Dynamic thermal error compensation method and system for double-moving-column numerical control gantry milling machine

Abstract

The invention relates to the technical field of numerical control planer mills, in particular to a dynamic thermal error compensation method and system for a double-moving-column numerical control planer miller, and the system comprises a plurality of temperature sensors which are distributed in a key heat sensitive area of a machine tool, and a laser range finder and an encoder which are used for monitoring the displacement of key parts. Synchronously collecting temperature data and displacement data of the machine tool in real time; according to the thermal error monitoring method and the thermal error monitoring system, through multi-sensor data fusion and Kalman filtering denoising, the reliability and the signal-to-noise ratio of thermal error monitoring data are remarkably improved; a nonlinear mapping relation between a temperature field and complex machine tool deformation can be accurately described by utilizing a thermal error prediction model, the problem that compensation is not timely due to temperature lag in a traditional method is fundamentally solved, a tool path is corrected in real time through a numerical control system, online and dynamic software compensation is achieved, and the real-time compensation precision is improved. Therefore, the machining precision and the consistency of the workpiece are greatly improved.

Description

Technical Field

[0001] This invention relates to the field of CNC gantry milling technology, and in particular to a dynamic thermal error compensation method and system for a double-moving-column CNC gantry milling machine. Background Technology

[0002] As a core piece of equipment for machining high-precision, large, and complex parts, the machining accuracy of CNC gantry milling machines is a key indicator for measuring the level of manufacturing. However, during long-term operation, heat sources inside the machine tool (such as the spindle motor, drive system, and friction pairs) generate a large amount of heat, causing uneven thermal deformation in key components such as the bed, column, crossbeam, and spindle. This thermal deformation directly changes the relative position of the tool and the workpiece, resulting in thermal errors, which has become a major bottleneck restricting the stability and consistency of machining accuracy.

[0003] Currently, most methods for compensating for thermal errors have limitations. For example, methods based on simplified physical models or empirical formulas are difficult to accurately describe the complex thermo-mechanical coupling dynamic characteristics of machine tools, resulting in low model accuracy and poor adaptability. On the other hand, some temperature-based compensation strategies suffer from inherent delays in heat conduction, causing the compensation signal to lag behind the actual error and failing to achieve true dynamic compensation, thus significantly reducing the compensation effect.

[0004] Therefore, overcoming the temperature hysteresis effect and achieving accurate prediction and real-time dynamic compensation of thermal errors is a technical problem that urgently needs to be solved in this field. Summary of the Invention

[0005] The purpose of this invention is to address the aforementioned shortcomings in the prior art by proposing a dynamic thermal error compensation method and system for a dual-moving column CNC gantry milling machine.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: The present invention discloses a dynamic thermal error compensation method for a dual-moving-column CNC gantry milling machine in its first aspect, comprising: The machine tool's temperature and displacement data are collected synchronously and in real time by using multiple temperature sensors distributed in the key heat-sensitive areas of the machine tool, as well as laser rangefinders and encoders used to monitor the displacement of key components. The displacement data is fused and denoised using the Kalman filter algorithm to obtain preprocessed displacement data. At the same time, the temperature data is standardized to obtain preprocessed temperature data. The preprocessed temperature data is input into the preset thermal error prediction model to calculate the real-time thermal error compensation. Based on the calculated thermal error compensation, control commands are generated to drive the machine tool CNC system to adjust the relative position of the tool and the workpiece in real time.

[0007] Furthermore, in the data acquisition step, the temperature sensor is a K-type thermocouple, and the temperature sensor is sequentially arranged on the spindle box, the lead screw and nut pair, and the bed guide rail.

[0008] Furthermore, the thermal error prediction model is a regression model based on machine learning, and its specific training steps are as follows: Collect historical temperature data and corresponding thermal error data of machine tools under different operating conditions to construct a training sample set; The temperature data in the training samples were standardized and dimensionality was reduced using principal component analysis to eliminate multicollinearity and reduce input variables. Using the principal component scores after dimensionality reduction as input and the corresponding thermal error as output, a thermal error prediction model is trained using the support vector regression algorithm, and the model parameters are optimized using the cross-validation method.

[0009] Furthermore, the principal component analysis method includes: Calculate the covariance matrix of the standardized temperature data and its eigenvalues ​​and eigenvectors; The number of principal components is selected based on the preset cumulative variance contribution rate not being lower than a threshold. Based on the selected principal components and their eigenvectors, the original temperature data is transformed into dimensionality-reduced principal component score vectors.

