A method and system for precision control of heavy-duty CNC machine tools

Abstract

This invention relates to the field of CNC machine tool precision control technology, and discloses a method and system for precision control of heavy-duty CNC machine tools. The method includes: constructing a unified state sample set and a basic error characterization model; performing spatial error partitioning modeling; performing partition boundary transition compensation processing on a preset machining trajectory; performing adaptive feedback correction based on a first compensation trajectory; correcting and controlling CNC commands during the formal machining process; and outputting a precision control result set. Compared to the existing technology that uses a single position compensation method, especially under the condition of machining large, irregularly shaped workpieces across the entire stroke, it cannot achieve the technical problem of boundary area compensation transition. This application, by constructing a unified state characterization model including spatial position, axis load, and workpiece weight distribution, combined with spatial error partitioning and boundary transition compensation, achieves partitioned, continuous, and dynamic precision control of heavy-duty CNC machine tools under the entire machining domain and multiple working conditions.

Description

Technical Field

[0001] This invention relates to the field of precision control technology for CNC machine tools, and in particular to a precision control method and system for heavy-duty CNC machine tools. Background Technology

[0002] Currently, the precision control capabilities of heavy-duty CNC machine tools in machining large, irregularly shaped workpieces, high-load-bearing workpieces, and complex trajectories spanning the entire machining stroke remain significantly insufficient. Existing technologies mostly employ a globally unified error table compensation method based on pre-calibrated machine tool geometric errors, or establish a single error compensation model solely based on the tool's spatial position. While these methods can provide some compensation for ordinary small to medium-sized workpieces, local machining areas, or conditions with minimal changes in operating conditions, when machining large, irregularly shaped workpieces, the large size, complex clamping methods, and significant shifts in the center of gravity, coupled with the fact that the machining area often spans the entire worktable and multiple coordinate axes, result in inconsistent guideway straightness errors, lead screw cumulative errors, transmission backlash, local structural stiffness differences, and load response characteristics across different spatial regions. This makes it difficult for a unified error model to accurately reflect the true error distribution across the entire machining domain.

[0003] For example, when machining large box-shaped, frame-shaped, or curved irregularly shaped structural parts on heavy-duty gantry milling machines or heavy-duty floor-type milling machines, the driving load and structural stress state of each axis of the machine tool will change significantly when the tool moves to the middle area, edge area, or near the heavy-duty support area of ​​the worktable. Correspondingly, the positioning error and contour error will also show uneven spatial distribution, local abrupt changes, and amplified errors in boundary areas. If the existing unified global error compensation table is still used for correction, it can often only provide coarse compensation for the overall average error, and it is difficult to take into account the continuity of compensation in areas of abrupt local error changes and boundary transition areas. This can easily lead to the problem that compensation is effective in one area but fails in adjacent areas.

[0004] Therefore, there is an urgent need for a heavy-duty CNC machine tool precision control method that can still achieve error modeling oriented towards spatial partitioning, continuous transition compensation for boundary regions, and online correction of compensation trajectory oriented towards dynamic feedback, even under conditions of uneven spatial error distribution across the entire machining domain, significant changes in workpiece weight distribution, and the coexistence of abrupt changes in local errors and boundary transitions. This method aims to improve the positional accuracy, contour accuracy, compensation smoothness, and full-stroke machining consistency of heavy-duty CNC machine tools under complex working conditions. Summary of the Invention

[0005] To address the aforementioned technical shortcomings, the purpose of this invention is to propose a precision control method for heavy-duty CNC machine tools, aiming to solve the technical problem that existing technologies using single-position compensation methods, especially when machining large, irregularly shaped workpieces across the entire stroke, cannot achieve compensation transitions in boundary areas.

[0006] To solve the above-mentioned technical problems, the present invention adopts the following technical solution: The present invention provides a method for precision control of heavy-duty CNC machine tools.

[0007] The precision control method for heavy-duty CNC machine tools includes: Step S10: Obtain the spatial position data, axis load status data, workpiece weight distribution data, and calibration error data of the heavy-duty CNC machine tool in the entire machining domain. Based on the spatial position data, axis load status data, workpiece weight distribution data, and calibration error data, a unified state representation construction mechanism is used to perform the basic error modeling task in the entire machining domain, and output a unified state sample set and a basic error representation model. Step S20: Based on the unified state sample set and the basic error characterization model, a spatial error partitioning modeling task is performed using a partitioning mechanism that couples position, load and weight distribution, and the result is an error partition set, a local error model and a boundary set. Step S30: Based on the error partition set, local error model and boundary set, the partition boundary transition compensation mechanism is used to perform the compensation trajectory preprocessing task and output the first compensation trajectory; Step S40: Based on the first compensation trajectory, an adaptive feedback correction mechanism is used to perform the compensation trajectory update task, and the second compensation trajectory is output; Step S50: Perform CNC instruction correction control during the formal machining process based on the second compensation trajectory, and output the accuracy control result set.

[0008] Preferably, step S10, which involves acquiring spatial position data, axis load state data, workpiece weight distribution data, and calibration error data of the heavy-duty CNC machine tool within the entire machining domain, and performing the basic error modeling task of the entire machining domain using a unified state representation construction mechanism based on the spatial position data, axis load state data, workpiece weight distribution data, and calibration error data, and outputting a unified state sample set and a basic error representation model, specifically includes: Step S101: Obtain the spatial position data, axis load status data, workpiece weight distribution data, and calibration error data of the tool center point of the heavy-duty CNC machine tool within the entire machining domain, and form an original sampling dataset based on the spatial position data, axis load status data, workpiece weight distribution data, and calibration error data; Step S102: Perform unified normalization and vectorization encoding on the original sampled dataset to output a unified state sample set; Step S103: Obtain the expected calibration error data, and use the weighted least squares regression method to perform state and error correlation mapping based on the expected calibration error data and the calibration error data, and output the basic error characterization model.

