A numerical control machine tool thermal error online monitoring method and system

Abstract

The application provides a numerical control machine tool thermal error online monitoring method and system, relates to the technical field of numerical control machine tool thermal error online monitoring, and collects temperature original data and infrared thermal imaging data of a guide rail mounting surface, a workbench supporting surface and a side surface of a machine bed; each area temperature field evaluation data is obtained by processing the temperature original data, visual image data is obtained by processing the infrared thermal imaging data; then, the consistency of the two is verified to obtain a collaborative data set, and a temperature field clustering result is determined by clustering; then, temperature and area data are extracted to determine a main influence area; finally, the data of the area is calculated by using a thermal error mapping algorithm to obtain a thermal error value, and the thermal error value can be obtained by multi-position temperature and infrared data collection, processing, verification, clustering and feature analysis, combined with the thermal error mapping algorithm, so that the online monitoring of the numerical control machine tool thermal error is realized.

Description

Technical Field

[0001] This application relates to the technical field of online monitoring of thermal errors in CNC machine tools, and in particular to a method and system for online monitoring of thermal errors in CNC machine tools. Background Technology

[0002] In high-precision manufacturing fields such as aerospace component processing and precision mold manufacturing, CNC machine tools, as core processing equipment, directly determine product quality through their machining accuracy. During CNC machine tool operation, critical components such as guide rail mounting surfaces, worktable support surfaces, and bed sides are susceptible to thermal deformation and thermal errors caused by factors such as motor heating, cutting heat, and ambient temperature fluctuations. This leads to deviations in the dimensional and positional accuracy of machined parts from design requirements. To avoid the impact of thermal errors on machining quality, online monitoring of CNC machine tool thermal errors is necessary. Real-time acquisition of thermal error data is required to assist in subsequent accuracy compensation or equipment adjustments. Therefore, developing reliable and accurate online monitoring methods for CNC machine tool thermal errors has become a key requirement in the high-precision manufacturing field.

[0003] Currently, existing technologies for online monitoring of thermal errors in CNC machine tools typically employ two methods: one involves deploying multiple temperature sensors on key components such as the CNC machine tool's guideways, worktable, and bed to collect raw temperature data in real time. The collected temperature data is then substituted into a pre-defined empirical error model to calculate the thermal error value. The other method utilizes an infrared thermal imager to scan the entire or a localized area of ​​the CNC machine tool, acquiring infrared thermal imaging data of the machine tool surface. By reading the temperature information of specific pixels in the infrared thermal imaging data and combining this with the corresponding machine tool location, the thermal error of that location can be estimated. Alternatively, the direction of thermal error change can be roughly determined based on the temperature distribution trend presented by the infrared thermal imaging data.

[0004] However, existing technologies have significant limitations in practical applications. Their most significant drawback lies in their inability to simultaneously ensure the comprehensiveness and reliability of temperature data, leading to insufficient accuracy in thermal error monitoring. Specifically, relying solely on temperature sensors to collect single-point temperature data fails to reflect the distribution characteristics and variation patterns of the temperature field across the entire CNC machine tool. It can only estimate thermal errors based on local temperature information, easily overlooking the comprehensive impact of temperature interactions between different parts on thermal deformation. Conversely, using infrared thermal imaging data alone lacks consistency verification between infrared data and temperature statistics, cannot eliminate deviations caused by environmental interference, and struggles to accurately locate key areas that dominate thermal deformation. Ultimately, this results in significant discrepancies between the calculated thermal error and the actual thermal deformation, failing to meet the accuracy requirements of high-precision manufacturing scenarios for thermal error monitoring. Summary of the Invention

[0005] The purpose of this application is to provide a method and system for online monitoring of thermal errors in CNC machine tools, so as to solve the problem of low accuracy of thermal error monitoring in the prior art.

[0006] To address the aforementioned technical problems, in a first aspect, this application provides a method for online monitoring of thermal errors in CNC machine tools, comprising:

[0007] The system collects raw temperature data of the CNC machine tool's guide rail mounting surface, worktable support surface, and bed side, as well as infrared thermal imaging data of the CNC machine tool online.

[0008] The raw temperature data is processed by image segmentation to obtain temperature segmentation data for each region. The average temperature and temperature variance of the temperature segmentation data are calculated to obtain temperature field evaluation data for each region. The infrared thermal imaging data is processed by pixel fusion to obtain visualized image data.

[0009] The temperature field assessment data and the visualization image data are checked for consistency to obtain a collaborative dataset. Based on the collaborative dataset, a region clustering algorithm is used to determine the temperature field clustering results of each clustering unit within the region.

[0010] Feature extraction is performed on the temperature field clustering results to obtain temperature data and area data of each cluster unit. Based on the temperature data and area data, the main influencing area of ​​thermal deformation of CNC machine tools is obtained.

[0011] A thermal error mapping algorithm is used to perform correlation calculation on the temperature data and area data corresponding to the main affected area to obtain the thermal error value.

[0012] Optionally, the step of using a thermal error mapping algorithm to correlate the temperature data and area data corresponding to the main affected area to obtain a thermal error value includes:

[0013] The scope of the parts of the CNC machine tool bed that need to be monitored is defined to identify the key locations that require focused monitoring.

[0014] Based on the spatial relationship between the key locations and the main areas of influence, an association is established between the key locations and the temperature data and area data of each cluster unit in the corresponding main areas of influence, resulting in a table of correspondence between locations and data.

[0015] Based on the correspondence table, a thermal error mapping algorithm is used to calculate the preliminary error value for each key position, and the preliminary error value is corrected to obtain the target error value.

[0016] The thermal error value is obtained by summing up the target error values ​​at all key locations.

[0017] Optionally, the step of calculating the preliminary error value for each key location based on the correspondence table using a thermal error mapping algorithm includes:

[0018] Obtain the temperature data, area data, and spatial characteristic parameters of each key location from the correspondence table;

[0019] The temperature deviation value is obtained by calculating the difference between the corresponding temperature data and the preset reference temperature. The corresponding area data is then converted according to the preset area weighting rule to obtain the area weighting value.

[0020] A thermal error mapping algorithm is used to multiply the temperature deviation value, the area weight value, and the thermal deformation sensitivity coefficient in the spatial characteristic parameters corresponding to the same key location to obtain a preliminary error base.

[0021] According to the preset error unit conversion rules, the preliminary error base corresponding to each key position is converted to obtain the preliminary error value of each key position.

[0022] Optionally, correcting the initial error value to obtain the target error value includes:

[0023] The initial error value is compared with a preset reasonable range of values ​​to obtain a first comparison result. The difference between the initial error value and the adjacent initial error value is calculated. The difference is compared with the preset value change range requirement to obtain a second comparison result.

[0024] The first comparison result and the second comparison result are combined to obtain the error check result corresponding to each key position. The error check results are classified according to the preset classification rules to obtain qualified results and results to be adjusted.

[0025] The preliminary error value corresponding to the qualified result is determined as the error value to be confirmed. The preliminary error value corresponding to the result to be adjusted is adjusted using a preset correction rule to obtain the error value to be confirmed.

[0026] The uniqueness of all the error values ​​to be confirmed is verified to obtain the target error value corresponding to each of the key positions.

[0027] Optionally, the step of verifying the consistency between the temperature field assessment data and the visualized image data to obtain a collaborative dataset, and using a region clustering algorithm based on the collaborative dataset to determine the temperature field clustering results of each clustering unit within the region, includes:

[0028] Temperature information is extracted from the visualized image data to obtain the image temperature information for each region;

[0029] The image temperature information is compared with the average temperature and temperature variance of the corresponding region in the temperature field evaluation data to obtain the verification result.

[0030] The verification results are judged to filter out valid area data, and all the valid area data are summarized to form a collaborative dataset;

[0031] A region clustering algorithm is used to divide the regions in the collaborative dataset whose temperature values ​​meet the preset temperature conditions and whose pixel positions are connected into one unit, resulting in multiple clustering units.

[0032] By summarizing all the clustering units and their corresponding temperature values ​​and pixel location information, the temperature field clustering result is obtained.

[0033] Optionally, feature extraction is performed on the temperature field clustering results to obtain temperature data and area data for each cluster unit. Based on the temperature data and the area data, the main influencing areas of thermal deformation of the CNC machine tool are obtained, including:

[0034] Temperature features are extracted from the temperature field clustering results to obtain temperature data for each cluster unit. Regional range statistics are performed on each cluster unit in the temperature field clustering results to obtain area data for each cluster unit.

[0035] The temperature data of all cluster units are sorted according to their numerical values ​​to obtain a temperature sequence. The cluster units corresponding to the temperatures at the top of the temperature sequence are selected as the first candidate units according to a preset temperature ratio.

[0036] According to the numerical value, the area data corresponding to all the first candidate units are sorted to obtain an area sequence. The clustering units corresponding to the areas at the top of the area sequence are selected as the second candidate units according to a preset area ratio.

[0037] Priority evaluation is performed on the second candidate units to obtain the priority level of each second candidate unit. The priority levels of all second candidate units are sorted. Based on the sorting results, the spatial range corresponding to the second candidate unit with the highest ranking is selected as the main influence area of ​​thermal deformation of the CNC machine tool.

[0038] Optionally, the pixel fusion processing of the infrared thermal imaging data to obtain visualized image data includes:

[0039] Pixel information is extracted from infrared thermal imaging data to obtain the temperature value and position coordinates of each pixel;

[0040] According to the preset temperature and display feature correspondence rules, the temperature value is converted to obtain the pixel display information of each pixel;

[0041] The position coordinates are fused with the corresponding pixel display information to obtain pixel fusion information;

[0042] According to the distribution order of the position coordinates on the CNC machine tool bed, all the pixel fusion information is arranged to obtain pixel arrangement information;

[0043] Integrate all the pixel display information in the pixel arrangement information to form visualized image data.

[0044] Secondly, this application provides an online thermal error monitoring system for CNC machine tools, comprising:

[0045] The data acquisition module is used to acquire raw temperature data of the CNC machine tool's guide rail mounting surface, worktable support surface, and bed side surface, as well as infrared thermal imaging data of the CNC machine tool online.

[0046] The fusion module is used to perform image segmentation processing on the raw temperature data to obtain temperature segmentation data for each region, calculate the average temperature and temperature variance of the temperature segmentation data to obtain temperature field evaluation data for each region, and perform pixel fusion processing on the infrared thermal imaging data to obtain visualized image data.

[0047] The verification module is used to verify the consistency between the temperature field assessment data and the visualization image data to obtain a collaborative dataset. Based on the collaborative dataset, a region clustering algorithm is used to determine the temperature field clustering results of each clustering unit within the region.

[0048] The extraction module is used to extract features from the temperature field clustering results to obtain temperature data and area data of each cluster unit, and based on the temperature data and area data, to obtain the main influencing area of ​​thermal deformation of CNC machine tools;

[0049] The calculation module is used to perform correlation calculations on the temperature data and area data corresponding to the main affected area using a thermal error mapping algorithm to obtain a thermal error value.

[0050] Thirdly, this application provides an electronic device, comprising:

[0051] Memory, used to store computer programs;

[0052] A processor is configured to execute the computer program to implement the steps of an online thermal error monitoring method for CNC machine tools as described in the first aspect above.

