data with geometric structural features, and generate an initial digital twin model; S2: Acquire real-time monitored cutting force and temperature data, perform dynamic correction in combination with the initial digital twin model, update the time-varying stiffness parameters of the model, and generate a corrected deformation prediction result; S3: performing error risk assessment on the deformation prediction result, identifying the risk area of ​​exceeding the standard in combination with the machining accuracy requirement of the bearing seat assembly surface, and generating the coordinates of the area of ​​exceeding the standard; S4: performing reverse solution of cutting parameters for the coordinates of the error-exceeding region, calculating the feed rate compensation amount according to the deformation gradient of the thin-walled region, and generating a pre-compensation parameter scheme, wherein the pre-compensation parameter scheme is used to instruct the numerical control system to adjust the cutting path.

2. The method according to claim 1, characterized in that Said S1 comprises: S11: extracting features from the cutting parameter sequence in the acquired historical processing data of the bogie, calling a time series analysis method to separate the process parameters from noise interference, and generating a historical process feature set; S12: topological mapping is performed on the bearing seat surface geometry in the acquired workpiece structural features, a corresponding relationship between the three-dimensional point cloud and the finite element mesh is established, and a geometric constraint matrix is ​​generated; S13: Perform physical property fusion on the historical process feature set and the geometric constraint matrix, couple the material stiffness attenuation model and the thermal expansion coefficient, and generate an initial digital twin model. The initial digital twin model is used to indicate the dynamic deformation behavior during the bogie processing.

3. The method according to claim 1, characterized in that The S2 includes: S21: Acquire real-time monitored cutting force data and perform dynamic load analysis, invert the instantaneous cutting force distribution through the spindle motor current signal, and generate a real-time cutting force vector field; S22: acquiring real-time monitored temperature data, performing multi-physics field coupling in combination with the real-time cutting force vector field, calling the initial digital twin model to calculate the force-heat combined effect, and generating a dynamic load matrix; S23: performing time-varying stiffness update processing on the dynamic load matrix, calling a recursive least squares algorithm to correct the stiffness attenuation coefficient of the material removal area, and generating a corrected deformation prediction result.

4. The method according to claim 1, wherein The S3 includes: S31: extracting the assembly surface area based on the deformation prediction result, locking the preset buffer range around the bearing seat installation hole, and generating a key analysis area; S32: performing tolerance zone comparison on the key analysis area, calculating the excess of profile deviation according to ISO 1101 standard, and generating an excess risk heat map; S33: Performing continuous region recognition processing on the risk heat map of exceeding the standard, extracting connected regions whose deviation values ​​exceed a preset exceeding standard threshold, and generating coordinates of error exceeding standard regions.

5. The method according to any one of claims 1 to 4, characterized in that The S4 includes: S41: performing deformation gradient analysis on the thin-walled structure in the coordinates of the error-exceeding region, calculating the sensitivity coefficients of the curvature radius and feed speed of the thin-walled region, and generating a dynamic compensation coefficient matrix; S42: performing parameter inverse solution on the dynamic compensation coefficient matrix, calculating the optimal feed rate adjustment amount of the compensation position point through the Jacobian matrix, and generating a theoretical compensation parameter set; S43: Performing a machine tool dynamic performance check on the theoretical compensation parameter set, verifying the acceleration limit and servo tracking capability of the axial motion of the numerical control system, and generating a pre-compensation parameter solution.

6. A bogie CNC machining error compensation system based on digital twin, characterized in that: The system comprises: The initial twin construction module is used to process the acquired historical processing data and workpiece structural features of the bogie, build an initial virtual processing environment by fusing material property data with geometric structural features, and generate an initial digital twin model; A model dynamic correction module is used to obtain real-time monitoring cutting force and temperature data, perform dynamic correction in combination with the initial digital twin model, update the time-varying stiffness parameters of the model, and generate a corrected deformation prediction result; An error risk assessment module is used to perform error risk assessment on the deformation prediction result, identify the risk area of ​​exceeding the standard in combination with the machining accuracy requirements of the bearing seat assembly surface, and generate the coordinates of the error exceeding the standard area; The cutting parameter compensation module is used to reversely solve the cutting parameters of the coordinates of the error-exceeding area, calculate the feed rate compensation amount according to the deformation gradient of the thin-walled area, and generate a pre-compensation parameter scheme. The pre-compensation parameter scheme is used to instruct the CNC system to adjust the cutting path.

7. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the method according to any one of claims 1 to 5 is implemented.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 5 is implemented.

Images