[0010] Furthermore, the method also includes an online model correction step: The preprocessed displacement data is used as the observed value of the real-time thermal error truth value; The compensation calculated by the thermal error prediction model is compared with the displacement data observations to generate residuals; Based on the residuals, the parameters of the thermal error prediction model are dynamically updated using the recursive least squares method to achieve adaptive online correction of the model.

[0011] Furthermore, it also includes a compensation effect evaluation step: after the compensation machining is completed, a coordinate measuring machine is used to inspect the reference workpiece, and the actual effectiveness of thermal error compensation is evaluated by comparing the machining errors.

[0012] The present invention discloses a dynamic thermal error compensation system for a dual-moving-column CNC gantry milling machine in a second aspect of the embodiments. The compensation system includes: The acquisition unit synchronously and in real time acquires the machine tool's temperature and displacement data through multiple temperature sensors distributed in the key heat-sensitive areas of the machine tool, as well as a laser rangefinder and encoder used to monitor the displacement of key components. The processing unit uses the Kalman filter algorithm to perform fusion and denoising processing on the displacement data to obtain preprocessed displacement data, and at the same time performs standardization processing on the temperature data to obtain preprocessed temperature data. The calculation unit inputs the preprocessed temperature data into the preset thermal error prediction model to calculate the real-time thermal error compensation amount. The control unit generates control commands based on the calculated thermal error compensation amount, driving the machine tool CNC system to adjust the relative position of the tool and the workpiece in real time.

[0013] Furthermore, it also includes an online model correction unit: The processing subunit uses the preprocessed displacement data as the observed value of the real-time thermal error truth value. The comparison sub-unit compares the compensation amount calculated by the thermal error prediction model with the displacement data observation value to generate a residual; The sub-unit is updated, and based on the residual, the parameters of the thermal error prediction model are dynamically updated using the recursive least squares method to achieve adaptive online correction of the model.

[0014] Furthermore, it also includes a compensation effect evaluation unit: The compensation effect evaluation unit is used to inspect the reference workpiece with a coordinate measuring machine after the compensation processing is completed, and to evaluate the actual effectiveness of thermal error compensation by comparing the processing errors.

[0015] The present invention discloses a dynamic thermal error compensation system for a dual-moving-column CNC gantry milling machine in a third aspect, the compensation system comprising: Memory containing executable program code; A processor coupled to the memory; The processor calls the executable program code stored in the memory to execute the dynamic thermal error compensation method for a dual-moving column CNC gantry milling machine as described above.

[0016] The present invention proposes a dynamic thermal error compensation method and system for a dual-moving column CNC gantry milling machine. The beneficial effects are as follows: The invention significantly improves the reliability and signal-to-noise ratio of thermal error monitoring data through multi-sensor data fusion and Kalman filtering denoising. Utilizing a thermal error prediction model, it accurately characterizes the nonlinear mapping relationship between the temperature field and complex machine tool deformation, achieving prediction and quantification of thermal errors. This fundamentally overcomes the problem of untimely compensation caused by temperature lag in traditional methods. Furthermore, by real-time correction of the tool path through the CNC system, online and dynamic software compensation is achieved, thereby significantly improving the machining accuracy and consistency of the workpiece and effectively solving the problem of high scrap rate caused by machine tool thermal deformation in high-precision machining. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating the dynamic thermal error compensation method for the dual-moving-column CNC gantry milling machine of the present invention. Figure 2 This is a schematic diagram of the training process for the thermal error prediction model of the present invention; Figure 3 This invention relates to a dynamic thermal error compensation system for a dual-moving-column CNC gantry milling machine. Figure 4 This invention provides another dynamic thermal error compensation system for a dual-moving column CNC gantry milling machine. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] It should be noted that the terms "first," "second," "third," "fourth," etc., used in the specification and claims of this invention are used to distinguish different objects, not to describe a specific order. The terms "comprising" and "having," and any variations thereof, in the embodiments of this invention are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to these processes, methods, products, or devices.