[0053] Fourthly, this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, can implement the steps of the online monitoring method for thermal errors of a CNC machine tool as described in the first aspect above.

[0054] This application provides an online monitoring method for thermal errors in CNC machine tools. The method involves online acquisition of raw temperature data from the CNC machine tool's guide rail mounting surface, worktable support surface, and bed side, as well as infrared thermal imaging data. The raw temperature data is processed through image segmentation to obtain temperature segmentation data for each region. The average temperature and temperature variance of the segmented temperature data are calculated to obtain temperature field assessment data for each region. The infrared thermal imaging data is then processed through pixel fusion to obtain visualized image data. The temperature field assessment data and the visualized image data are checked for consistency to obtain a collaborative dataset. Based on the collaborative dataset, a region clustering algorithm is used to determine the temperature field clustering results for each cluster unit within the region. Feature extraction is performed on the temperature field clustering results to obtain temperature data and area data for each cluster unit. Based on the temperature data and area data, the main influencing regions of the CNC machine tool's thermal deformation are identified. Finally, a thermal error mapping algorithm is used to perform correlation calculations on the temperature data and area data corresponding to the main influencing regions to obtain thermal error values.

[0055] The technical solution of this application has the following beneficial effects:

[0056] This application covers critical areas of CNC machine tools sensitive to thermal deformation, simultaneously acquiring both "single-point temperature data" and "global temperature distribution data." This avoids the limitations of single-data types, providing comprehensive and complete raw data support for subsequent multi-dimensional data collaborative processing, ensuring no omissions in the monitoring range. By transforming scattered raw temperature data into temperature field assessment data reflecting the overall characteristics of the regional temperature, and simultaneously converting infrared thermal imaging data into intuitive visual image data, it achieves the transformation from "abstract data" to "usable data." This facilitates subsequent data verification and provides the possibility for manual, intuitive observation of temperature distribution. Consistency checks eliminate abnormal deviations in single data points, ensuring data reliability. Furthermore, region clustering groups data with similar temperatures and locations... The interconnected regions are divided into clustering units to achieve a structured division of the machine tool temperature field, avoiding the chaotic nature of the overall data and laying the foundation for subsequent feature extraction. Key features are extracted from the clustering results, and the regions that play a dominant role in thermal deformation are accurately located through dual-dimensional filtering based on "temperature level" and "region size". This avoids non-critical region data interfering with subsequent error calculations and achieves the monitoring goal of "focusing on core influencing factors". Error calculations are performed on the core data of the main influencing regions, rather than the overall data, which reduces computational redundancy and ensures that the algorithm focuses on the factors that contribute the most to thermal deformation. The accuracy of thermal error calculation is improved through the correlation operation of "temperature and area", making the results more consistent with the actual thermal deformation of the machine tool.

[0057] Furthermore, this application defines the scope of the parts of the CNC machine tool bed that need to be monitored to determine the key locations that need to be monitored. Then, based on the spatial relationship between the key locations and the main affected areas, it establishes the correlation between the temperature data and area data of each cluster unit in the corresponding main affected areas, and obtains a correspondence table between locations and data. Next, based on this correspondence table, it uses a thermal error mapping algorithm to calculate the preliminary error value of each key location, and corrects the preliminary error value to obtain the target error value. Finally, it summarizes the target error values ​​of all key locations to obtain the thermal error value.

[0058] This application focuses on key monitoring points by "delineating critical locations" to avoid dispersing computing resources for data from non-critical locations; it uses a "correspondence table between location and data" to accurately bind critical locations with core data, preventing error calculation deviations caused by misalignment of data and location; it eliminates unreliable values ​​caused by transient interference through "preliminary error correction" to ensure the accuracy of error values ​​at each critical location; and finally, it summarizes the target errors at critical locations to obtain thermal error values, thus achieving accurate monitoring of thermal errors in key parts of the machine tool.

[0059] These or other aspects of this application will become more apparent in the following description of the embodiments. Attached Figure Description

[0060] To more clearly illustrate the technical solutions of the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0061] Figure 1 A flowchart illustrating an online thermal error monitoring method for CNC machine tools provided in this application embodiment;

[0062] Figure 2 This is a schematic diagram illustrating a specific implementation of an online thermal error monitoring method for CNC machine tools provided in this application.

[0063] Figure 3 A schematic diagram of the structure of an online thermal error monitoring system for CNC machine tools provided in this application embodiment;

[0064] Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0065] In high-precision manufacturing fields such as aerospace components and precision molds, the thermal error of CNC machine tools directly affects the machining accuracy of parts. Therefore, online monitoring of thermal errors is crucial to ensuring product quality. However, existing online thermal error monitoring technologies for CNC machine tools have significant limitations: some technologies only collect single-point temperature data from parts such as guideways and worktables using temperature sensors, reflecting only local temperature conditions and failing to present the distribution and variation patterns of the temperature field across the entire machine tool. This can easily lead to inaccurate thermal error estimations due to neglecting overall temperature interactions. Other technologies rely solely on infrared thermal imaging data, which can acquire the overall temperature distribution. However, infrared data is easily affected by environmental factors such as dust obstruction and light reflection, and lacks verification with temperature statistical data, making it difficult to determine data reliability. Furthermore, it cannot accurately locate key areas that play a dominant role in thermal deformation, ultimately resulting in thermal error monitoring results that do not match the actual situation, failing to meet the demands of high-precision manufacturing.

[0066] To address the problem of insufficient accuracy in thermal error monitoring due to the inability of existing technologies to simultaneously ensure the comprehensiveness and reliability of temperature data, this application proposes an online thermal error monitoring method for CNC machine tools. This method employs the following technical concept: First, it collects raw temperature data at single points on the CNC machine tool's guide rail mounting surface, worktable support surface, and bed side, while simultaneously collecting overall infrared thermal imaging data of the machine tool, achieving data complementarity between "precise local temperature measurement and comprehensive temperature visualization." Next, the raw temperature data is processed to obtain the average temperature and temperature fluctuation of each region. The infrared data is processed and transformed into intuitive images, and reliability is ensured by checking the consistency of the two types of data. Then, the verified valid data is divided into regions based on similar temperatures and adjacent locations, accurately locating the main areas with the greatest impact on thermal deformation. Finally, for these key areas, the thermal error value is calculated by combining their temperature and area data. This solution solves the data reliability problem through "complementary data acquisition and consistency verification," and solves the comprehensive monitoring problem through "comprehensive data processing and key area location," effectively overcoming the shortcomings of existing technologies and providing accurate and reliable thermal error monitoring results for high-precision manufacturing.

[0067] To enable those skilled in the art to better understand the present application, the present application will be further described in detail below with reference to the accompanying drawings and specific embodiments. Obviously, the described embodiments are merely some embodiments of the present application, and not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0068] The core of this application is to provide an online monitoring method for thermal errors in CNC machine tools, and a flowchart of one specific implementation is shown below. Figure 1 As shown, the method includes:

[0069] S101: Online acquisition of raw temperature data of CNC machine tool guide rail mounting surface, worktable support surface and bed side, as well as infrared thermal imaging data of CNC machine tool.

[0070] In the above scheme, the raw temperature data refers to the real-time temperature information of three key heat-deformation sensitive parts of the CNC machine tool: the guide rail mounting surface, the worktable support surface, and the bed side surface. This data is used for subsequent analysis of the temperature change patterns and temperature field distribution characteristics of each area. The infrared thermal imaging data refers to the temperature distribution image data of the entire CNC machine tool or key areas acquired by an infrared thermal imager. This data can intuitively present the temperature differences of different parts of the machine tool and is used for subsequent visualization of the temperature distribution and collaborative verification with the raw temperature data. The guide rail mounting surface refers to the contact surface connecting the CNC machine tool guide rail and the bed. The worktable support surface refers to the contact surface that supports the worktable and ensures its stable movement. The bed side surface refers to the side area of ​​the bed near the machine tool's power components.

[0071] In this application example, firstly, raw temperature data is acquired: contact temperature sensors are selected, and based on the structural characteristics of the CNC machine tool's guide rail mounting surface, worktable support surface, and bed side, sensors are deployed at heat deformation sensitive points in various locations. Bolt fixing or high-temperature resistant adhesive bonding is used to ensure tight contact between the sensors and the monitored areas. All sensors are connected to the data acquisition module via signal cables, and the data acquisition frequency is set. The data acquisition module receives and stores the temperature signals acquired by each sensor in real time, converting them into raw temperature data suitable for subsequent processing. Next, after acquiring the raw temperature data, infrared thermal imaging data is acquired: a non-contact infrared thermal imager is selected and positioned around the machine tool in an unobstructed location that covers the guide rail mounting surface, worktable support surface, and bed side. The infrared thermal imager is fixed using a tripod, and the lens angle and focal length are adjusted to ensure complete coverage of the three key monitoring areas. The image resolution and frame rate of the thermal imager are set, and the thermal imager is connected to the data processing terminal via a data cable. The thermal imager acquires infrared thermal images of the monitored areas in real time, transmits the image data to the terminal, and stores it, forming infrared thermal imaging data.

[0072] In practical applications, during the production of precision parts on CNC machine tools, the initial step is to collect raw temperature data. Two platinum resistance temperature sensors are placed along the slider's motion path on the X-axis guideway mounting surface and at the center of the Y-axis guideway mounting surface. One sensor is placed at each of the four corners of the worktable support surface, and two sensors are placed on the side of the bed near the spindle box and the X-axis motor, for a total of 10 sensors. M3 bolts are used to secure the sensors to their monitoring locations. The sensors are connected to a data acquisition module via shielded cables. The module communicates with the machine tool's industrial computer, and the data acquisition frequency is set to 1 second per acquisition. The machining time for this part is 2 hours, which translates to 7200 seconds. Based on the total data volume calculation formula... Where Q is the total amount of raw temperature data, N is the number of sensors, f is the acquisition frequency, and T is the acquisition duration, substituting the values, we get Q = 10 × 1 × 7200 = 72000 data points. The industrial computer receives and stores these 72000 temperature data points in real time. These data will be used for image segmentation processing of the raw temperature data in step S102.

[0073] Next, infrared thermal imaging data acquisition was conducted: The infrared thermal imager was fixed on a tripod 1.8 meters above the machine tool's operating side. The imager's lens angle was adjusted to cover all the aforementioned temperature monitoring areas. The imager resolution was set to 640×512 pixels and the frame rate to 10 frames per second. The total number of infrared thermal imaging frames was calculated using the formula... Where F is the total number of frames. Let F be the frame rate and T be the acquisition duration. Substituting the values, we get F = 10 × 7200 = 72000 frames; The formula for calculating the number of pixels per frame is... Where P is the number of pixels per frame, W is the number of pixels in the image width, and H is the number of pixels in the image height, substituting the values, we get P = 640 × 512 = 327680 pixels; Formula for calculating the total pixel data volume of infrared thermal imaging. Substituting the values, we get D = 72000 × 327680 = 23592960000 pixels. The thermal imager is then connected to an industrial computer via a USB cable. The industrial computer receives and stores these 72,000 frames of infrared thermal imaging image data in real time. These data will be used for pixel fusion processing of the infrared thermal imaging data in step S102.