[0009] Preferably, step S20, which involves performing spatial error partitioning modeling based on a unified state sample set and a fundamental error characterization model using a partitioning mechanism that couples location, load, and weight distribution, and outputting an error partition set, a local error model, and a boundary set, specifically includes: Step S201: Input the unified state sample set into the basic error representation model, and the basic error representation model outputs the basic error distribution results corresponding to each sampling position; Step S202: Based on the basic error distribution results, the sampling locations in the entire processing domain are divided into regions using error gradient constraint clustering to generate an error partition set; Step S203: For the error partition set, local error fitting and modeling processing is performed using partitioned Kriging interpolation to output the local error model. The local error model is used to characterize the local positioning error law caused by changes in spatial position, shaft load, and workpiece weight distribution within the corresponding error partition. The boundary region between adjacent error partitions in the error partition set is extracted as the boundary set. Finally, the error partition set, the local error model, and the boundary set are output.

[0010] Preferably, in step S202, the sampling locations in the entire processing domain are divided into regions using error gradient constraint clustering based on the basic error distribution results to generate an error partition set, specifically including: Step S2021: Based on the basic error distribution results, the spatial difference gradient analysis method is used to extract the error gradient and output the error gradient matrix. The error gradient matrix includes the internal state gradient difference of the partition and the error gradient at the boundary. The internal state gradient difference is used to characterize the consistency of the error change of adjacent sampling positions within the same region, and the error gradient at the boundary is used to characterize the intensity of the error abrupt change at the junction of adjacent regions. Step S2022: Based on the basic error distribution results, perform clustering partitioning according to the priority principle of small internal state gradient difference and large error gradient at the boundary to obtain the initial partitioning results; Step S2023: Verify the connectivity and error mutation continuity of adjacent partitions in the initial partitioning result, and re-merge or split the sampling positions that do not meet the preset connectivity conditions and preset error mutation continuity conditions to generate an error partition set.

[0011] Preferably, step S30, which involves performing a preprocessing task for the compensation trajectory based on the error partition set, the local error model, and the boundary set using a partition boundary transition compensation mechanism, and outputting the first compensation trajectory, specifically includes: Step S301: Obtain the trajectory position data corresponding to each discrete tool position point in the preset machining trajectory, and determine the error partition set or boundary set to which each discrete tool position point belongs based on the error partition set and trajectory position data using a region inclusion determination method based on spatial coordinate index and boundary threshold comparison. Step S302: When the discrete tool point is located inside a single error partition, the local error model corresponding to the error partition is called to calculate the partition compensation amount of the discrete tool point; when the discrete tool point is located within the boundary set, the distance-weighted interpolation method is used to calculate the transition compensation amount based on the positional relationship between the discrete tool point and the two sides of the boundary. Step S303: The partition compensation amount and transition compensation amount corresponding to each discrete tool position point are spliced ​​and smoothed according to the preset machining trajectory, and the first compensation trajectory corresponding to the preset machining trajectory is generated and output.

[0012] Preferably, step S40, which involves performing a compensation trajectory update task based on the first compensation trajectory using an adaptive feedback correction mechanism and outputting the second compensation trajectory, specifically includes: Step S401: Control the heavy-duty CNC machine tool to perform no-load tracking operation according to the first compensation trajectory, and simultaneously collect the actual position feedback data, servo following error data and cutting load response data during the no-load tracking operation; Step S402: Use the extended Kalman filter method to perform residual error calculation and analysis on the actual position feedback data, servo follow error data and cutting load response data with the first compensation trajectory, and output the residual error data corresponding to each discrete tool position. Step S403: Based on the residual error data, the compensation parameters of each discrete tool position in the first compensation trajectory are updated using a recursive Bayesian update method to form the second compensation trajectory and output it.

[0013] Preferably, step S50, which involves correcting CNC commands during the formal machining process based on the second compensation trajectory and outputting a set of accuracy control results, specifically includes: Step S501: During the formal machining process, read the target compensation parameters corresponding to the current tool position in the second compensation trajectory; Step S502: Superimpose the target compensation parameters onto the original trajectory instruction corresponding to the current CNC program to generate the corrected CNC control instruction, and send the corrected CNC control instruction to the machine tool servo execution terminal for machining control; Step S503: After the machining control is completed, collect the actual machining accuracy detection data corresponding to the machining trajectory, and generate a precision control result set including position accuracy results, contour accuracy results and area compensation effect results based on the actual machining accuracy detection data.

[0014] The present invention also provides a precision control system for heavy-duty CNC machine tools, comprising: The basic error modeling module is used to acquire spatial position data, axis load state data, workpiece weight distribution data, and calibration error data of heavy CNC machine tools in the entire machining domain. Based on the spatial position data, axis load state data, workpiece weight distribution data, and calibration error data, it uses a unified state representation construction mechanism to perform basic error modeling tasks in the entire machining domain and outputs a unified state sample set and a basic error representation model. The spatial error partitioning module is used to perform spatial error partitioning modeling tasks based on a unified set of state samples and a basic error characterization model, using a partitioning mechanism that couples position, load and weight distribution. It outputs an error partition set, a local error model and a boundary set. The compensation trajectory preprocessing module is used to perform the compensation trajectory preprocessing task based on the error partition set, local error model and boundary set, and adopts the partition boundary transition compensation mechanism to output the first compensation trajectory. The compensation trajectory update module is used to perform the compensation trajectory update task based on the first compensation trajectory using an adaptive feedback correction mechanism, and output the second compensation trajectory. The instruction correction control module is used to perform CNC instruction correction control during the formal machining process based on the second compensation trajectory, and output a set of accuracy control results.

[0015] The present invention also provides a heavy-duty CNC machine tool precision control device, comprising: a memory, a processor, and a heavy-duty CNC machine tool precision control program stored in the memory and executable on the processor. When the heavy-duty CNC machine tool precision control program is executed by the processor, a heavy-duty CNC machine tool precision control method is implemented.

[0016] The present invention also provides a computer program product, including a heavy-duty CNC machine tool precision control program, which, when executed by a processor, implements the heavy-duty CNC machine tool precision control method.