[0020] Example 1 Reference Figure 1 This invention discloses a flowchart illustrating a dynamic thermal error compensation method for a dual-moving-column CNC gantry milling machine. (See attached diagram.) Figure 1 As shown, the dynamic thermal error compensation method for the dual-moving column CNC gantry milling machine may include the following steps: 101. By using multiple temperature sensors distributed in the key heat-sensitive areas of the machine tool, as well as laser rangefinders and encoders for monitoring the displacement of key components, the temperature and displacement data of the machine tool are collected synchronously and in real time. 102. The displacement data is fused and denoised using the Kalman filter algorithm to obtain preprocessed displacement data. At the same time, the temperature data is standardized to obtain preprocessed temperature data. 103. Input the preprocessed temperature data into the preset thermal error prediction model and calculate the real-time thermal error compensation amount; 104. Based on the calculated thermal error compensation amount, control commands are generated to drive the CNC system of the machine tool to adjust the relative position of the tool and the workpiece in real time.

[0021] As an optional implementation, in this embodiment of the invention, in the data acquisition step, the temperature sensor is a K-type thermocouple, and the temperature sensor is sequentially arranged on the spindle box, the lead screw and nut pair, and the bed guide rail.

[0022] Specifically, by placing the sensor on the core internal heat source of the spindle box, the invention can directly capture the key temperature variables that cause thermal elongation and tilting of the spindle, providing a core basis for accurate prediction and compensation of thermal errors in the Z-axis direction. By placing points on the lead screw and nut pair, the frictional temperature rise can be monitored in real time, thereby converting the thermal expansion of the lead screw into the corresponding positioning compensation amount. In addition, by placing points on the bed guide rail, which is the basis of the machine tool's geometric accuracy, the overall temperature field information transmitted by various heat sources and causing the bed to undergo complex bending and twisting can be captured.

[0023] As an optional implementation, in this embodiment of the invention, a laser rangefinder, as a high-precision non-contact sensor, is mounted on a fixed part of the machine tool to directly measure the thermal expansion of moving parts, such as the spindle box, in the vertical direction; while the encoder is used to monitor the actual position of each feed axis, and the difference between its reading and the machine tool command position includes the error component caused by thermal deformation. As an optional implementation, in this embodiment of the invention, a Kalman filter algorithm is used to process the displacement data from the laser rangefinder and encoder. Kalman filtering is an optimal estimation algorithm that can effectively fuse data from different sensors and filter out random noise introduced by machine tool vibration, electrical noise, and other interference. Through an iterative process of state prediction and measurement update, it outputs a smooth, reliable, and optimally estimated displacement data sequence, greatly improving the signal-to-noise ratio and reliability of thermal error observations.

[0024] Temperature data standardization: The collected raw temperature data is simultaneously standardized. This process converts the temperature values ​​at each measurement point into a distribution with a mean of 0 and a standard deviation of 1. The formula is as follows: in, This is the original temperature value. It is the average of historical data. It is the standard deviation of historical data; standardization eliminates the influence of different dimensions and numerical ranges caused by differences in physical location and sensor characteristics of different temperature measurement points, so that all temperature features are on the same scale when input into the prediction model. This not only accelerates the convergence of the model, but also improves the prediction accuracy and stability of the model.

[0025] As an optional implementation, in this embodiment of the invention, the system collects temperature data in real time using temperature sensors pre-installed in key parts such as the spindle box and lead screw; at the same time, it collects displacement data in real time using a laser rangefinder and an encoder. The acquired raw displacement data is fused and denoised using a Kalman filter algorithm to obtain a smooth and reliable displacement signal; the temperature data is standardized to eliminate the influence of dimensions. The preprocessed standardized temperature data is input into the preset thermal error prediction model; Based on the calculated compensation amount, control commands are generated and the machine tool CNC system is driven to adjust the tool path in real time by means of position offset. By actively offsetting thermal deformation, the correct relative position between the tool and the workpiece is maintained, thereby directly improving the machining accuracy.

[0026] The method also includes an online model correction step: 105. Use the preprocessed displacement data as the observed value of the real-time thermal error truth value; As an optional implementation, in this embodiment of the invention, the "preprocessed displacement data" refers to the displacement signal acquired by a direct measurement device such as a laser rangefinder and processed by a Kalman filter algorithm. The system identifies this displacement data as the physical quantity closest to the actual thermal deformation at the current moment and defines it as the observed value of the real-time thermal error truth. This is based on the fact that the data directly originates from the physical displacement measurement of the machine tool structure, rather than from indirect calculations or predictions. By establishing this observed value, the system provides an objective, reliable, and quantifiable benchmark for subsequently evaluating the accuracy of the prediction model.