[0074] The aforementioned S101 overall solution, by simultaneously collecting raw temperature data from three thermally sensitive key parts of the CNC machine tool and infrared thermal imaging data covering the key areas, achieves both accurate and real-time monitoring of the temperature of local core parts and a visual representation of the overall temperature distribution of the machine tool through infrared thermal imaging data. These two types of data complement each other, providing a comprehensive and reliable raw data foundation for subsequent temperature data processing, verification, and thermal error analysis. Furthermore, the precise placement and parameter settings for key parts ensure that the collected data accurately reflects the temperature changes during CNC machine tool operation, avoiding deviations in subsequent analysis due to missed monitoring points or incomplete data collection. This lays a data foundation for the accuracy of the entire online thermal error monitoring method.

[0075] S102. Perform image segmentation processing on the raw temperature data to obtain temperature segmentation data for each region, calculate the average temperature and temperature variance of the temperature segmentation data to obtain temperature field evaluation data for each region, and perform pixel fusion processing on the infrared thermal imaging data to obtain visualized image data.

[0076] The regions are divided based on three key thermal deformation-sensitive areas: the CNC machine tool guide rail mounting surface, the worktable support surface, and the bed side. These areas form a direct or substructure-specific correspondence with the three core areas, essentially transforming the "physical locations" of the S101 data acquisition into "data units" for subsequent data processing. This manifests as complete spatial matching, direct binding to data sources, and high adaptability to subsequent processing, achieving a complete closed loop from "physical location, data acquisition, region division to data processing." The number of regions is based on the principle of "covering all key locations and meeting the specific requirements of data processing," with a basic number of three regions, each corresponding one-to-one with the three key locations: the guide rail mounting surface, the worktable support surface, and the bed side. This is suitable for scenarios with moderate monitoring accuracy requirements. For higher precision monitoring, sub-regions can be further divided under the three basic regions, increasing the number of regions accordingly. However, all sub-regions still belong to the three core key locations, always revolving around the guide rail mounting surface, the worktable support surface, and the bed side, ensuring that data processing is both accurate and relevant to the actual physical locations.

[0077] Optionally, step S102 involves performing pixel fusion processing on the infrared thermal imaging data to obtain visualized image data, including:

[0078] Step 1021: Extract pixel information from the infrared thermal imaging data to obtain the temperature value and position coordinates of each pixel.

[0079] Step 1022: According to the preset temperature and display feature correspondence rules, the temperature value is converted to obtain the pixel display information of each pixel.

[0080] Step 1023: Merge the position coordinates with the corresponding pixel display information to obtain pixel fusion information.

[0081] Step 1024: Arrange all the pixel fusion information according to the distribution order of the position coordinates on the CNC machine tool bed to obtain pixel arrangement information.

[0082] Step 1025: Integrate all the pixel display information in the pixel arrangement information to form visualized image data.

[0083] In the above scheme, temperature segmentation data refers to the set of temperature data from all temperature acquisition points within each region after dividing the raw temperature data according to the actual structural area of ​​the CNC machine tool. This data is used to further calculate the overall temperature characteristics of each region. Average temperature refers to the arithmetic mean of the temperature segmentation data within each region, reflecting the overall temperature level of that region. Temperature variance refers to the arithmetic mean of the squares of the differences between the temperature segmentation data and the average temperature within each region, reflecting the degree of temperature fluctuation within that region. Temperature field evaluation data refers to the set of data composed of the average temperature and temperature variance of each region. This data is used to quantitatively evaluate the temperature field distribution state of each region, providing a basis for subsequent consistency verification with infrared data. The temperature-display feature correspondence rule refers to the pre-set temperature range and visual display... The correspondence of features is used to transform abstract temperature values ​​into intuitive visual information; pixel display information refers to the visual feature data obtained after the temperature values ​​are transformed by the above rules, which is used for the subsequent visualization of images; pixel fusion information refers to the data unit formed by binding the position coordinates of each pixel with the corresponding pixel display information, ensuring that visual features correspond one-to-one with spatial positions; pixel arrangement information refers to the data sequence formed by arranging all pixel fusion information according to the actual spatial distribution order of the CNC machine tool bed, ensuring that the arrangement order is consistent with the actual structure of the machine tool; the visualized image data refers to the image file formed by integrating the pixel display information in all pixel arrangement information. This data is used to intuitively present the temperature distribution of each area of ​​the CNC machine tool, and at the same time provides a visual basis for subsequent consistency verification with temperature field evaluation data.

[0084] In this application example, image segmentation processing is performed on the raw temperature data: a region growing algorithm is used, with the structural boundaries of the CNC machine tool guide rail mounting surface, worktable support surface, and bed side as the segmentation basis. The raw temperature data and machine tool structural drawings are imported into the data processing terminal, and temperature acquisition point data belonging to the same structural region are grouped into one category to obtain temperature segmentation data for each region; subsequently, the average temperature and temperature variance of the temperature segmentation data are calculated: for each group of temperature segmentation data, the arithmetic mean method is first used to calculate the average temperature, and the calculation formula is as follows: ,in, The value represents the average temperature, and n represents the number of temperature monitoring points in the area. This represents the temperature value at each sampling point; then, the temperature variance is calculated using a variance calculation method, the formula being: ,in, This represents the temperature variance; by combining the average temperature and temperature variance of each region, we obtain the temperature field assessment data for each region.

[0085] First, pixel information is extracted in step 1021. A data parsing tool supporting the infrared thermal imager is called in the data processing terminal to read the pixel matrix of each frame of the infrared thermal imaging data. The temperature value of each pixel is parsed according to the temperature-gray scale correspondence relationship preset by the thermal imager. At the same time, according to the calibration parameters of the relative position between the thermal imager and the machine tool, the pixel coordinates of the pixel in the image are converted into the position coordinates in the machine tool bed coordinate system, and the temperature value and position coordinates corresponding to each pixel are obtained. Then, after the pixel information extraction is completed, the pixel display information is converted in step 1022. The preset temperature and display feature correspondence rule is called to match the temperature value of each pixel to the corresponding visual display feature, and the pixel display information of each pixel is obtained. Then, in step 1023, the pixel information is fused. The position coordinates of each pixel are bound to the corresponding pixel display information to form pixel fusion information. After that, in step 1024, the pixel fusion information is arranged in order. According to the actual spatial distribution order from left to right and from top to bottom of the machine tool bed, all the pixel fusion information is sorted to ensure that the sorted pixel fusion information is consistent with the actual positions of the key areas of the machine tool in space, and pixel arrangement information is obtained. Finally, in step **1025**, the visualization image data is integrated. All the pixel display information in the pixel arrangement information is integrated into a continuous image pixel matrix according to the arrangement order, and an image file that can be viewed in the data processing terminal is generated to obtain the visualization image data.

[0086] In practical applications, in the production scenario of machining precision parts on a numerically controlled machine tool, temperature raw data processing is first carried out: The region growing algorithm tool in the data processing terminal is called, and the temperature raw data of 10 sensors collected in step S101 and the structure drawing of the numerically controlled machine tool are imported. According to the guide rail installation surface, workbench support surface, and bed side boundary marked in the drawing, the 4 sensor data on the X-axis and Y-axis guide rail installation surfaces are divided into guide rail area temperature segmentation data, the 4 sensor data on the workbench support surface are divided into workbench area temperature segmentation data, and the 2 sensor data on the bed side are divided into bed side area temperature segmentation data. For the 4 temperature data at a certain moment in the guide rail area, which are 25°C, 26°C, 25°C, and 26°C respectively, the average temperature formula is used for calculation. Among them, is the average temperature, n is the number of temperature acquisition points and takes the value 4, is the temperature value of each acquisition point and takes the values 25°C, 26°C, 25°C, and 26°C in sequence. Substituting into the formula for calculation, we get °C; then the formula is used to calculate the temperature variance. Among them, is the temperature variance, and the values of n, , and are the same as those in the average temperature calculation. Substituting into the formula for calculation, we get Similarly, the average temperature of the workbench area was calculated to be 26.2℃ with a temperature variance of 0.3℃², and the average temperature of the bed side area was calculated to be 28.1℃ with a temperature variance of 0.15℃². The average temperature and temperature variance of the three areas were combined to obtain the temperature field assessment data at that moment.

[0087] Next, infrared thermal imaging data processing is carried out: the infrared thermal imaging data with a resolution of 640×512 pixels and a total of 72,000 frames acquired in step S101 is imported, and the pixel matrix of each frame image is read through the analysis tool of the thermal imager. The temperature value of each pixel is analyzed according to the temperature-grayscale correspondence preset by the thermal imager. At the same time, according to the relative position calibration parameters between the thermal imager and the machine tool, the row and column coordinates of the pixels in the image are converted into position coordinates in the machine tool bed coordinate system. For example, the pixel in the 100th row and 200th column of a certain frame image is analyzed to have a temperature value of 25.5℃, and the corresponding bed position coordinates after conversion are X=150mm and Y=220mm. The preset temperature and display feature correspondence rule is invoked, which sets 25-26℃ to blue, 26-27℃ to green, 27-28℃ to yellow, and 28-30℃ to red. The temperature value of 25.5℃ is matched with blue to obtain the pixel display information of the pixel. Then, the position coordinates X=150mm and Y=220mm are bound with the blue display information to form the pixel fusion information of the pixel.

[0088] Then, according to the actual spatial distribution order of the left guide rail, the middle worktable, and the right side of the machine tool bed, the fused information of 327,680 pixels in each frame image is sorted to obtain pixel arrangement information, where 327,680 is the number of pixels in a single frame image, obtained through the formula... The calculations show that P is the number of pixels per frame, W is the image width in pixels (640), and H is the image height in pixels (512). All pixel display information in the pixel arrangement information is integrated into a continuous image pixel matrix according to the arrangement order, generating visualized image data that can be viewed on the data processing terminal. The temperature field assessment data and visualized image data obtained in this step will be used for consistency verification in step S103 to verify the reliability of the two types of data and provide accurate data support for subsequent temperature field clustering and thermal error calculation.

[0089] The aforementioned S102 overall scheme, through the segmentation of raw temperature data and the calculation of temperature characteristics, transforms fragmented temperature acquisition data into temperature field assessment data that quantifies the temperature levels and fluctuations in each region, providing a quantitative basis for subsequent assessment of the reliability of temperature data. Simultaneously, through pixel-level processing of infrared thermal imaging data, abstract infrared data is transformed into visualized image data corresponding to the actual structure of the machine tool, intuitively presenting the differences in temperature distribution across regions and facilitating rapid identification of abnormal temperature areas. This simultaneous processing of both types of data compensates for the deficiency of single temperature data in intuitively presenting the overall distribution, as well as the lack of quantified temperature characteristics in single infrared data. This lays a reliable foundation for subsequent consistency verification of temperature field assessment data and visualized image data, ensuring that the data upon which subsequent thermal error calculations rely is both accurate and comprehensive.

[0090] S103. Perform consistency verification between the temperature field assessment data and the visualization image data to obtain a collaborative dataset. Based on the collaborative dataset, use a region clustering algorithm to determine the temperature field clustering results of each clustering unit within the region.