[0017] The beneficial effects of this invention are as follows: By incorporating spatial position data, shaft load state data, and workpiece weight distribution data into the error modeling process, and further constructing error partition sets, local error models, and boundary sets, the error compensation of heavy-duty CNC machine tools is no longer limited to a globally unified error table. Instead, it can perform partitioned modeling and compensation for error differences in different areas within the entire machining domain. This is especially suitable for application scenarios involving the machining of large, irregularly shaped workpieces across the entire stroke, significant center of gravity offset, and complex load variations. Therefore, it can significantly improve positional accuracy and contour accuracy, and enhance the consistency and stability of machining results.

[0018] This invention introduces boundary transition compensation and adaptive feedback correction mechanisms after the compensation trajectory is generated, enabling the compensation parameters to be dynamically updated by combining the actual position feedback data, servo following error data, and cutting load response data during the no-load tracking operation. This effectively avoids the problems of boundary compensation jumps, local error mismatches, and residual error accumulation that are prone to occur in traditional static compensation methods, and further improves the real-time correction capability, regional compensation smoothness, and accuracy maintenance capability of heavy CNC machine tools in the formal machining process. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention 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 the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a flowchart illustrating the first embodiment of a precision control method for heavy-duty CNC machine tools according to the present invention.

[0021] Figure 2 This is a schematic diagram of the equipment for a precision control method for heavy-duty CNC machine tools according to the present invention. Detailed Implementation

[0022] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0023] Example 1: As Figure 1 The diagram shown is a flowchart of the first embodiment of the precision control method for heavy-duty CNC machine tools of the present invention, which presents the first embodiment of the precision control method for heavy-duty CNC machine tools of the present invention.

[0024] In the first embodiment, the precision control method for the heavy-duty CNC machine tool includes: Step S10: Obtain the spatial position data, axis load status data, workpiece weight distribution data, and calibration error data of the heavy-duty CNC machine tool in the entire machining domain. Based on the spatial position data, axis load status data, workpiece weight distribution data, and calibration error data, a unified state representation construction mechanism is used to perform the basic error modeling task in the entire machining domain, and output a unified state sample set and a basic error representation model. It should be noted that "spatial position data" refers to data that characterizes the spatial position of the tool center point, spindle endpoint, or actual machining point within the entire machining domain of a heavy-duty CNC machine tool. This includes, but is not limited to, X-axis coordinates, Y-axis coordinates, Z-axis coordinates, the trajectory segment number of the tool position point, the spatial interval between adjacent tool position points, the approach state at the end of the stroke, and positional relationship information relative to the center area, edge area, or support sensitive area of ​​the worktable. "Axis load status data" refers to data that reflects the force, drive, and motion impedance changes of each motion axis at its corresponding spatial position. This includes, but is not limited to, servo drive current, drive torque, load rate, feed resistance response, guideway friction change, lead screw load change, and spindle load. The load-bearing state of the box or slide in its current position; "Workpiece weight distribution data" refers to data that can characterize the overall quality characteristics and local off-center load characteristics of the large irregular workpiece to be processed, including but not limited to the total mass of the workpiece, the coordinates of the workpiece's center of gravity, the offset of the workpiece's center of gravity relative to the center of the worktable, the distribution of clamping points, the distribution of support points, the distribution of suspended areas, the proportion of local thick areas, and the resulting local load differences on the worktable; "Calibration error data" refers to data collected by a measuring device at the preset sampling position of the machine tool to characterize the degree of deviation of the actual position of the machine tool from the theoretical position, including but not limited to positioning error, repeatability error, backlash influence, trajectory following deviation, local geometric deviation, and the deviation of the actual machining point from the theoretical tool position.

[0025] Understandably, by collecting information that was originally scattered across spatial location, drive load, workpiece bearing, and error measurement levels, and constructing a unified set of state samples that can directly participate in subsequent modeling and analysis, the input information for machine tool error analysis becomes more complete. Since large, irregularly shaped workpieces faced by heavy-duty CNC machine tools often have obvious areas of concentrated mass, off-center loading, and local support differences, the actual error performance of the machine tool can change even under the same tool position simply due to differences in the workpiece's center of gravity offset direction, clamping method, or load state of each axis. This step incorporates these state variables closely related to error formation into a unified representation system, enabling subsequent error modeling to no longer rely solely on single positional factors but to be based on multi-dimensional information that more closely approximates actual working conditions. The direct technical effect of this is that the subsequent basic error representation model can more realistically reflect the non-uniformity, regional differences, and working condition correlations of error distribution throughout the entire machining domain, providing a more accurate and targeted input basis for subsequent error partitioning, boundary extraction, and compensation trajectory generation. For example, on a heavy-duty gantry CNC machine tool used to process large, irregularly shaped box-shaped parts, the workpiece is long, heavy, and has an asymmetrical internal cavity structure, causing its center of gravity to deviate significantly from the geometric center of the worktable. If traditional methods are used, only tool position coordinates and calibration error data are collected within the entire machining domain according to a fixed grid, resulting in a global error mapping result based on spatial position. This type of result may still have some reference value in the central area of ​​the worktable, where the load is relatively uniform. However, when the tool moves to the edge area near the off-center load side, the actual load state of each axis of the machine tool changes significantly compared to the central area, and the guide rail contact state and structural stress state also change accordingly. At this point, the error model established based on position coordinates may suffer from insufficient or excessive compensation. Conversely, in this step, in addition to collecting the spatial position data of the tool center point within the entire machining domain, the load state data of each axis at the corresponding position, as well as weight distribution data such as the total weight of the workpiece, center of gravity offset, and support point distribution, are simultaneously collected. These data, combined with the calibration error data at the corresponding position, form the original sampling dataset, and a unified state sample set is further constructed. In this way, even if two sampling points are geometrically close, they can be distinguished in the unified state representation as long as they correspond to different shaft load states or different local bearing states. The basic error representation model established in this way can more accurately distinguish the error differences under "same location but different load states".For example, when machining a large frame-like workpiece with unevenly distributed thick-walled stiffeners, a certain area may experience localized mass concentration due to the dense distribution of stiffeners, resulting in a significant increase in force on the corresponding side of the worktable. If a traditional global error table is still used for compensation, the actual positioning accuracy near this area is often unstable. However, in this step, because the weight distribution and shaft load states corresponding to the mass concentration area are pre-included in the sample set, the subsequent basic error characterization model can reflect the error difference between this area and the ordinary area at this stage, providing a more sufficient basis for the error partitioning set division in the subsequent step S20.