[0027] 106. Compare the compensation amount calculated by the thermal error prediction model with the displacement data observation value to generate residuals; As an optional implementation, in this embodiment of the invention, the theoretical thermal error compensation amount calculated in real time based on temperature data by the thermal error prediction model is compared in real time with the observed value established in step 106. This comparison is achieved through arithmetic operations, specifically calculating the difference between the theoretical compensation amount and the observed value. The system defines this difference as a residual, which is a signed numerical value whose absolute value characterizes the degree of deviation between the predicted output of the thermal error prediction model at the current moment and the measured physical deformation. Its positive or negative sign indicates the direction of deviation. Therefore, the residual serves as a direct input signal driving the update of model parameters.

[0028] If the calculation results of the thermal error prediction model are compared with the displacement observations in real time, the prediction residuals can be obtained: in, These are displacement observations. These are model predictions.

[0029] 107. Based on the residuals, the parameters of the thermal error prediction model are dynamically updated using the recursive least squares method to achieve adaptive online correction of the model.

[0030] As can be seen, through this process, the thermal error prediction model is no longer a static model fixed after leaving the factory, but an intelligent model that can dynamically adapt to performance changes, wear, and different processing conditions throughout the entire life cycle of the machine tool. This self-correcting capability greatly improves the long-term stability, robustness, and adaptability to complex and changing production environments of the compensation system, ensuring that the accuracy of thermal error compensation can be maintained at a high level for a long time.

[0031] 108. It also includes a compensation effect evaluation step: After the compensation processing is completed, a coordinate measuring machine is used to inspect the reference workpiece, and the actual effectiveness of thermal error compensation is evaluated by comparing the processing errors.

[0032] As an optional implementation, this embodiment of the invention also includes a compensation effect evaluation step: after the compensation machining is completed, a coordinate measuring machine is used to inspect the reference workpiece, and the actual effectiveness of thermal error compensation is evaluated by comparing the machining errors. The reference workpiece is a specially designed standard test piece containing multiple geometric features that reflect thermal errors in various axes of the machine tool, such as stepped holes, grid-like hole arrays, and precision inclined surface structures distributed along the X, Y, and Z axes. By measuring the deviation between the actual machined dimensions and the theoretical design dimensions of these features, the positioning accuracy improvement rate and accuracy stability index are calculated, and the actual effect of the compensation system is quantitatively evaluated. This evaluation method can objectively verify the effectiveness of the thermal error compensation system, provide data support for the optimization of compensation parameters, and establish a long-term performance monitoring mechanism to ensure that the compensation system maintains stable compensation accuracy throughout its entire life cycle.

[0033] As an optional implementation method, in this embodiment of the invention, by comparing the machining error data before and after compensation, the degree of improvement in the positioning accuracy of each axis by the thermal error compensation system can be accurately calculated. For example, by measuring the positional deviation of the precision hole system on the reference workpiece, the positioning error values ​​of the X, Y, and Z axes before and after compensation can be obtained, and then the accuracy improvement rate of each axis can be calculated. This quantitative evaluation method can not only verify the immediate effect of the compensation system, but also promptly detect the degradation trend of system performance by tracking the changes of these indicators over a long period of time, providing a basis for preventive maintenance. At the same time, based on the evaluation results, the parameters of the thermal error prediction model can be optimized in reverse, forming a complete closed loop from compensation to evaluation to optimization, continuously improving the compensation accuracy and stability of the system.

[0034] As can be seen, this invention significantly improves the reliability and signal-to-noise ratio of thermal error monitoring data through multi-sensor data fusion and Kalman filtering denoising. The thermal error prediction model can accurately characterize the nonlinear mapping relationship between the temperature field and complex machine tool deformation, realizing the prediction and quantification of thermal errors. It fundamentally overcomes the problem of untimely compensation caused by temperature lag in traditional methods. By correcting the tool path in real time through the CNC system, online and dynamic software compensation is realized, thereby greatly improving the machining accuracy and consistency of the workpiece and effectively solving the problem of scrap rate caused by machine tool thermal deformation in high-precision machining.