[0091] Optionally, step S103 involves verifying the consistency between the temperature field assessment data and the visualization image data to obtain a collaborative dataset, and using a region clustering algorithm based on the collaborative dataset to determine the temperature field clustering results of each clustering unit within the region, including:

[0092] Step 1031: Extract temperature information from the visualized image data to obtain image temperature information for each region.

[0093] Step 1032: Perform consistency verification processing on the image temperature information and the average temperature and temperature variance of the corresponding area in the temperature field evaluation data to obtain the verification result.

[0094] Step 1033: Determine the verification results to filter out valid area data, summarize all the valid area data, and form a collaborative dataset.

[0095] Step 1034: Using a region clustering algorithm, regions in the collaborative dataset whose temperature values ​​meet preset temperature conditions and whose pixel positions are connected are divided into a unit, resulting in multiple clustering units.

[0096] Step 1035: Summarize all the clustering units and their corresponding temperature values ​​and pixel location information to obtain the temperature field clustering results.

[0097] In the above scheme, image temperature information refers to the temperature feature data of each region extracted from the visualized image data, including the average temperature and temperature range calculated from the temperature values ​​of all pixels in each region. This data is used to compare with the corresponding indicators in the temperature field assessment data to verify the consistency of the two types of data. The consistency verification process refers to the process of comparing and analyzing the image temperature information with the average temperature and temperature variance of the corresponding region in the temperature field assessment data. By judging whether the difference between the two is within a preset reasonable range, the reliability of the data is verified. The collaborative dataset refers to the unified dataset formed after summarizing all valid region data. This dataset integrates the verified temperature features and image association information for subsequent region clustering calculations. The input to the algorithm is as follows: Region clustering algorithm refers to an algorithm used to divide regions with similar characteristics in a dataset into unified units, thereby achieving a structured division of the temperature field; preset temperature conditions refer to the pre-defined temperature value range used for clustering, ensuring that the temperature characteristics within the cluster units are similar; a cluster unit refers to an independent region unit obtained by the region clustering algorithm that meets the preset temperature conditions and has connected pixel positions. Each cluster unit corresponds to a set of temperature values ​​and pixel position information; the temperature field clustering result refers to the result set formed by summarizing all cluster units and their corresponding temperature values ​​and pixel position information. This result clearly presents the structured distribution of the temperature field, providing a basis for subsequent extraction of the main areas affecting thermal deformation.

[0098] In this application example, firstly, image temperature information is extracted through step 1031: an image analysis tool is called in the data processing terminal to import the visualized image data obtained in step S102. Based on the structural correspondence between the CNC machine tool guide rail mounting surface, the worktable support surface, and the side of the bed, multiple regions corresponding to the temperature field evaluation data are divided in the visualized image. For each divided region, the temperature value corresponding to all pixels in that region is extracted, and the arithmetic mean method is used to calculate the average image temperature of that region. At the same time, the maximum and minimum values ​​of the pixel temperature values ​​in that region are counted to determine the image temperature range. The two together constitute the image temperature information of each region.

[0099] Next, after obtaining the image temperature information for each region, a consistency verification process is performed through step 1032: For each region, the difference between the average image temperature obtained in step 1031 and the average temperature of the corresponding region in the temperature field evaluation data obtained in step S102 is calculated to determine whether the difference is within the preset reasonable temperature difference range; at the same time, based on the average temperature and temperature variance of the corresponding region in the temperature field evaluation data, the temperature fluctuation range of the region is derived, and the image temperature range obtained in step 1031 is compared with the fluctuation range to determine whether the image temperature range is within an acceptable matching degree; the verification result for each region is formed by combining the temperature difference judgment result and the temperature range matching result.

[0100] Then, after completing the consistency verification process and obtaining the verification results, the verification results are judged and a collaborative dataset is formed through 1033 steps: the verification result judgment rules are preset, and areas where "the temperature difference is within a reasonable range and the temperature range matching degree meets the requirements" are judged as valid areas, while areas where "the temperature difference exceeds a reasonable range or the temperature range matching degree does not meet the requirements" are judged as invalid areas; all valid areas are selected from the verification results, and the temperature field evaluation data corresponding to each valid area is combined with the image temperature information to form valid area data; all valid area data are summarized to form a unified collaborative dataset.

[0101] After obtaining the collaborative dataset, clustering units are divided through 1034 steps: The region clustering algorithm is called in the data processing terminal, the collaborative dataset is imported into the algorithm, and a preset temperature condition is set to clarify the allowable difference range of temperature values ​​within the clustering unit; the algorithm first selects an unclassified pixel in the collaborative dataset as an initial point, finds all pixels whose surrounding pixel positions are connected to the initial point and whose temperature values ​​meet the preset temperature condition, and classifies these pixels into a temporary unit; the above search and classification process is repeated for unclassified pixels in the collaborative dataset until all pixels in the collaborative dataset are classified into the corresponding units, resulting in multiple independent clustering units.

[0102] Finally, after completing the clustering unit division, the temperature field clustering results are obtained by summarizing through 1035 steps: for each clustering unit, the average temperature of all pixel temperature values ​​in the unit is calculated to determine the temperature value corresponding to the unit, and the pixel position coordinate range of the unit in the CNC machine tool bed coordinate system is recorded to determine the position information corresponding to the unit; all clustering units and their corresponding temperature values ​​and position information are summarized to form the temperature field clustering results.

[0103] In practical applications, in production scenarios where CNC machine tools are used to process precision parts, the process begins by using the image analysis tool in the data processing terminal to import the visualized image data obtained in step S102. Based on the structural correspondence between the CNC machine tool guideway mounting surface, the worktable support surface, and the bed side, three corresponding regions are divided in the image. For the image region corresponding to the guideway mounting surface, the temperature values ​​of 10,000 pixels within that region are extracted, and the image average temperature formula is applied. Calculate, where, Let m represent the average temperature of the i-th image region, and m represent the number of pixels in that region. This represents the temperature value of the j-th pixel in the region. Substituting m=10000 and the temperature values ​​of each pixel, we can calculate... =25.6℃, and the maximum pixel temperature in this area was calculated to be 26.2℃ and the minimum pixel temperature was calculated to determine the image temperature range; similarly, the average temperature of the image area corresponding to the workbench support surface was calculated. =26.3℃, temperature range 26-26.6℃, average temperature of the image area corresponding to the side of the bed. =28.0℃, temperature range 27.8-28.2℃, obtained image temperature information for three regions.

[0104] Next, with a preset reasonable temperature difference range of 0.3℃ and a temperature range tolerance of 0.2℃, for the guide rail mounting surface area, the difference between the calculated average temperature of 25.6℃ and the average temperature of 25.5℃ obtained from the temperature field evaluation data in step S102 is 0.1℃. 0.1℃ < 0.3℃, which is within a reasonable range. Based on the average temperature of 25.5℃ and the temperature variance of 0.25℃², the formula is used... Calculate the temperature fluctuation range, where, The average temperature representing the temperature field assessment data, The temperature variance, representing the temperature field assessment data, is substituted into the equation to obtain... ℃ to The image temperature range is 25-26.2℃, which exceeds the fluctuation range by 0.2℃. This excess value is within the tolerance range, and the overall verification result of this area is determined to be consistent. Similarly, the verification results of the workbench support surface area and the bed side area are determined to be consistent, and the verification results of the three areas are recorded.

[0105] Then, the temperature field evaluation data and image temperature information corresponding to the three consistent regions are combined into effective region data, and the three effective region data are summarized to form a collaborative dataset. A density clustering algorithm is used, with the preset temperature condition that the pixel temperature difference within a unit does not exceed 0.8℃. The collaborative dataset is imported into the algorithm. The algorithm selects the pixel with a temperature of 25.5℃ in the guide rail mounting surface area as the initial point, finds adjacent pixels with temperatures between 25.5-26.3℃ and assigns them to cluster unit 1. Similarly, the algorithm selects the pixel with a temperature of 26.3℃ in the workbench support surface area as the initial point, finds adjacent pixels with temperatures between 26.3-27.1℃ and assigns them to cluster unit 2. Finally, the algorithm selects pixels with temperatures of 27.8℃, 28.0℃, and 28.2℃ in the bed side area as initial points, assigning them to cluster units 3, 4, and 5 respectively, resulting in 5 cluster units.

[0106] Finally, the average temperature of each cluster unit is calculated, with cluster unit 1 at 25.8℃, unit 2 at 26.5℃, unit 3 at 27.9℃, unit 4 at 28.0℃, and unit 5 at 28.1℃. Simultaneously, the coordinate range of each unit in the machine tool bed coordinate system is recorded: X=50-150mm and Y=100-200mm for cluster unit 1, X=200-300mm and Y=100-200mm for unit 2, X=350-400mm and Y=50-100mm for unit 3, X=350-400mm and Y=100-150mm for unit 4, and X=350-400mm and Y=150-200mm for unit 5. The average temperature and coordinate range of all cluster units are then summarized to obtain the temperature field clustering result. This result is subsequently used in step S104 to extract the temperature and area data of each cluster unit.

[0107] The above-mentioned S103 overall solution effectively filters out reliable and valid regional data by verifying the consistency between temperature field assessment data and visualization image data, eliminating unreliable information caused by single data anomalies, and ensuring the accuracy of the data foundation for subsequent processing. By summarizing the valid regional data to form a collaborative dataset, it integrates the two types of verified data, providing comprehensive and reliable input for region clustering. Using a region clustering algorithm, it divides regions with similar temperatures and connected locations in the collaborative dataset into clustering units, transforming the originally scattered temperature data into structured temperature field clustering results, clearly presenting the distribution characteristics of the temperature field, avoiding inefficiency or bias caused by data clutter during subsequent feature extraction, and laying a structured data foundation for accurately locating the main influencing areas of thermal deformation of CNC machine tools. At the same time, the entire process takes into account both data reliability and structured processing, improving the accuracy of subsequent thermal error calculation.

[0108] S104. Extract features from the temperature field clustering results to obtain temperature data and area data of each cluster unit. Based on the temperature data and area data, obtain the main influence area of ​​thermal deformation of CNC machine tool.

[0109] Among them, the "mainly affected area" is related to the "clustering unit", "region", and "CNC machine tool guide rail mounting surface, worktable support surface and bed side" in a hierarchical and focused manner. Specifically, the "CNC machine tool guide rail mounting surface, worktable support surface and bed side" are the most basic heat deformation sensitive physical parts of the CNC machine tool. The "region" in S102 is a data processing unit divided based on these three physical parts. The "clustering unit" in S103 is a structured unit with similar temperature and connected position, divided by the region clustering algorithm after consistency verification of the "region" in S102. The "mainly affected area" is the spatial range corresponding to the highest priority clustering unit selected from all "clustering units" after temperature sorting, area sorting, and priority evaluation. Its spatial range essentially still belongs to one or more basic physical parts of the "CNC machine tool guide rail mounting surface, worktable support surface and bed side". It is a further precise focus on the basic physical parts, "region" and "clustering unit", and finally locks the core spatial range that contributes the most to the heat deformation of the CNC machine tool.