[0026] Step S20: Based on the unified state sample set and the basic error characterization model, a spatial error partitioning modeling task is performed using a partitioning mechanism that couples position, load and weight distribution, and the result is an error partition set, a local error model and a boundary set. It should be noted that the "position, load, and weight distribution coupled partitioning mechanism" refers to using the spatial position features, shaft load state features, and workpiece weight distribution features from the unified state sample set output in step S10 as the partitioning basis to jointly analyze the variation law of error distribution within the entire machining domain, thereby forming a spatial error partitioning result that better matches the real error field. Here, the "error partition set" refers to a set of sub-regions with relatively consistent error characteristics, relatively continuous error variation trends within the region, and relatively obvious error differences between regions, obtained by dividing the region based on the basic error distribution results and their variation characteristics; the "local error model" refers to a localized error characterization model corresponding to each error partition, used to describe the relationship between spatial position changes, shaft load changes, workpiece weight distribution changes, and positioning errors within that partition; the "boundary set" refers to the set of boundary regions located between adjacent error partitions, where error gradient changes are significant, and compensation parameters may switch or transition. Furthermore, the "spatial error partitioning modeling task" not only includes determining the partitioning of sampling locations, but also refitting the error distribution patterns within each partition, and separately identifying and preserving error abrupt change regions at the partition edges, so as to provide clear partitioning call criteria and boundary transition criteria for the subsequent compensation trajectory preprocessing stage.

[0027] It should be understood that compared to traditional techniques that typically use regular grids, fixed spatial coordinate segments, or empirical zoning methods for error region division, the improvements in technical effectiveness of this step are mainly reflected in the following aspects. First, traditional spatial division often only considers geometric proximity, assuming that "regions with close proximity have similar errors." However, this assumption is not always valid for heavy-duty CNC machine tools. This is because even if two tool positions are geometrically adjacent, if there are significant differences in the corresponding axis load states, or if the workpiece's center of gravity shifts and local support states differ, their actual positioning errors may also differ. Traditional methods of dividing regions solely based on geometric position easily classify points with fundamentally different errors into the same region, leading to averaging, fuzzification, or even distortion in subsequent compensation models. This step, however, considers spatial position, load state, and weight distribution simultaneously during the zoning stage, expanding the region division basis from "spatial proximity" to "similar states," thus more realistically reflecting the error mechanism of heavy-duty machine tools. Secondly, traditional techniques often lack separate identification of boundary regions, typically involving a direct hard switch between adjacent regions. This can lead to abrupt compensation jumps at the boundaries when different compensation models are subsequently called. This step extracts the boundary regions between adjacent error partitions as a boundary set during the modeling stage. This means that boundaries are no longer considered negligible transition zones but are retained separately as an important basis for subsequent compensation transitions, thus helping to reduce trajectory fluctuations caused by discontinuous model switching. Thirdly, traditional techniques often use a single global model to describe the error distribution of the entire processing domain. This can lower the overall compensation effect when there is "good performance in the central region, distorted errors in the boundary regions, and insufficient fitting in the heavily loaded regions." This step, however, establishes independent local error models for each error partition, which can separately preserve the error evolution patterns within different regions, thereby significantly enhancing the model's ability to express the error characteristics of local regions.

[0028] For example, in a scenario where a heavy-duty gantry CNC machine tool is machining a large, irregularly shaped frame component throughout its entire stroke, if the workpiece is long and has uneven mass distribution at both ends, with a thick-walled reinforced structure on the left end and a hollow, weight-reducing structure on the right end, then even if the tool moves continuously along the X-axis, the table load-bearing state and the drive response of each axis will show significant differences at different positions. If traditional techniques are used, dividing the machining area into several regular grids based solely on the X, Y, and Z coordinates might classify tool points near the thick-walled area on the left end and tool points in the central load-equalizing area as belonging to the same grid region, simply because these points are geometrically close. However, in reality, due to localized mass concentration in the thick-walled area on the left end, the load on that side of the table increases, and the force state of the guide rail and the drive load response change. The corresponding positioning error distribution is often significantly different from that in the central load-equalizing area. In this case, traditional regular grid division causes the subsequent local compensation model to simultaneously fit two different types of data, ultimately resulting in the model only outputting a compromise result, failing to accurately compensate for the thick-walled area or fully compensate for the central area. In this step, because position, axial load, and weight distribution information are considered simultaneously during partitioning, it can be identified that although some tool positions are spatially adjacent, the load state of the left-end thick-walled region is higher and the workpiece center of gravity offset is more obvious. Therefore, the corresponding basic error distribution results differ significantly from those of the central load-equalizing region. Thus, the left-end thick-walled region is classified as a separate error partition, while the central load-equalizing region is classified as another type of error partition. Boundary sets are further extracted at the boundary between the two. For example, when machining large box-shaped parts, the structural stiffness near the opening edge is lower, while the local support stiffness near the densely reinforced rib area is higher. If a global error model is uniformly used, insufficient compensation often occurs near the opening edge, while overcompensation occurs in the densely reinforced rib area. This step, by performing error gradient constraint clustering on the basic error distribution results, can identify the rapidly changing error area near the opening edge as a partition with independent error characteristics and distinguish it from the surrounding relatively stable error areas. In this way, when generating the first compensation trajectory in the subsequent step S30, different local error models can be called according to different error partitions, and transition processing can be performed at the boundary set. This avoids the compensation mismatch problem caused by directly applying the same compensation parameters between regions with significantly different error characteristics, as is the case in traditional techniques. Therefore, this step not only improves the targeting of spatial error modeling but also directly creates conditions for subsequent compensation trajectory generation and smooth boundary transitions, enabling the entire precision control chain to have higher continuity and stability under complex heavy-load conditions.