[0035] Example 2 Please see Figure 2 , Figure 2 This is a schematic diagram of the training process for the thermal error prediction model of the dynamic thermal error compensation method for a dual-moving column CNC gantry milling machine disclosed in an embodiment of the present invention; The thermal error prediction model described in this invention is a regression model based on machine learning, and its specific training steps are as follows: 201. Collect historical temperature data and corresponding thermal error data of the machine tool under different operating conditions, and construct a training sample set; 202. Standardize the temperature data in the training samples and use principal component analysis to reduce dimensionality in order to eliminate multicollinearity and reduce input variables; 203. Using the principal component scores after dimensionality reduction as input and the corresponding thermal error as output, a thermal error prediction model is trained using the support vector regression algorithm, and the model parameters are optimized using the cross-validation method.

[0036] The principal component analysis method includes: 204. Calculate the covariance matrix of the standardized temperature data and its eigenvalues ​​and eigenvectors; 205. Select the number of principal components based on the preset cumulative variance contribution rate not being lower than the threshold; 206. Based on the selected principal components and their eigenvectors, the original temperature data is converted into dimensionality-reduced principal component score vectors.

[0037] As an optional implementation, this embodiment of the invention collects temperature data and thermal error data of the machine tool under different operating conditions to construct a training sample set. Temperature data is obtained from sensors at various thermal key points, and thermal error data is obtained through a high-precision measurement device, ensuring that the sample covers typical operating conditions.

[0038] As an optional implementation, the temperature data is standardized in this embodiment of the invention: i = (x i - μ) / σ Where x i σ represents the original temperature value, μ represents the mean, and σ represents the standard deviation.

[0039] In addition, principal component analysis is used to eliminate multicollinearity among temperature variables, including: 1. Calculate the covariance matrix of the standardized data; 2. Perform eigenvalue decomposition to obtain eigenvalues ​​and eigenvectors; 3. Select the number of principal components k based on a cumulative variance contribution rate ≥ 85%; 4. Transform the data to the principal component space: T = XV k .

[0040] Finally, using principal component scores as input and thermal error as output, a support vector regression algorithm is employed to train the prediction model. Model parameters, including the penalty parameter C and kernel function parameters, are optimized through k-fold cross-validation to ensure good prediction accuracy and generalization ability.

[0041] As can be seen, this invention constructs a training sample set by collecting historical data under different working conditions, ensuring that the trained prediction model can adapt to various processing conditions, and significantly improving the model's generalization ability and reliability in practical applications. The use of principal component analysis to reduce the dimensionality of temperature data effectively eliminates the correlation interference between various temperature measurement points, retaining the main temperature change information while avoiding the instability problem caused by the high correlation of input variables. The use of support vector regression algorithm combined with cross-validation parameter optimization can effectively establish a complex nonlinear mapping relationship between temperature and thermal error, effectively avoiding overfitting and ensuring that the model has both excellent prediction accuracy and generalization stability.

[0042] Example 3 Please see Figure 3 , Figure 3 This invention discloses a dynamic thermal error compensation system for a dual-moving-column CNC gantry milling machine. For example... Figure 3 The compensation system may include: The acquisition unit 301 synchronously and in real time acquires the temperature and displacement data of the machine tool through multiple temperature sensors distributed in the key heat-sensitive areas of the machine tool, as well as a laser rangefinder and encoder for monitoring the displacement of key components. Processing unit 302 uses the Kalman filter algorithm to perform fusion and denoising processing on the displacement data to obtain preprocessed displacement data, and at the same time performs standardization processing on the temperature data to obtain preprocessed temperature data. The calculation unit 303 inputs the preprocessed temperature data into the preset thermal error prediction model to calculate the real-time thermal error compensation amount. The control unit 304 generates control commands based on the calculated thermal error compensation amount, driving the machine tool CNC system to adjust the relative position of the tool and the workpiece in real time.

[0043] As an optional implementation, in this embodiment of the invention, in the data acquisition step, the temperature sensor is a K-type thermocouple, and the temperature sensor is sequentially arranged on the spindle box, the lead screw and nut pair, and the bed guide rail.

[0044] Specifically, by placing the sensor on the core internal heat source of the spindle box, the invention can directly capture the key temperature variables that cause thermal elongation and tilting of the spindle, providing a core basis for accurate prediction and compensation of thermal errors in the Z-axis direction. By placing points on the lead screw and nut pair, the frictional temperature rise can be monitored in real time, thereby converting the thermal expansion of the lead screw into the corresponding positioning compensation amount. In addition, by placing points on the bed guide rail, which is the basis of the machine tool's geometric accuracy, the overall temperature field information transmitted by various heat sources and causing the bed to undergo complex bending and twisting can be captured.