[0110] Optionally, in step S104, feature extraction is performed on the temperature field clustering results to obtain temperature data and area data of each cluster unit. Based on the temperature data and the area data, the main influencing areas of thermal deformation of the CNC machine tool are obtained, including:

[0111] Step 1041: Extract temperature features from the temperature field clustering results to obtain temperature data for each cluster unit. Perform regional range statistics on each cluster unit in the temperature field clustering results to obtain area data for each cluster unit.

[0112] Step 1042: Sort the temperature data of all cluster units according to their numerical values ​​to obtain a temperature sequence. Select the cluster units corresponding to the highest temperatures in the temperature sequence according to a preset temperature ratio as the first candidate units. It should be understood that this first candidate unit may also be called a high-temperature candidate unit.

[0113] Step 1043: Sort the area data corresponding to all the first candidate units according to their numerical values ​​to obtain an area sequence. Select the clustering units corresponding to the areas at the top of the area sequence according to a preset area ratio as the second candidate units. It should be understood that the second candidate unit may also be called a large area candidate unit.

[0114] Step 1044: Prioritize the large-area candidate units to obtain the priority level of each large-area candidate unit. Sort the priority levels of all large-area candidate units. Based on the sorting results, select the spatial range corresponding to the large-area candidate unit with the highest ranking as the main area affected by thermal deformation of the CNC machine tool.

[0115] In the above scheme, temperature data refers to the temperature feature data of each cluster unit extracted from the temperature field clustering results, including the average temperature, maximum temperature, and minimum temperature of all pixels within each cluster unit. This data reflects the temperature level of each cluster unit and provides a basis for screening high-temperature units. Area data refers to the area value obtained by statistically analyzing the regional range of each cluster unit, calculated from the coordinate range of the cluster unit in the coordinate system of the CNC machine tool bed. This data reflects the spatial size of each cluster unit and provides a basis for screening large-area units. Temperature sequence refers to the ordered data set formed by sorting the temperature data of all cluster units from largest to smallest or smallest to largest, used to visually present the temperature ranking of each unit. Preset temperature ratio refers to the pre-set threshold ratio for screening high-temperature candidate units, which determines the number of units with the highest temperatures to be retained. High-temperature candidate units refer to the cluster units with the highest temperature values ​​selected from the temperature sequence according to the preset temperature ratio. These units have a higher potential impact on thermal deformation due to their higher temperatures. Area sequence refers to the total area of ​​all high-temperature candidate units. The area data of candidate units are sorted by numerical value from largest to smallest or smallest to largest to form an ordered data set, used to visually present the size ranking of high-temperature candidate unit areas; the preset area ratio refers to the pre-set threshold ratio for screening large-area candidate units, which determines the number of units with the highest area to be retained; large-area candidate units refer to high-temperature candidate units with the highest area values ​​selected from the area sequence according to the preset area ratio. These units combine high temperature and large spatial range, and are key potential sources of thermal deformation; priority assessment refers to the process of calculating the priority score of each unit by combining the temperature data and area data of large-area candidate units according to preset weight rules. Units with higher temperature and larger area usually have higher priority; priority level refers to the result of classifying large-area candidate units according to priority scores, used to distinguish the importance of each unit to thermal deformation; the main influence area refers to the spatial range corresponding to the large-area candidate unit with the highest priority level. This area is the core area that contributes the most to the thermal deformation of CNC machine tools, and is the key object to focus on in subsequent thermal error calculations.

[0116] In this application example, firstly, temperature data and statistical area data are extracted through step 1041: the feature extraction tool is called in the data processing terminal, the temperature field clustering results obtained in step S103 are imported, and for each cluster unit, the recorded temperature value and temperature fluctuation related data are extracted and combined to form the temperature data of each cluster unit; at the same time, based on the coordinate range of the CNC machine tool bed coordinate system recorded by each cluster unit, the length of the unit in the X-axis direction and the length in the Y-axis direction are calculated, and the area value of the unit is obtained by multiplying the two lengths, thus completing the area statistics of all units and obtaining the area data of each cluster unit.

[0117] Next, after obtaining the temperature and area data of each cluster unit, high-temperature candidate units are screened through 1042 steps: In the data processing terminal, a data sorting algorithm is used to sort the temperature data of all cluster units from largest to smallest to form a temperature sequence; a preset temperature ratio is called, and the number of units to be retained is calculated based on the total number of cluster units and the ratio; the units that are ranked first in the temperature sequence and meet the quantity requirements are selected to obtain high-temperature candidate units.

[0118] Then, after determining the high-temperature candidate units, based on the area data corresponding to the high-temperature candidate units, large-area candidate units are screened through 1043 steps: using the same data sorting algorithm as temperature sorting, the area data of all high-temperature candidate units are sorted from largest to smallest to form an area sequence; a preset area ratio is called, and the number of units to be retained is calculated based on the total number of high-temperature candidate units and the ratio. Units that are ranked first in the area sequence and whose number meets the requirements are selected to obtain large-area candidate units.

[0119] Finally, after obtaining large-area candidate units, the priorities are evaluated and the main influencing areas are determined through 1044 steps: a priority evaluation rule is preset, and the weights of temperature data and area data are set. For each large-area candidate unit, the standardized value of its temperature data is multiplied by the temperature weight, and the standardized value of its area data is multiplied by the area weight. The two are added together to calculate the priority score of the unit. Based on the priority score, all large-area candidate units are divided into priority levels. The priority levels are sorted, and the large-area candidate unit with the highest priority is selected. The spatial range of the CNC machine tool bed coordinate system corresponding to this unit is extracted as the main influencing area of ​​thermal deformation of the CNC machine tool.

[0120] In practical applications, in the production scenario of CNC machine tool machining precision parts, firstly, the feature extraction tool in the data processing terminal is called to import the temperature field clustering results of the five clustering units obtained in step S103, and the average temperature of each unit is extracted as temperature data, namely Unit 1: 25.8℃, Unit 2: 26.5℃, Unit 3: 27.9℃, Unit 4: 28.0℃, and Unit 5: 28.1℃; at the same time, the area calculation formula is used. Calculate the area data for each cluster, where S is the area of ​​the cluster unit. The length of the element along the X-axis is equal to the element's ending X-axis coordinate minus its starting X-axis coordinate. The length of the unit along the Y-axis is equal to the unit's Y-axis termination coordinate minus its Y-axis start coordinate. Unit 1: X-axis start coordinate 50mm, termination coordinate 150mm. =150-50=100mm, Y-axis start coordinate 100mm, end coordinate 200mm. =200-100=100mm, substituting into the formula, we get... =100×100=10000mm²; Unit 2: X-axis starting coordinate 200mm, ending coordinate 300mm. =300-200=100mm, Y-axis start coordinate 100mm, end coordinate 200mm. =200-100=100mm, therefore... =100×100=10000mm²; Unit 3: X-axis starting coordinate 350mm, ending coordinate 400mm =400-350=50mm, Y-axis start coordinate 50mm, end coordinate 100mm =100-50=50mm, therefore... =50×50=2500mm²; Unit 4: X-axis starting coordinate 350mm, ending coordinate 400mm =400-350=50mm, Y-axis start coordinate 100mm, end coordinate 150mm =150-100=50mm, therefore... =50×50=2500mm²; Unit 5: X-axis starting coordinate 350mm, ending coordinate 400mm =400-350=50mm, Y-axis start coordinate 150mm, end coordinate 200mm =200-150=50mm, therefore... =50×50=2500mm².

[0121] Next, the temperature data of the five units are sorted from largest to smallest using the quicksort algorithm, forming a temperature sequence of 28.1℃ (unit 5) > 28.0℃ (unit 4) > 27.9℃ (unit 3) > 26.5℃ (unit 2) > 25.8℃ (unit 1). With a preset temperature ratio of 40%, the number of units to be retained is calculated to be 5 × 40% = 2. Units 5 and 4, ranking first and second respectively, are selected as high-temperature candidate units. Then, the area data of the high-temperature candidate units are sorted from largest to smallest using the quicksort algorithm, forming an area sequence of 2500mm² (unit 5) = 2500mm² (unit 4). With a preset area ratio of 100%, the number of units to be retained is calculated to be 2 × 100% = 2. Units 5 and 4 are selected as large-area candidate units.

[0122] Finally, a priority evaluation rule was preset, with temperature weighting at 0.6 and area weighting at 0.4. First, the temperature and area data were standardized, mapping them to the 0-1 range. Unit 5, with a temperature of 28.1℃, had the highest standardized value of 1, and its area of ​​2500mm² was the highest among the high-temperature candidate units, also with a standardized value of 1, resulting in a priority score of 1×0.6+1×0.4=1.0. Unit 4, with a temperature of 28.0℃, had a standardized value of 0.9, and its area of ​​2500mm² had a standardized value of 1, resulting in a priority score of 0.9×0.6+1×0.4=0.94. Priority levels were assigned based on the scores: Unit 5 was classified as Level 1, and Unit 4 as Level 2. Unit 5 was ranked first in priority. The coordinate range X=350-400mm and Y=150-200mm corresponding to Unit 5 in the bed coordinate system was extracted as the main influence area. This area was subsequently used in step S105 for correlation calculation using the thermal error mapping algorithm to obtain the thermal error value.

[0123] The aforementioned S104 overall scheme extracts features from the temperature field clustering results to accurately obtain temperature and area data for each cluster unit, providing a quantitative basis for subsequent screening of key areas. Through a dual screening process of "temperature sorting to screen high-temperature units → area sorting to screen large-area units," it gradually focuses on candidate units that combine high temperature and large spatial range, eliminating units with low temperature or small area and limited impact on thermal deformation, thus reducing redundant objects in subsequent thermal error calculations. By prioritizing candidate units according to their importance, the unit with the highest ranking is finally determined as the main influencing area, ensuring that subsequent thermal error calculations can accurately focus on the core area that contributes the most to thermal deformation, avoiding the reduction in accuracy or inefficiency caused by scattered calculation of the entire data area, laying a key foundation for improving the accuracy and targeting of thermal error monitoring.

[0124] S105. A thermal error mapping algorithm is used to perform correlation calculation on the temperature data and area data corresponding to the main affected area to obtain the thermal error value.

[0125] Optionally, step S105 involves using a thermal error mapping algorithm to perform correlation calculations on the temperature data and area data corresponding to the main affected area to obtain a thermal error value, including:

[0126] Step 1051: Define the scope of the parts of the CNC machine tool bed that need to be monitored to identify the key locations requiring focused monitoring. The parts of the bed to be monitored are related to the "CNC machine tool guide rail mounting surface, worktable support surface, and bed side" through a "core inclusion and precise focus" relationship: the former is based on these three core areas because the guide rail mounting surface directly affects the guide rail's motion accuracy, the worktable support surface determines the worktable's load-bearing stability, and the bed side is often close to the power components and prone to generating concentrated heat sources. All three are key sources of thermal deformation in the bed, therefore the parts of the bed to be monitored must cover the core areas of these three areas. Simultaneously, the parts of the bed to be monitored will further focus on more critical sub-areas within the scope of these three areas. Essentially, this involves selecting "smaller monitoring points that have a more direct impact on thermal deformation and machining accuracy" from "larger key areas." This approach does not deviate from the scope of these three core areas, but through precise focus, ensures that the monitoring data of subsequent key locations directly reflects the thermal deformation state of the core areas, providing accurate physical location basis for subsequent correlation of major influencing areas and calculation of thermal errors.