[0029] Step S30: Based on the error partition set, local error model and boundary set, the partition boundary transition compensation mechanism is used to perform the compensation trajectory preprocessing task and output the first compensation trajectory; It should be noted that the "partition boundary transition compensation mechanism" in this step means that, based on the error partition set, local error model and boundary set obtained in step S20, a single compensation model is no longer uniformly called for the entire preset machining trajectory, nor is a hard switch made directly between adjacent partitions. Instead, it first identifies which error partition and which boundary set each discrete tool position in the preset machining trajectory is located in, and then selects the corresponding local error model or boundary transition compensation method according to the actual ownership of the tool position, so as to generate a compensation trajectory that corresponds one-to-one with the original machining trajectory. "Preset machining trajectory" refers to the theoretical tool motion trajectory generated by the CNC program before the error compensation amount is superimposed, including but not limited to straight trajectory segments, circular arc trajectory segments, spline interpolation trajectory segments, contour finishing trajectory segments, and continuous composite trajectory segments spanning multiple spatial regions; "Discrete tool position points" refers to the set of trajectory points extracted from the preset machining trajectory according to the preset step distance, interpolation cycle, or control sampling cycle for the purpose of facilitating the calculation of compensation amount; "First compensation trajectory" refers to the initial correction trajectory formed after superimposing the partition compensation amount or boundary transition compensation amount on each discrete tool position point in the preset machining trajectory, which is essentially the input basis for the subsequent step S40 to perform adaptive feedback correction.

[0030] Understandably, this step truly transforms the structured spatial error partitioning result obtained in step S20 into a compensation trajectory result that can be applied to the machining path, establishing a direct and continuous correspondence between "error partitioning modeling" and "NC trajectory correction." Through this step, on the one hand, it ensures that tool positions within the same error partition call local error models that match the error characteristics of that partition, thereby improving the targeting of local compensation. On the other hand, for tool positions located within the boundary set, the switching method of "terminating the previous partition model and immediately taking over the next partition model" is no longer used. Instead, boundary transition compensation allows the compensation amount to continuously change within the boundary range, ensuring that the trajectory correction amount is spatially gradual. The resulting technical effect is that the first compensation trajectory not only better matches the real error characteristics of different regions but also maintains good continuity and smoothness at the region boundaries, avoiding tool position jumps, interpolation disturbances, or servo following anomalies caused by sudden changes in compensation amount. Furthermore, since this step generates the first compensation trajectory based on the complete set of error partitions and boundary sets, the subsequent step S40, when performing idle tracking operation and feedback correction, will no longer face a "coarse compensation trajectory that has not been adapted to the region", but an initial compensation trajectory that already has the characteristics of regional difference response and smooth boundary transition. This will significantly improve the convergence efficiency and correction effectiveness of the subsequent dynamic correction stage.

[0031] For example, in a scenario where a heavy-duty gantry CNC machine tool is machining the contour of a large, irregularly shaped sidewall part, the preset machining trajectory needs to gradually move from the central area of ​​the worktable to the off-center loading area near the right edge, and then cross into an adjacent area with a large change in stiffness. If a traditional global uniform compensation method is used, the entire trajectory usually only has one set of average compensation. The compensation in the central area may be acceptable, but when the tool enters the right off-center loading area, the actual error deviates significantly from the characteristics of the central area due to the greater influence of the workpiece's center of gravity shift. The uniform compensation trajectory will then be insufficient. Further into the adjacent area of ​​the opening, due to the abrupt change in structural stiffness, the uniform compensation may result in local overcompensation, ultimately leading to large fluctuations in the accuracy of the entire contour edge. In this step, the discrete tool positions on the preset machining trajectory are first determined according to the error partition set: the tool positions located in the central area belong to the central stable error partition, the tool positions located in the right off-center loading area belong to the off-center loading error partition, and the tool positions located between the off-center loading area and the adjacent area of ​​the opening belong to the boundary set. For tool positions in the central region, the local error model corresponding to the central region is used to calculate the partition compensation amount; for tool positions in the right-side off-load region, the local error model of the off-load region is used to calculate the corresponding compensation amount; for tool positions within the boundary set, the switching is not direct, but the transition compensation amount is calculated based on the positional relationship between the tool position and the two sides of the boundary. In this way, as the tool moves along the trajectory from the center into the off-load region and then into the opening adjacent region, the compensation amount does not change abruptly, but gradually transitions along the boundary range, thus ensuring that the generated first compensation trajectory maintains high continuity throughout the entire movement. For example, when machining the top contour of a large box-shaped part, a certain tool trajectory segment will continuously cross three error partitions: the first partition is the support adjacent area with high stiffness, the second partition is the central load-sharing region, and the third partition is the flexible region near the overhanging edge. Traditional hard-switching methods for partitioning create two compensation abrupt changes at the boundaries of the first and second partitions and the second and third partitions, which can easily cause the tool to generate local trajectory inflection points near the boundaries. However, this step introduces transition compensation at both boundaries and performs sequential smoothing on the entire trajectory. The resulting first compensation trajectory is numerically continuous and gradually changing, and physically, the servo command changes are smoother and the trajectory following is more stable.

[0032] Step S40: Based on the first compensation trajectory, an adaptive feedback correction mechanism is used to perform the compensation trajectory update task, and the second compensation trajectory is output; It should be noted that the "adaptive feedback correction mechanism" in this step refers to the following: based on the first compensation trajectory generated in step S30, the first compensation trajectory is not directly used as the final formal machining trajectory for production. Instead, the heavy-duty CNC machine tool is first controlled to perform no-load tracking operation or low-load verification operation according to the first compensation trajectory. Simultaneously, the actual position feedback data, servo following error data, and cutting load response data of the machine tool during the trajectory execution process are collected. Then, based on the collected feedback information, the compensation parameters corresponding to each discrete tool position in the first compensation trajectory are re-analyzed, re-corrected, and re-updated to form a second compensation trajectory that is closer to the actual dynamic response state of the machine tool. Here, "actual position feedback data" refers to the actual motion position data fed back in real time by the grating ruler, encoder, laser measurement unit, closed-loop feedback link, or other position detection unit during the execution of the first compensation trajectory by the machine tool; "servo following error data" refers to the deviation data formed between the theoretical command position and the actual execution position of the servo system after receiving the control command corresponding to the first compensation trajectory, including but not limited to position following error, speed following error, acceleration and deceleration response deviation, and corner transition hysteresis; "cutting load response data" in this step can be either the equivalent load response data estimated based on historical process parameters under no-load tracking state, or the dynamic load response data reflected by drive current, shaft torque, vibration response, or spindle load change during low-load verification operation.