[0045] As an optional implementation, in this embodiment of the invention, a laser rangefinder, as a high-precision non-contact sensor, is mounted on a fixed part of the machine tool to directly measure the thermal expansion of moving parts, such as the spindle box, in the vertical direction; while the encoder is used to monitor the actual position of each feed axis, and the difference between its reading and the machine tool command position includes the error component caused by thermal deformation. As an optional implementation, in this embodiment of the invention, a Kalman filter algorithm is used to process the displacement data from the laser rangefinder and encoder. Kalman filtering is an optimal estimation algorithm that can effectively fuse data from different sensors and filter out random noise introduced by machine tool vibration, electrical noise, and other interference. Through an iterative process of state prediction and measurement update, it outputs a smooth, reliable, and optimally estimated displacement data sequence, greatly improving the signal-to-noise ratio and reliability of thermal error observations.

[0046] Temperature data standardization: The collected raw temperature data is simultaneously standardized. This process converts the temperature values ​​at each measurement point into a distribution with a mean of 0 and a standard deviation of 1. The formula is as follows: in, This is the original temperature value. It is the average of historical data. It is the standard deviation of historical data; standardization eliminates the influence of different dimensions and numerical ranges caused by differences in physical location and sensor characteristics of different temperature measurement points, so that all temperature features are on the same scale when input into the prediction model. This not only accelerates the convergence of the model, but also improves the prediction accuracy and stability of the model.

[0047] As an optional implementation, in this embodiment of the invention, the system collects temperature data in real time using temperature sensors pre-installed in key parts such as the spindle box and lead screw; at the same time, it collects displacement data in real time using a laser rangefinder and an encoder. The acquired raw displacement data is fused and denoised using a Kalman filter algorithm to obtain a smooth and reliable displacement signal; the temperature data is standardized to eliminate the influence of dimensions. The preprocessed standardized temperature data is input into the preset thermal error prediction model; Based on the calculated compensation amount, control commands are generated and the machine tool CNC system is driven to adjust the tool path in real time by means of position offset. By actively offsetting thermal deformation, the correct relative position between the tool and the workpiece is maintained, thereby directly improving the machining accuracy.

[0048] As an optional implementation, this embodiment of the invention further includes an online model correction unit 305: Processing subunit 3051 uses the preprocessed displacement data as the observed value of the real-time thermal error truth value; As an optional implementation, in this embodiment of the invention, the "preprocessed displacement data" refers to the displacement signal acquired by a direct measurement device such as a laser rangefinder and processed by a Kalman filter algorithm. The system identifies this displacement data as the physical quantity closest to the actual thermal deformation at the current moment and defines it as the observed value of the real-time thermal error truth. This is based on the fact that the data directly originates from the physical displacement measurement of the machine tool structure, rather than from indirect calculations or predictions. By establishing this observed value, the system provides an objective, reliable, and quantifiable benchmark for subsequently evaluating the accuracy of the prediction model.

[0049] Comparison subunit 3052 compares the compensation amount calculated by the thermal error prediction model with the displacement data observation value to generate residuals; As an optional implementation, in this embodiment of the invention, the theoretical thermal error compensation amount calculated in real time based on temperature data by the thermal error prediction model is compared in real time with the observed value established in step 106. This comparison is achieved through arithmetic operations, specifically calculating the difference between the theoretical compensation amount and the observed value. The system defines this difference as a residual, which is a signed numerical value whose absolute value characterizes the degree of deviation between the predicted output of the thermal error prediction model at the current moment and the measured physical deformation. Its positive or negative sign indicates the direction of deviation. Therefore, the residual serves as a direct input signal driving the update of model parameters.

[0050] If the calculation results of the thermal error prediction model are compared with the displacement observations in real time, the prediction residuals can be obtained: in, These are displacement observations. These are model predictions.

[0051] The updated subunit 3053 dynamically updates the parameters of the thermal error prediction model based on the residual using the recursive least squares method, thereby achieving adaptive online correction of the model.

[0052] As can be seen, through this process, the thermal error prediction model is no longer a static model fixed after leaving the factory, but an intelligent model that can dynamically adapt to performance changes, wear, and different processing conditions throughout the entire life cycle of the machine tool. This self-correcting capability greatly improves the long-term stability, robustness, and adaptability to complex and changing production environments of the compensation system, ensuring that the accuracy of thermal error compensation can be maintained at a high level for a long time.