[0127] Step 1052: Based on the spatial relationship between the key location and the main influence area, establish the association between the key location and the temperature data and area data of each cluster unit in the corresponding main influence area, and obtain a correspondence table between location and data.

[0128] Step 1053: Based on the correspondence table, use the thermal error mapping algorithm to calculate the preliminary error value of each key position, correct the preliminary error value, and obtain the target error value.

[0129] Step 1053 may specifically include the following steps: obtaining temperature data, area data, and spatial characteristic parameters corresponding to each key location from the correspondence table; calculating the difference between the corresponding temperature data and a preset reference temperature to obtain a temperature deviation value; converting the corresponding area data according to a preset area weighting rule to obtain an area weight value; using a thermal error mapping algorithm, multiplying the temperature deviation value, the area weight value, and the thermal deformation sensitivity coefficient in the spatial characteristic parameters corresponding to the same key location to obtain a preliminary error base; converting the preliminary error base corresponding to each key location according to a preset error unit conversion rule to obtain a preliminary error value for each key location; and then converting the temperature deviation value, area weight value, and thermal deformation sensitivity coefficient in the spatial characteristic parameters corresponding to the key location according to the preset error unit conversion rule to obtain a preliminary error value for each key location. The initial error value is compared with a preset reasonable range to obtain a first comparison result. The difference between the initial error value and adjacent initial error values ​​is calculated, and the difference is compared with a preset value change range requirement to obtain a second comparison result. The first comparison result and the second comparison result are combined to obtain the error check result corresponding to each key position. The error check results are classified according to a preset classification rule to obtain qualified results and results to be adjusted. The initial error value corresponding to the qualified result is determined as the error value to be confirmed. The initial error value corresponding to the result to be adjusted is adjusted using a preset correction rule to obtain the error value to be confirmed. The uniqueness of all the error values ​​to be confirmed is verified to obtain the target error value corresponding to each key position.

[0130] The physical meaning of adjacent preliminary error values ​​refers to the preliminary error values ​​corresponding to key locations that are spatially adjacent in the physical layout of the CNC machine tool bed, i.e., close in distance and directly related to the transmission of thermal deformation. These spatially adjacent key locations, being located in the same area of ​​the bed or distributed along the thermal deformation transmission path, exhibit continuous thermal deformation states, and their corresponding preliminary error values ​​should also show reasonable correlation characteristics. The core function of this value is to verify the spatial rationality of the preliminary error values ​​of the target key location by calculating the difference with the preliminary error values ​​of the target key location, ensuring that it conforms to the physical law of continuous transmission of thermal deformation of the bed, and providing a spatial dimension basis for subsequent error correction.

[0131] Step 1054: Summarize the target error values ​​at all key locations to obtain the thermal error value.

[0132] In the above scheme, the critical location refers to the bed parts that have a significant impact on the machining accuracy of CNC machine tools and require key monitoring. These are identified by analyzing the correlation between machine tool structure and accuracy, serving as the core objects for focusing thermal error calculations. Spatial correlation refers to the spatial relationship between the critical location and each cluster unit in the main affected area, including whether the critical location is located within a cluster unit or is adjacent to one, used to determine the monitoring data corresponding to the critical location. The correspondence table is a table formed by binding the temperature data, area data, and spatial characteristic parameters of the critical location with the corresponding cluster unit. Spatial characteristic parameters refer to the inherent attribute data of the critical location within the machine tool bed, including the distance between the location and moving parts, and the material's thermal deformation sensitivity coefficient. The parameters include: thermal deformation sensitivity coefficient, which reflects the influence of material temperature changes on deformation; temperature deviation value, which is the difference between the temperature data at key locations and the preset reference temperature, used to quantify the deviation of the actual temperature from the normal state; area weighting rule, which is a pre-set standard for converting area data into weight values, used to reflect the difference in the impact of different area sizes on thermal error; area weight value, which is the value obtained after area data is converted through the area weighting rule, used to participate in thermal error correlation calculation; preliminary error base, which is the basic error value obtained by multiplying the temperature deviation value, area weight value, and thermal deformation sensitivity coefficient through the thermal error mapping algorithm, reflecting the original calculation result of thermal error; and error unit conversion rule, which is the rule for converting the initial error base into a basic error value. The initial error base is converted to the standard units commonly used in machine tool accuracy monitoring to ensure that the error values ​​meet the actual application requirements; the preliminary error value refers to the error value obtained after unit conversion of the initial error base, providing initial data for subsequent error correction; the reasonable range of values ​​refers to the pre-set normal range in which the preliminary error value should be, used to determine whether the preliminary error is abnormal; the first comparison result refers to the conclusion obtained after comparing the preliminary error value with the reasonable range of values, indicating whether it conforms to or exceeds the range; adjacent preliminary error values ​​refer to the preliminary error values ​​corresponding to key locations that are spatially adjacent, used to determine whether the error change is stable; the numerical change amplitude requirement refers to the pre-set maximum allowable difference between the preliminary error values ​​of adjacent key locations, used to determine whether the error change is stable. The first comparison result determines whether the error change is reasonable; the second comparison result refers to the conclusion that the difference between adjacent preliminary error values ​​meets or exceeds the required range of numerical change; the error check result refers to the comprehensive judgment on the reliability of the preliminary error value formed by combining the first and second comparison results; the classification rule refers to the standard for dividing the error check result into qualified results and results to be adjusted; a qualified result refers to the result corresponding to the preliminary error value that the error check result determines is reliable; a result to be adjusted refers to the result corresponding to the preliminary error value that the error check result determines is abnormal; the correction rule refers to the method for correcting the preliminary error value that is in the result to be adjusted, such as interpolation correction based on adjacent qualified error values.The error value to be confirmed refers to the preliminary error value corresponding to the qualified result or the error value obtained after the result to be adjusted, providing data for uniqueness verification; uniqueness verification refers to checking whether the error value to be confirmed is duplicated or contradictory, ensuring that each key position corresponds to a unique and reliable error value; the target error value refers to the final error value of each key position determined after uniqueness verification, which has high reliability; the thermal error value refers to the data set reflecting the overall thermal error state of the CNC machine tool formed by summarizing the target error values ​​of all key positions, used to reflect the degree of influence of machine tool thermal deformation on machining accuracy.

[0133] In this application example, firstly, the key locations are defined through step 1051: the machine tool structure analysis tool is called in the data processing terminal, the structural drawings and accuracy influence analysis report of the CNC machine tool are imported, and the correlation between these parts and thermal deformation is analyzed in conjunction with the precision-sensitive parts during machine tool processing, such as the periphery of the spindle box, the key nodes of the guide rail slider movement, and the load-bearing center of the worktable. The key locations that need to be monitored are defined, and the coordinate range of each key location in the machine tool bed coordinate system is clarified, thus completing the scope definition of the key locations.

[0134] Next, after delineating the key locations, a correspondence table is established through step 1052: the spatial location analysis tool is called, and the key location coordinates obtained in step 1051, the main influence area determined in step S104, and the temperature and area data of each cluster unit in the area are imported. The spatial relationship between each key location and each cluster unit in the main influence area is analyzed, such as whether the key location coordinates are within the coordinate range of a certain cluster unit, or whether the distance from the coordinate range of the cluster unit is less than a preset threshold. The key locations with spatial relationships are bound to the temperature and area data of the corresponding cluster units. At the same time, the spatial characteristic parameters corresponding to each key location in the machine tool bed are supplemented. All the binding relationships of "key location-temperature data-area data-spatial characteristic parameters" are organized into a table to obtain the correspondence table between location and data.

[0135] Then, after obtaining the correspondence table, the preliminary error values ​​are calculated and corrected through 1053 steps: First, extract the temperature data, area data, and spatial characteristic parameters corresponding to each key location from the correspondence table; Second, calculate the difference between the extracted temperature data and the preset reference temperature to obtain the temperature deviation value of each key location, and convert the area data into area weight values ​​according to the preset area weight rules; Third, use a thermal error mapping algorithm to multiply the temperature deviation value and area weight value of the same key location by the thermal deformation sensitivity coefficient in the spatial characteristic parameters to obtain the preliminary error base value of that key location, and then convert the preliminary error base value according to the preset error unit conversion rules to obtain the preliminary error value of each key location; Fourth, compare each preliminary error value with the preset reasonable range of values ​​to determine whether the preliminary error value is within the normal range, obtain the first comparison result, and calculate the relationship between the preliminary error value and the spatial location. The first step involves comparing the difference between the initial error values ​​corresponding to adjacent key locations with a preset value change range requirement to determine whether the error change is stable and reasonable, thus obtaining a second comparison result. The second step combines the first and second comparison results to form the error check result for each key location. According to a preset classification rule, the error check results are divided into qualified results and results requiring adjustment. The sixth step directly determines the initial error value corresponding to the qualified result as the error value to be confirmed. For the initial error values ​​in the results requiring adjustment, a preset correction rule (such as linear interpolation based on the error values ​​to be confirmed of adjacent qualified key locations) is used for adjustment to obtain the adjusted error value to be confirmed. The seventh step performs a uniqueness verification on all error values ​​to be confirmed, checking for situations where multiple error values ​​correspond to the same key location, or where there are logical contradictions in the error values ​​to be confirmed of different key locations. After verification, the target error value corresponding to each key location is determined.

[0136] Finally, after determining the target error values ​​for all key positions, the thermal error values ​​are summarized through step 1054: the data summarization tool is called to organize the target error values ​​for all key positions obtained in step 1053 according to the spatial distribution order of the key positions in the machine tool bed, and the key position name and coordinate information corresponding to each target error value are labeled to form a summary table containing the correspondence between "key position - target error value - coordinates". This summary table is the thermal error value that reflects the overall thermal error status of the CNC machine tool.

[0137] In practical applications, in the production scenario of CNC machine tools machining precision parts, the machine tool structure analysis tool in the data processing terminal is first called to import the structural drawings and accuracy impact analysis report of the CNC machine tool. Combining the significant impact of the spindle box and guide rail slider on the machining accuracy when the machine tool is machining parts, four key positions are identified: Key Position 1 (below the spindle box, coordinates X=370mm, Y=180mm), Key Position 2 (front slider of the X-axis guide rail, coordinates X=100mm, Y=150mm), Key Position 3 (right slider of the Y-axis guide rail, coordinates X=250mm, Y=180mm), and Key Position 4 (below the center of the worktable, coordinates X=200mm, Y=150mm).

[0138] Next, the spatial location analysis tool was used to import the coordinates of four key locations, the main influence area determined in step S104 (unit 5, coordinate range X=350-400mm, Y=150-200mm, temperature data 28.1℃, area data 2500mm²), and related data of cluster units within this area. Analysis revealed that the coordinates of key location 1 are within the coordinate range of unit 5, indicating a spatial relationship with unit 5. The other key locations are associated with other cluster units obtained in step S103. Each key location was bound to the temperature and area data of the corresponding cluster unit. At the same time, the spatial characteristic parameters of each key location were supplemented, and their thermal deformation sensitivity coefficients were all set to 0.02mm / (℃・weight unit). A correspondence table between location and data was compiled, in which the temperature data of key location 1 is 28.1℃, the area data is 2500mm², and the thermal deformation sensitivity coefficient is 0.02mm / (℃・weight unit).