[0033] It should be understood that, compared to the traditional approach of directly generating compensation trajectories after offline error calibration and then calling them all at once during formal machining, the improvements in technical effectiveness of this step are mainly reflected in the following aspects. First, traditional techniques often assume that the compensation trajectory built offline can be ideally implemented in actual machine tool execution. Therefore, once the compensation trajectory is generated, it is directly applied to formal machining without considering whether the machine tool servo system generates additional following errors, transition lags, or local amplification errors on the trajectory during actual operation. This approach is acceptable for small and medium-sized machine tools, short-stroke, and low-inertia conditions. However, for heavy-duty CNC machine tools, due to the large inertia of each axis, long structural chains, and significant response lags, the theoretical validity of the first compensation trajectory does not necessarily mean that it is equally accurate at the physical execution level. This step, by introducing no-load tracking operation and feedback acquisition processes, verifies the question of "whether the theoretical compensation trajectory can be actually executed" before formal machining, thus significantly reducing the risk of "correct trajectory design but large deviation in execution effect" in traditional techniques. Secondly, traditional techniques typically treat the compensation trajectory as a static lookup result, requiring recalibration only during long-term maintenance. They lack the ability to rapidly adapt and correct for individual machining tasks, single workpiece clamping states, or specific load conditions. This step, however, incorporates actual position feedback data, servo follower error data, and cutting load response data into the compensation parameter update process. This enables dynamic adaptation at the task, condition, and trajectory levels, better addressing dynamic error changes caused by workpiece off-center loading, large trajectory spans, and frequent area switching in the machining of large, irregularly shaped workpieces. Thirdly, even when considering error closed-loop analysis, traditional techniques usually only focus on reducing the overall error mean, lacking fine-grained correction of the error distribution at local discrete points on the trajectory. This often results in a situation where "overall accuracy appears to improve, but errors in key local areas remain excessive." This step, by outputting the residual error data corresponding to each discrete tool position point and updating the compensation parameters accordingly, enables fine-grained correction of key areas, boundary transition sections, and error-sensitive sections. Therefore, the improvement of this step compared to traditional technology is not only reflected in the addition of a feedback process, but also in the upgrading of the original static compensation logic to a trajectory closed-loop optimization logic oriented towards the dynamic response characteristics of heavy machine tools. This has a direct and significant effect on improving the accuracy maintenance capability under complex working conditions.

[0034] For example, in a scenario where a heavy-duty floor-type boring and milling machine is performing long-stroke contour machining on a large, irregularly shaped shell part, step S30 has already generated a first compensation trajectory based on the error partition set. Theoretically, this first compensation trajectory can call the corresponding local error model for different error partitions and complete the compensation transition in the boundary region, so its overall compensation quality is significantly better than the traditional unified error table method. However, in actual trial operation, due to the long extension of the boring bar, the large inertia of the Z-axis slide, and the local drive lag during the long-stroke movement of the X-axis, the machine tool still exhibits a phenomenon where the actual position lags behind the theoretical compensation position in some trajectory segments when executing the first compensation trajectory, especially near the partition boundary and near the direction switch. If we follow the traditional approach, we would usually assume that since the first compensation trajectory is generated based on calibration data and partition model, it can be directly used for formal machining. However, in this case, the aforementioned dynamic following error will still be superimposed on the machining result during formal cutting, leading to local errors exceeding the tolerance at the contour edge. In this step, the machine tool is first controlled to perform no-load tracking operation according to the first compensation trajectory. Simultaneously, the actual position feedback data and servo following error data corresponding to each discrete tool position are recorded. Further, combined with the historical load response characteristics of the workpiece machining task, residual error calculation is performed on the dynamic error of each trajectory segment. Analysis revealed that several discrete tool positions near the positive end of the X-axis exhibit persistent lag deviations, while the boundary transition compensation amount is slightly insufficient at the junction of two error zones. Based on these residual error data, this step further updates the compensation parameters of the corresponding discrete tool positions in the first compensation trajectory to obtain the second compensation trajectory. Thus, during formal machining, the trajectory compensation amount near the end of the long stroke region will be appropriately increased to address the actual lag phenomenon, while the transition section compensation curve at the boundary junction will be further optimized, making the trajectory used in the final formal machining more closely match the actual execution state of the machine tool. For example, when machining large frame-type workpieces on a heavy-duty gantry milling machine, although some cross-region trajectory segments have achieved smooth boundary transitions in step S30, they still exhibit slight oscillations during actual no-load operation due to the servo system's response inertia to continuous small-amplitude compensation commands. If the machining process proceeds directly, these slight oscillations could have a cumulative impact on the surface profile quality. However, the adaptive feedback correction mechanism in this step can absorb and correct these residual errors caused by the servo dynamic response before the actual machining process, making the second compensation trajectory smoother and more in line with the machine tool's dynamic execution capabilities than the first compensation trajectory.

[0035] Step S50: Perform CNC instruction correction control during the formal machining process based on the second compensation trajectory, and output the accuracy control result set.

[0036] It should be noted that the "NC command correction control during formal machining based on the second compensation trajectory" in this step means that after the second compensation trajectory has been generated based on the first compensation trajectory and actual feedback information in step S40, the uncorrected theoretical trajectory command in the original NC program is no longer directly sent to the machine tool servo execution end. Instead, the target compensation parameter corresponding to the current tool position, the current interpolation cycle, or the current trajectory segment in the second compensation trajectory is read, and the target compensation parameter is superimposed on the original trajectory command to form the corrected NC control command. The machine tool control system then drives each coordinate axis, spindle, slide, or worktable to perform formal machining based on this command.