[0053] As an optional implementation, this embodiment of the invention further includes a compensation effect evaluation unit 306: The compensation effect evaluation unit is used to inspect the reference workpiece with a coordinate measuring machine after the compensation processing is completed, and to evaluate the actual effectiveness of thermal error compensation by comparing the processing errors.

[0054] As an optional implementation, this embodiment of the invention also includes a compensation effect evaluation step: after the compensation machining is completed, a coordinate measuring machine is used to inspect the reference workpiece, and the actual effectiveness of thermal error compensation is evaluated by comparing the machining errors. The reference workpiece is a specially designed standard test piece containing multiple geometric features that reflect thermal errors in various axes of the machine tool, such as stepped holes, grid-like hole arrays, and precision inclined surface structures distributed along the X, Y, and Z axes. By measuring the deviation between the actual machined dimensions and the theoretical design dimensions of these features, the positioning accuracy improvement rate and accuracy stability index are calculated, and the actual effect of the compensation system is quantitatively evaluated. This evaluation method can objectively verify the effectiveness of the thermal error compensation system, provide data support for the optimization of compensation parameters, and establish a long-term performance monitoring mechanism to ensure that the compensation system maintains stable compensation accuracy throughout its entire life cycle.

[0055] As an optional implementation method, in this embodiment of the invention, by comparing the machining error data before and after compensation, the degree of improvement in the positioning accuracy of each axis by the thermal error compensation system can be accurately calculated. For example, by measuring the positional deviation of the precision hole system on the reference workpiece, the positioning error values ​​of the X, Y, and Z axes before and after compensation can be obtained, and then the accuracy improvement rate of each axis can be calculated. This quantitative evaluation method can not only verify the immediate effect of the compensation system, but also promptly detect the degradation trend of system performance by tracking the changes of these indicators over a long period of time, providing a basis for preventive maintenance. At the same time, based on the evaluation results, the parameters of the thermal error prediction model can be optimized in reverse, forming a complete closed loop from compensation to evaluation to optimization, continuously improving the compensation accuracy and stability of the system.

[0056] As can be seen, this invention significantly improves the reliability and signal-to-noise ratio of thermal error monitoring data through multi-sensor data fusion and Kalman filtering denoising. The thermal error prediction model can accurately characterize the nonlinear mapping relationship between the temperature field and complex machine tool deformation, realizing the prediction and quantification of thermal errors. It fundamentally overcomes the problem of untimely compensation caused by temperature lag in traditional methods. By correcting the tool path in real time through the CNC system, online and dynamic software compensation is realized, thereby greatly improving the machining accuracy and consistency of the workpiece and effectively solving the problem of scrap rate caused by machine tool thermal deformation in high-precision machining.

[0057] Example 4 Please see Figure 4 , Figure 4 This is another dynamic thermal error compensation system for a dual-moving-column CNC gantry milling machine disclosed in an embodiment of the present invention. For example... Figure 4 The compensation system may include Memory 401 storing executable program code; Processor 402 coupled to the memory 401; The processor 402 calls the executable program code stored in the memory 401 to execute the dynamic thermal error compensation method for the dual-moving column CNC gantry milling machine as described above.

[0058] This invention discloses a computer-readable storage medium storing a computer program, wherein the computer program causes a computer to execute... Figures 1-2 A dynamic thermal error compensation method for any type of double-moving column CNC gantry milling machine.

[0059] This invention also discloses a computer program product, wherein when the computer program product is run on a computer, the computer performs some or all of the steps of the methods described in the above method embodiments.

[0060] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, including read-only memory (ROM), random access memory (RAM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), one-time programmable read-only memory (OTPROM), electrically-Erasable Programmable Read-Only Memory (EEPROM), compactdisc read-only memory (CD-ROM) or other optical disc storage, disk storage, magnetic tape storage, or any other computer-readable medium capable of carrying or storing data.

[0061] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A dynamic thermal error compensation method for a dual moving column numerical control gantry milling machine, characterized in that, The method comprises the following steps: Real-time temperature data and displacement data of the machine tool are synchronously collected through a plurality of temperature sensors distributed in key heat-sensitive areas of the machine tool, and a laser range finder and an encoder for monitoring displacement of key components; The displacement data is fused and denoised by using a Kalman filtering algorithm to obtain preprocessed displacement data, and the temperature data is standardized to obtain preprocessed temperature data; The preprocessed temperature data is input into a preset thermal error prediction model to calculate a real-time thermal error compensation amount; A control instruction is generated according to the calculated thermal error compensation amount to drive the machine tool numerical control system to adjust the relative position of a tool and a workpiece in real time.