[0139] Then, the first step is to extract the temperature data (28.1℃), area data (2500mm²), and thermal deformation sensitivity coefficient (0.02mm / (℃・weight unit)) of key location 1 from the correspondence table; the second step is to preset the reference temperature to 25℃, calculate the temperature deviation as 28.1℃ minus 25℃ equals 3.1℃, and preset the area weight rule as 1 weight unit per 2500mm², calculate the area weight as 2500mm² divided by 2500mm² equals 1; the third step is to use the thermal error mapping algorithm to calculate the initial error base, the calculation formula is as follows: ,in, As the initial error base, This is the temperature deviation value. This is the area weight value. The coefficient for sensitivity to thermal deformation is given by substituting it into the numerical calculation. The preset error unit conversion rule is that the initial error base directly corresponds to the millimeter unit, resulting in an initial error value of 0.062mm for key position 1. In the fourth step, the preset reasonable range for the value is 0.02mm to 0.08mm. 0.062mm falls within this range, and the first comparison result meets the range requirement. The adjacent key position for key position 1 is key position 3, and the initial error value for key position 3 is 0.058mm. The difference between the two is calculated as 0.062mm minus 0.058mm, which equals 0.004mm. The preset requirement for the value variation is no more than 0.01mm, and this difference meets the requirement. The second comparison result meets the variation range requirements; the fifth step combines the two comparison results, and the error check result of key position 1 is qualified; the sixth step determines the preliminary error value of key position 1, 0.062mm, as the error value to be confirmed; the seventh step verifies the uniqueness of the error value to be confirmed of key position 1, and no duplication or contradiction is found, so the target error value of key position 1 is determined to be 0.062mm. Similarly, the target error values ​​of key position 2, 0.035mm, key position 3, 0.058mm, and key position 4, 0.042mm, are calculated and verified.

[0140] Finally, the data aggregation tool is called to organize the target error values ​​in the order of key position 1 to key position 4, and the key position name and coordinates corresponding to each target error value are marked to form a summary table. This summary table is the thermal error value when the CNC machine tool is machining the part, which will be used for accuracy compensation and adjustment in the subsequent machining process of the part.

[0141] The aforementioned S105 overall solution focuses on the main areas affected by thermal deformation of CNC machine tools, accurately delineates key locations, and establishes a "location-data" correspondence, avoiding interference from non-critical location data in thermal error calculation and ensuring the targeted nature of the calculation. Through a thermal error mapping algorithm, it correlates and calculates multi-dimensional data such as temperature, area, and material properties, combined with error value rationality checks and anomaly corrections, effectively eliminating unreliable results caused by single data fluctuations or calculation deviations, thus improving the accuracy of initial error values. Uniqueness verification further ensures the reliability of the target error value. The final summarized thermal error value comprehensively and accurately reflects the thermal error state of the CNC machine tool, providing reliable data support for subsequent machine tool accuracy compensation and machining process optimization, meeting the accuracy and practicality requirements of thermal error monitoring in high-precision manufacturing scenarios.

[0142] The following is a complete example for steps S101 to S105, such as Figure 2As shown, in a production scenario where precision parts are machined on a CNC machine tool, the first step is to perform online data acquisition: Two platinum resistance temperature sensors are placed next to the slider on the X-axis guide rail mounting surface and in the middle of the Y-axis guide rail mounting surface; one sensor is placed at each of the four corners of the worktable support surface; and two sensors are placed on the side of the bed near the spindle box and the X-axis motor, for a total of 10 sensors. The sensors are connected to an industrial computer via a data acquisition module, with a acquisition frequency set to 1 second / time, and a part processing time of 2 hours (equivalent to 7200 seconds). The total amount of raw temperature data is calculated using the formula... Where Q is the total amount of raw temperature data, N is the number of sensors, f is the acquisition frequency, and T is the acquisition duration. Substituting these values, we can calculate... These 72,000 raw temperature data points were subsequently used for image segmentation in the data processing steps. Simultaneously, an infrared thermal imager was fixed on a tripod 1.8 meters above the machine tool's operating side, with the image resolution set to 640×512 pixels and the frame rate set to 10 frames per second. The total number of infrared thermal imaging frames was calculated using the formula... Where F is the total number of infrared thermal imaging frames. Let T be the frame rate and T be the duration of the data acquisition. Substituting the values ​​into the input... Frame; calculated according to the formula for the number of pixels in a single frame image. Where P is the number of pixels in a single frame, W is the image width in pixels, and H is the image height in pixels, the values ​​substituted into the input value are... Pixels; according to the formula for calculating the total pixel data volume in infrared thermal imaging, D=F×P, where D is the total pixel data volume, substituting the values... These infrared thermal imaging data are subsequently used in pixel fusion processing, a data processing step.

[0143] Next, the data processing steps are executed: The region growing algorithm tool in the industrial computer is invoked, importing the aforementioned 72,000 raw temperature data points and the CNC machine tool structural drawings. According to the drawings, the temperature data is divided into three regions: the guide rail mounting surface region, the worktable support surface region, and the bed side region. For the four temperature data points at a specific moment in the guide rail mounting surface region (25℃, 26℃, 25℃, 26℃), the average temperature calculation formula is used. ,in Let n be the average temperature of the region, and n be the number of temperature monitoring points in the region. Substituting the temperature values ​​from each collection point into the numerical calculation yields... The temperature variance calculation formula is used. ,in, The variance of the regional temperature is denoted as , and the numerical values ​​are substituted to obtain . Similarly, the average temperature of the workbench support surface area is calculated to be 26.2℃ with a temperature variance of 0.3℃², and the average temperature of the bed side area is calculated to be 28.1℃ with a temperature variance of 0.15℃². These temperature field assessment data will be used in the subsequent consistency verification step. Meanwhile, for the aforementioned 72,000 frames of infrared thermal imaging data, the temperature value and machine bed coordinates of each pixel in each frame are extracted using the thermal imager's accompanying analysis tool. For example, the temperature value of the pixel in the 100th row and 200th column of a certain frame is 25.5℃, which corresponds to the machine bed coordinates X=150mm and Y=220mm after coordinate calibration. The preset temperature and display feature correspondence rule is called, which sets 25℃-26℃ to blue, 26℃-27℃ to green, 27℃-28℃ to yellow, and 28℃-30℃ to red, converting the 25.5℃ temperature value into blue display information. The machine bed coordinates are bound to the corresponding display information and sorted in the order of the left guide rail, the middle worktable, and the right side of the machine bed, and integrated into visual image data. This data is then checked for consistency with the temperature field evaluation data.

[0144] Subsequently, consistency verification and clustering steps were performed: temperature information for each region was extracted from the visualized image data. The average temperature of the guide rail mounting surface region was 25.6℃, with a temperature range of 25℃-26.2℃; the average temperature of the workbench support surface region was 26.3℃, with a temperature range of 26℃-26.6℃; and the average temperature of the bed side region was 28.0℃, with a temperature range of 27.8℃-28.2℃. Comparing this image temperature information with the aforementioned temperature field assessment data, the temperature difference in the guide rail mounting surface region was 25.6℃-25.5℃=0.1℃. The preset reasonable temperature difference range was 0.3℃, and 0.1℃ is within the reasonable range. The temperature fluctuation range was calculated based on the temperature field assessment data using the formula... ,in The average temperature of the region. Substituting the regional temperature variance into the data for the guide rail mounting surface area, we get... The image temperature range of 25℃-26.2℃ exceeded the fluctuation range by 0.2℃, which is within the preset temperature range tolerance value of 0.2℃, thus the data in this area was determined to be consistent. Similarly, the data in the other two areas were determined to be consistent. The temperature field assessment data and image temperature information from the three areas were combined to form a collaborative dataset, which was subsequently used for region clustering. A density clustering algorithm in an industrial computer was called, and the collaborative dataset was imported. The preset temperature condition was that the pixel temperature difference within a unit did not exceed 0.8℃. The algorithm divided the collaborative data into 5 clustering units, recording the average temperature and coordinate range of each unit. Unit 5 had an average temperature of 28.1℃ and a coordinate range of X=350-400mm and Y=150-200mm. These temperature field clustering results were subsequently used in the step of determining the main influencing regions.

[0145] Next, the main influence area determination step is performed: feature extraction is performed on the five cluster units, extracting temperature and area data for each unit. The area is calculated using a formula... Where S is the area of ​​the cluster unit. The length of the element along the X-axis. Length of element 5 in the Y-axis direction. =400-350=50mm, =200-150=50mm, substituting the values, we get S=50×50=2500mm². Sort the temperature data of the 5 cluster units from largest to smallest: unit 528.1℃, unit 428.0℃, unit 327.9℃, unit 226.5℃, and unit 125.8℃. With a preset temperature ratio of 40%, the number of units to be retained is calculated to be 5×40%=2. Units 5 and 4 are selected as high-temperature candidate units. Sort the area data of the high-temperature candidate units from largest to smallest. Units 5 and 4 both have an area of ​​2500mm². With a preset area ratio of 100%, the number of units to be retained is calculated to be 2×100%=2. Units 5 and 4 are selected as large-area candidate units. A preset priority evaluation rule is used, with temperature weighting at 0.6 and area weighting at 0.4. Temperature and area data for large-area candidate units are standardized. Unit 5 has a standardized temperature value of 1 and a standardized area value of 1. The priority score calculation formula is: Score = Standardized Temperature Value × 0.6 + Standardized Area Value × 0.4. Substituting the values, the score is: Score = 1 × 0.6 + 1 × 0.4 = 1.0. Unit 4 has a standardized temperature value of 0.9 and a standardized area value of 1, so the score is calculated as: 0.9 × 0.6 + 1 × 0.4 = 0.94. Unit 5 is ranked first based on its score, thus determining the area corresponding to Unit 5 as the main area affected by thermal deformation of the CNC machine tool. This area will be used for subsequent thermal error mapping calculations.

[0146] Finally, the thermal error calculation steps were performed: Four key locations were identified: Key Location 1, below the spindle box (X=370mm, Y=180mm); Key Location 2, at the front slider of the X-axis guide rail (X=100mm, Y=150mm); Key Location 3, at the right slider of the Y-axis guide rail (X=250mm, Y=180mm); and Key Location 4, below the center of the worktable (X=200mm, Y=150mm). Key Location 1 is located within the main influence area. A correspondence table was established between the key locations and the main influence area data. Key Location 1 corresponds to a temperature of 28.1℃, an area of ​​2500mm², and a thermal deformation sensitivity coefficient of 0.02mm / (℃・weight unit). The temperature deviation value was calculated, with a preset reference temperature of 25℃. The formula for calculating the temperature deviation value is as follows: =Actual temperature - Reference temperature, substitute the value =28.1-25=3.1℃; The preset area weighting rule is that every 2500mm² corresponds to 1 weighting unit, and the area weighting value is calculated using the following formula: =Divide the area data by 2500, and substitute the result. =2500÷2500=1. The initial error base is calculated using the thermal error mapping algorithm, with the formula: ,in As the initial error base, This is the temperature deviation value. This is the area weight value. Let be the heat deformation sensitivity coefficient, and substitute the values ​​into the equation. =3.1×1×0.02=0.062mm. The preset error unit conversion rule is that the initial error base directly corresponds to millimeters, resulting in an initial error value of 0.062mm for key position 1. Comparing this initial error value with the preset reasonable range of 0.02-0.08mm, 0.062mm falls within this range, resulting in a first comparison result of compliance. The difference between the initial error values ​​of key position 1 and the adjacent key position 3 is calculated. The initial error value of key position 3 is 0.058mm, and the difference is 0.062-0.058=0.004mm. The preset requirement for the variation range is no more than 0.01mm, so the difference meets the requirement, resulting in a second comparison result of compliance. Combining the two comparison results, the initial error value of key position 1 is determined to be acceptable and is identified as the target error value. Similarly, the target error values ​​for key position 2 are calculated to be 0.035mm, key position 3 to be 0.058mm, and key position 4 to be 0.042mm. The target error values ​​of the four key positions are summed to form the thermal error value, which will be used for accuracy compensation and adjustment when CNC machine tools are used to process precision parts.