[0037] It should be understood that compared to traditional technologies that typically involve directly machining with the original CNC program or simply loading a set of static compensation tables into the control system for unified correction, the improvements in technical effectiveness of this step are mainly reflected in the following aspects. First, even if an error compensation model has been established in traditional technologies, the compensation table is often simply used as a fixed, long-term parameter of the machine tool, making it difficult to implement targeted CNC command corrections based on a single workpiece clamping state, specific working condition trajectory, or current zone boundary conditions. For heavy-duty CNC machine tools, this fixed, global compensation calling method often cannot adapt to the error variations of large, irregularly shaped workpieces under different regions, postures, and load conditions. This step, however, uses the second compensation trajectory as a direct basis, reading the target compensation parameters point by point or segment by segment during formal machining control and integrating them with the current original trajectory commands, thereby giving the compensation control significant task-specific and region-specific characteristics. Secondly, in traditional techniques, CNC control commands are typically generated only around the theoretical machining path, while compensation is often added uniformly at the machine tool parameter level. This easily leads to a disconnect between the theoretical path and the compensation execution path, especially in machining complex contours across regions, boundaries, and long strokes. This often results in discontinuous trajectories or unstable accuracy at boundary segments after the local compensation amounts are superimposed. This step, by directly generating corrected CNC control commands based on the second compensation trajectory, ensures that the compensation amount and the trajectory itself are within a unified control logic during the generation stage. This helps guarantee the continuity of the formal machining path and the consistency of compensation execution. Thirdly, traditional techniques often limit the evaluation of machining results to overall dimensional deviations or a small amount of detection point data, lacking a systematic output of the compensation effect in different error zones, boundary transition stability, and local contour accuracy differences. This makes it difficult to determine which areas the compensation control is effective in and which areas still have shortcomings. This step outputs a precision control result set after the formal machining is completed, allowing the results of positional accuracy, contour accuracy, regional compensation effect, and boundary stability to be recorded and analyzed separately. This not only helps verify the control effect in the current machining task but also provides a basis for further optimization of unified state representation, error zones, and compensation trajectories in subsequent tasks. Therefore, the improvement of this step compared to traditional technology is not just "using the compensation trajectory for processing", but integrating the compensation result, control command and processing evaluation into a complete execution unit, so that precision control truly forms a verifiable, traceable and iteratively optimized engineering implementation chain.

[0038] Example 2: Furthermore, the present invention provides a heavy-duty CNC machine tool precision control system, which employs a heavy-duty CNC machine tool precision control method from the above embodiments, and can solve a technical problem related to the precision control of heavy-duty CNC machine tools. The beneficial effects of the heavy-duty CNC machine tool precision control system provided by the present invention are the same as those of the heavy-duty CNC machine tool precision control method provided in the above embodiments, and other technical features in the heavy-duty CNC machine tool precision control system are the same as those disclosed in the methods of the above embodiments, and will not be repeated here.

[0039] Example 3: This invention provides a precision control device for heavy-duty CNC machine tools. Please refer to... Figure 2A heavy-duty CNC machine tool precision control device includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, which are executed by the at least one processor to enable the at least one processor to execute the heavy-duty CNC machine tool precision control method of Embodiment 1 above. The heavy-duty CNC machine tool precision control device in this embodiment may include, but is not limited to, mobile terminals such as mobile phones, laptops, digital radio receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Description), PMPs (Portable Media Players), vehicle terminals (e.g., vehicle navigation terminals), and fixed terminals such as digital TVs and desktop computers. This heavy-duty CNC machine tool precision control device is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of this invention. The heavy-duty CNC machine tool precision control device may include a processing device 1001 (e.g., a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory 1002 or a program loaded from a storage device 1003 into a random access memory 1004. Random access memory 1004 also stores various programs and data required for the operation of a heavy-duty CNC machine tool precision control device. Processing unit 1001, read-only memory 1002, and random access memory 1004 are interconnected via bus 1005. I / O interface 1006 is also connected to the bus. Typically, the following systems can be connected to I / O interface 1006: input devices 1007 including, for example, touchscreens, touchpads, keyboards, mice, image sensors, microphones, accelerometers, gyroscopes, etc.; output devices 1008 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 1003 including, for example, magnetic tapes, hard disks, etc.; and communication devices 1009. Communication device 1009 allows a heavy-duty CNC machine tool precision control device to communicate wirelessly or wiredly with other devices to exchange data. Although a heavy-duty CNC machine tool precision control device with various systems is shown in the figure, it should be understood that it is not required to implement or possess all the systems shown. More or fewer systems may be implemented alternatively.

[0040] Example 4: This invention also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the above-described method for precision control of a heavy-duty CNC machine tool. The computer program product provided by this invention can solve the technical problem of precision control of a heavy-duty CNC machine tool. Compared with the prior art, the beneficial effects of the computer program product provided by this invention are the same as the beneficial effects of the precision control method for a heavy-duty CNC machine tool provided in the above embodiments, and will not be repeated here.

[0041] In particular, according to the embodiments disclosed in this invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this invention include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device, or installed from storage device 1003, or installed from read-only memory 1002. When the computer program is executed by processing device 1001, it performs the functions defined in the methods of the embodiments disclosed in this invention.

[0042] It should be understood that the various parts disclosed in this invention can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics may be combined in any suitable manner in one or more embodiments or examples.

[0043] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A method for precision control of heavy-duty CNC machine tools, characterized in that, The methods include: Step S10: Obtain the spatial position data, axis load status data, workpiece weight distribution data, and calibration error data of the heavy-duty CNC machine tool in the entire machining domain. Based on the spatial position data, axis load status data, workpiece weight distribution data, and calibration error data, a unified state representation construction mechanism is used to perform the basic error modeling task in the entire machining domain, and output a unified state sample set and a basic error representation model. Step S20: Based on the unified state sample set and the basic error characterization model, a spatial error partitioning modeling task is performed using a partitioning mechanism that couples position, load and weight distribution, and the result is an error partition set, a local error model and a boundary set. Step S30: Based on the error partition set, local error model and boundary set, the partition boundary transition compensation mechanism is used to perform the compensation trajectory preprocessing task and output the first compensation trajectory; Step S40: Based on the first compensation trajectory, an adaptive feedback correction mechanism is used to perform the compensation trajectory update task, and the second compensation trajectory is output; Step S50: Perform CNC instruction correction control during the formal machining process based on the second compensation trajectory, and output the accuracy control result set.