2. The dynamic thermal error compensation method of a dual moving column numerical control gantry milling machine according to claim 1, characterized in that: In the data collection step, the temperature sensors are K-type thermocouples, and the temperature sensors are arranged on a spindle box, a screw nut pair and a bed rail in sequence.

3. The dynamic thermal error compensation method of a dual moving column numerical control gantry milling machine according to claim 1, wherein: The thermal error prediction model is a regression model based on machine learning, and the specific training steps are as follows: Collect historical temperature data and corresponding thermal error data of the machine tool under different working conditions to construct a training sample set; The temperature data in the training sample is standardized, and principal component analysis is used for dimension reduction to eliminate multicollinearity and reduce input variables; The principal component scores after dimension reduction are used as input, and the corresponding thermal error is used as output, and a support vector regression algorithm is used to train the thermal error prediction model, and the model parameters are optimized by a cross-validation method.

4. The dynamic thermal error compensation method of a dual moving column numerical control gantry milling machine according to claim 3, characterized in that: The principal component analysis method comprises: The covariance matrix of the standardized temperature data and its eigenvalues and eigenvectors are calculated; The number of principal components is selected according to a preset cumulative variance contribution rate not less than a threshold value; Based on the selected principal components and their eigenvectors, the original temperature data is converted into a principal component score vector after dimension reduction.

5. The dynamic thermal error compensation method of a dual moving column numerical control gantry milling machine according to claim 1, wherein: The method further comprises a model online correction step: The preprocessed displacement data are used as observation values of real-time thermal error true values; The compensation amount calculated by the thermal error prediction model is compared with the displacement data observation values to generate a residual error; Based on the residual error, the parameters of the thermal error prediction model are dynamically updated by using a recursive least squares method to realize adaptive online correction of the model.

6. The dynamic thermal error compensation method of a dual moving column numerical control gantry milling machine according to claim 1, wherein: The method further comprises a compensation effect evaluation step: after completing compensation processing, a three-coordinate measuring machine is used to detect a reference workpiece, and the actual performance of thermal error compensation is evaluated by comparing processing errors.

7. A dynamic thermal error compensation system for a dual moving column CNC gantry mill, characterized by, The method comprises the following steps: A collection unit synchronously collects real-time temperature data and displacement data of the machine tool through a plurality of temperature sensors distributed in key heat-sensitive areas of the machine tool, and a laser range finder and an encoder for monitoring displacement of key components; A processing unit fuses and denoises the displacement data by using a Kalman filtering algorithm to obtain preprocessed displacement data, and standardizes the temperature data to obtain preprocessed temperature data; A calculation unit inputs the preprocessed temperature data into a preset thermal error prediction model to calculate a real-time thermal error compensation amount; A control unit generates a control instruction according to the calculated thermal error compensation amount to drive the machine tool numerical control system to adjust the relative position of a tool and a workpiece in real time.

8. The dynamic thermal error compensation system of a twin-kinematic- column numerically controlled gantry milling machine according to claim 7, wherein: The method further comprises a model online correction unit: a processing subunit, taking the pre-processed displacement data as an observation value of real-time thermal error true value; a comparison subunit, comparing the compensation amount calculated by the thermal error prediction model with the displacement data observation value to generate a residual; an updating subunit, dynamically updating the parameters of the thermal error prediction model based on the residual by using a recursive least square method to realize adaptive online correction of the model.

9. The dynamic thermal error compensation system of a twin-kinematic-structure CNC gantry milling machine according to claim 7, wherein: Further comprising a compensation effect evaluation unit: The compensation effect evaluation unit is used to detect the reference workpiece after completing the compensation processing using a three-coordinate measuring machine, and evaluate the actual performance of the thermal error compensation by comparing the processing errors.

10. A dynamic thermal error compensation system for a dual moving column CNC gantry mill, characterized by, including; a memory storing executable program codes; a processor coupled with the memory; The processor calls the executable program codes stored in the memory to execute the dynamic thermal error compensation method of the double-moving-column numerical control gantry milling machine according to any one of claims 1-6.

Images