[0147] Figure 3 This is a schematic diagram of a specific embodiment of an online thermal error monitoring system for CNC machine tools provided in this application. (Refer to...) Figure 3 The system may include:

[0148] The acquisition module 31 is used to acquire raw temperature data of the CNC machine tool guide rail mounting surface, worktable support surface and bed side surface, as well as infrared thermal imaging data of the CNC machine tool online.

[0149] The fusion module 32 is used to perform image segmentation processing on the raw temperature data to obtain temperature segmentation data for each region, calculate the average temperature and temperature variance of the temperature segmentation data to obtain temperature field evaluation data for each region, and perform pixel fusion processing on the infrared thermal imaging data to obtain visualized image data.

[0150] The verification module 33 is used to verify the consistency between the temperature field evaluation data and the visualization image data to obtain a collaborative dataset, and to determine the temperature field clustering result of each clustering unit in the region based on the collaborative dataset using a region clustering algorithm.

[0151] The extraction module 34 is used to extract features from the temperature field clustering results to obtain temperature data and area data of each cluster unit, and based on the temperature data and area data, to obtain the main influence area of ​​thermal deformation of CNC machine tool.

[0152] The calculation module 35 is used to perform correlation calculation on the temperature data and the area data corresponding to the main affected area using a thermal error mapping algorithm to obtain a thermal error value.

[0153] This application provides an online thermal error monitoring system for CNC machine tools to implement the aforementioned online thermal error monitoring method for CNC machine tools. Therefore, the specific implementation of the online thermal error monitoring system for CNC machine tools can be found in the embodiment section of the online thermal error monitoring method for CNC machine tools described above. The specific implementation can be referred to the description of the corresponding embodiments, and will not be repeated here.

[0154] like Figure 4 As shown, this application also provides an electronic device, including: a memory 41 for storing a computer program; and a processor 42 for executing the computer program to implement the steps of any of the above-described online monitoring methods for thermal errors of CNC machine tools.

[0155] This application also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of any of the above-described methods for online monitoring of thermal errors in CNC machine tools.

[0156] In one exemplary embodiment, the aforementioned computer-readable storage medium may include, but is not limited to, various media capable of storing computer programs, such as USB flash drives, read-only memory, random access memory, portable hard drives, magnetic disks, or optical disks.

[0157] Embodiments of the present invention also provide a computer program product, which includes a computer program that, when executed by a processor, implements the steps in any of the embodiments of the online monitoring method for thermal errors of CNC machine tools described above.

[0158] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0159] The above provides a detailed description of the online thermal error monitoring method and system for CNC machine tools provided in this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the embodiments above are merely for the purpose of helping to understand the method and its core ideas. It should be noted that those skilled in the art can make various improvements and modifications to this application without departing from its principles, and these improvements and modifications also fall within the protection scope of this application.

Claims

1. A method for online monitoring of thermal errors of a numerically controlled machine tool, characterized in that, The method comprises the following steps: Collecting temperature raw data of guide rail mounting surface, worktable supporting surface and side surface of the bed of the numerical control machine tool and infrared thermal imaging data of the numerical control machine tool; Performing image segmentation processing on the temperature raw data to obtain temperature segmentation data of each region, calculating average temperature and temperature variance of the temperature segmentation data to obtain temperature field evaluation data of each region, and performing pixel fusion processing on the infrared thermal imaging data to obtain visual image data; Performing consistency verification on the temperature field evaluation data and the visual image data to obtain a collaborative data set, and determining temperature field clustering results of each clustering unit in the region based on the collaborative data set by using a region clustering algorithm; Extracting features of the temperature field clustering results to obtain temperature data and area data of each clustering unit, and obtaining a main influence region of thermal deformation of the numerical control machine tool based on the temperature data and the area data; Performing associated calculation and processing on the temperature data and the area data corresponding to the main influence region by using a thermal error mapping algorithm to obtain a thermal error value.

2. The method of claim 1, wherein, The method of performing associated calculation and processing on the temperature data and the area data corresponding to the main influence region by using a thermal error mapping algorithm to obtain a thermal error value comprises the following steps: Delimiting a range of a part of the bed of the numerical control machine tool to be monitored to obtain a key position to be monitored; Based on a spatial correlation relationship between the key position and the main influence region, an association between the key position and the temperature data and the area data of each clustering unit in the main influence region is established to obtain a corresponding relationship table of position and data; Based on the corresponding relationship table, a preliminary error value of each key position is calculated by using a thermal error mapping algorithm, and the preliminary error value is corrected to obtain a target error value; The target error values of all key positions are summarized to obtain a thermal error value.

3. The method of claim 2, wherein, The method of calculating a preliminary error value of each key position based on the corresponding relationship table by using a thermal error mapping algorithm comprises the following steps: Obtaining temperature data corresponding to each key position, corresponding area data and spatial characteristic parameters corresponding to the numerical control machine tool bed from the corresponding relationship table; Performing difference calculation on the corresponding temperature data and a preset reference temperature to obtain a temperature deviation value, and performing conversion processing on the corresponding area data according to a preset area weight rule to obtain an area weight value; Performing product calculation on the temperature deviation value, the area weight value and a thermal deformation sensitive coefficient in the spatial characteristic parameters corresponding to the same key position by using a thermal error mapping algorithm to obtain a preliminary error base number; Performing conversion processing on the preliminary error base number corresponding to each key position according to a preset error unit conversion rule to obtain a preliminary error value of each key position.

4. The method of claim 2, wherein, The method of correcting the preliminary error value to obtain a target error value comprises the following steps: The preliminary error value is compared with a preset reasonable value interval to obtain a first comparison result, and a difference value between the preliminary error value and a neighboring preliminary error value is calculated, and the difference value is compared with a preset value change amplitude requirement to obtain a second comparison result; The first comparison result and the second comparison result are combined to obtain an error checking result corresponding to each key position, and the error checking result is classified according to a preset classification rule to obtain a qualified result and an adjustment result; The preliminary error value corresponding to the qualified result is determined as a to-be-confirmed error value, and the preliminary error value corresponding to the adjustment result is adjusted using a preset correction rule to obtain a to-be-confirmed error value; All the to-be-confirmed error values are subjected to uniqueness verification to obtain a target error value corresponding to each key position.

5. The method of claim 1, wherein, The temperature field evaluation data and the visual image data are subjected to consistency verification to obtain a collaborative data set, and a regional clustering algorithm is used to determine a temperature field clustering result of each clustering unit in a region based on the collaborative data set, including: Temperature information of the visual image data is extracted to obtain image temperature information of each region; The image temperature information is subjected to consistency verification processing with an average temperature and a temperature variance of a corresponding region in the temperature field evaluation data to obtain a verification result; The verification result is determined to filter out effective region data, and all the effective region data are summarized to form a collaborative data set; A regional clustering algorithm is used to divide a region in which temperature values in the collaborative data set meet a preset temperature condition and pixel positions are connected into a unit to obtain a plurality of clustering units; All the clustering units and corresponding temperature values and pixel position information are summarized to obtain a temperature field clustering result.

6. The method of claim 1, wherein, The temperature field clustering result is subjected to feature extraction to obtain temperature data and area data of each clustering unit, and based on the temperature data and the area data, a main influence region of thermal deformation of the numerical control machine tool is obtained, including: Temperature features of the temperature field clustering result are extracted to obtain temperature data of each clustering unit, and a region range of each clustering unit in the temperature field clustering result is counted to obtain area data of each clustering unit; The temperature data of all the clustering units are sorted according to the value size to obtain a temperature sequence, and a clustering unit corresponding to a temperature in a leading position in the temperature sequence is selected as a first candidate unit according to a preset temperature proportion; The area data corresponding to all the first candidate units are sorted according to the value size to obtain an area sequence, and a clustering unit corresponding to an area in a leading position in the area sequence is selected as a second candidate unit according to a preset area proportion; The second candidate units are subjected to priority evaluation to obtain a priority level of each second candidate unit, and the priority levels of all the second candidate units are sorted, and based on the sorting result, a space range corresponding to the second candidate unit in a leading position is selected as a main influence region of thermal deformation of the numerical control machine tool.

7. The method of claim 1, wherein, The pixel fusion processing on the infrared thermal imaging data obtains visual image data, including: Pixel information extraction is performed on the infrared thermal imaging data to obtain a temperature value corresponding to each pixel and a position coordinate; According to a preset temperature and display feature corresponding rule, the temperature value is converted to obtain pixel display information of each pixel; The position coordinate and the corresponding pixel display information are fused to obtain pixel fusion information; According to a distribution order of the position coordinate on the bed body of the numerical control machine tool, all the pixel fusion information is arranged to obtain pixel arrangement information; All the pixel display information in the pixel arrangement information is integrated to form visual image data.

8. A thermal error on-line monitoring system for a numerically controlled machine tool, characterized by Comprising: The acquisition module is used for online acquisition of temperature original data of a numerical control machine tool guide rail mounting surface, a workbench support surface and a bed body side surface and infrared thermal imaging data of the numerical control machine tool; The fusion module is used for image segmentation processing on the temperature original data to obtain temperature segmentation data of each region, calculation of average temperature and temperature variance of the temperature segmentation data to obtain temperature field evaluation data of each region, pixel fusion processing on the infrared thermal imaging data to obtain visual image data; The verification module is used for consistency verification of the temperature field evaluation data and the visual image data to obtain a collaborative data set, determination of a temperature field clustering result of each clustering unit in a region based on the collaborative data set by using a region clustering algorithm; The extraction module is used for feature extraction on the temperature field clustering result to obtain temperature data and area data of each clustering unit, and based on the temperature data and the area data, a main influence region of thermal deformation of the numerical control machine tool is obtained; The calculation module is used for associated calculation processing on the temperature data and the area data corresponding to the main influence region by using a thermal error mapping algorithm to obtain a thermal error value.

9. An electronic device, comprising: Comprising: The memory is used for storing a computer program; The processor is used for executing the computer program to realize the steps of the numerical control machine tool thermal error online monitoring method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer program is stored in the computer readable storage medium, and the computer program is executed by the processor to realize the numerical control machine tool thermal error online monitoring method according to any one of claims 1 to 7.

Images