2. The method for precision control of a heavy-duty CNC machine tool as described in claim 1, characterized in that, Step S10 involves acquiring spatial position data, axis load state data, workpiece weight distribution data, and calibration error data of the heavy-duty CNC machine tool within the entire machining domain. Based on the spatial position data, axis load state data, workpiece weight distribution data, and calibration error data, a unified state representation construction mechanism is used to perform the basic error modeling task for the entire machining domain, outputting a unified state sample set and a basic error representation model. Specifically, this includes: Step S101: Obtain the spatial position data, axis load status data, workpiece weight distribution data, and calibration error data of the tool center point of the heavy-duty CNC machine tool within the entire machining domain, and form an original sampling dataset based on the spatial position data, axis load status data, workpiece weight distribution data, and calibration error data; Step S102: Perform unified normalization and vectorization encoding on the original sampled dataset to output a unified state sample set; Step S103: Obtain the expected calibration error data, and use the weighted least squares regression method to perform state and error correlation mapping based on the expected calibration error data and the calibration error data, and output the basic error characterization model.

3. The method for precision control of a heavy-duty CNC machine tool as described in claim 1, characterized in that, Step S20 involves performing spatial error partitioning modeling based on a unified state sample set and a fundamental error characterization model, using a partitioning mechanism that couples location, load, and weight distribution. The steps to output the error partition set, local error model, and boundary set specifically include: Step S201: Input the unified state sample set into the basic error representation model, and the basic error representation model outputs the basic error distribution results corresponding to each sampling position; Step S202: Based on the basic error distribution results, the sampling locations in the entire processing domain are divided into regions using error gradient constraint clustering to generate an error partition set; Step S203: For the error partition set, local error fitting and modeling processing is performed using partitioned Kriging interpolation to output the local error model. The local error model is used to characterize the local positioning error law caused by changes in spatial position, shaft load, and workpiece weight distribution within the corresponding error partition. The boundary region between adjacent error partitions in the error partition set is extracted as the boundary set. Finally, the error partition set, the local error model, and the boundary set are output.

4. The method for precision control of a heavy-duty CNC machine tool as described in claim 1, characterized in that, In step S202, based on the basic error distribution results, error gradient constrained clustering is used to divide the sampling locations in the entire processing domain into regional clusters, generating an error partition set, specifically including: Step S2021: Based on the basic error distribution results, the spatial difference gradient analysis method is used to extract the error gradient and output the error gradient matrix. The error gradient matrix includes the internal state gradient difference of the partition and the error gradient at the boundary. The internal state gradient difference is used to characterize the consistency of the error change of adjacent sampling positions within the same region, and the error gradient at the boundary is used to characterize the intensity of the error abrupt change at the junction of adjacent regions. Step S2022: Based on the basic error distribution results, perform clustering partitioning according to the priority principle of small internal state gradient difference and large error gradient at the boundary to obtain the initial partitioning results; Step S2023: Verify the connectivity and error mutation continuity of adjacent partitions in the initial partitioning result, and re-merge or split the sampling positions that do not meet the preset connectivity conditions and preset error mutation continuity conditions to generate an error partition set.

5. The precision control method for heavy-duty CNC machine tools as described in claim 1, characterized in that, Step S30, which involves performing a preprocessing task for the compensation trajectory based on the error partition set, the local error model, and the boundary set using a partition boundary transition compensation mechanism, and outputting the first compensation trajectory, specifically includes: Step S301: Obtain the trajectory position data corresponding to each discrete tool position point in the preset machining trajectory, and determine the error partition set or boundary set to which each discrete tool position point belongs based on the error partition set and trajectory position data using a region inclusion determination method based on spatial coordinate index and boundary threshold comparison. Step S302: When the discrete tool point is located inside a single error partition, the local error model corresponding to the error partition is called to calculate the partition compensation amount of the discrete tool point; when the discrete tool point is located within the boundary set, the distance-weighted interpolation method is used to calculate the transition compensation amount based on the positional relationship between the discrete tool point and the two sides of the boundary. Step S303: The partition compensation amount and transition compensation amount corresponding to each discrete tool position point are spliced ​​and smoothed according to the preset machining trajectory, and the first compensation trajectory corresponding to the preset machining trajectory is generated and output.

6. The method for precision control of a heavy-duty CNC machine tool as described in claim 1, characterized in that, Step S40, which involves performing the compensation trajectory update task based on the first compensation trajectory using an adaptive feedback correction mechanism and outputting the second compensation trajectory, specifically includes: Step S401: Control the heavy-duty CNC machine tool to perform no-load tracking operation according to the first compensation trajectory, and simultaneously collect the actual position feedback data, servo following error data and cutting load response data during the no-load tracking operation; Step S402: Use the extended Kalman filter method to perform residual error calculation and analysis on the actual position feedback data, servo follow error data and cutting load response data with the first compensation trajectory, and output the residual error data corresponding to each discrete tool position. Step S403: Based on the residual error data, the compensation parameters of each discrete tool position in the first compensation trajectory are updated using a recursive Bayesian update method to form the second compensation trajectory and output it.

7. The method for precision control of a heavy-duty CNC machine tool as described in claim 1, characterized in that, Step S50, which involves correcting CNC commands during the formal machining process based on the second compensation trajectory and outputting a set of accuracy control results, specifically includes: Step S501: During the formal machining process, read the target compensation parameters corresponding to the current tool position in the second compensation trajectory; Step S502: Superimpose the target compensation parameters onto the original trajectory instruction corresponding to the current CNC program to generate the corrected CNC control instruction, and send the corrected CNC control instruction to the machine tool servo execution terminal for machining control; Step S503: After the machining control is completed, collect the actual machining accuracy detection data corresponding to the machining trajectory, and generate a precision control result set including position accuracy results, contour accuracy results and area compensation effect results based on the actual machining accuracy detection data.

8. A precision control system for a heavy-duty CNC machine tool, applied to the precision control method for a heavy-duty CNC machine tool according to any one of claims 1 to 7, characterized in that, The precision control system for the heavy-duty CNC machine tool includes: The basic error modeling module is used to acquire spatial position data, axis load state data, workpiece weight distribution data, and calibration error data of heavy CNC machine tools in the entire machining domain. Based on the spatial position data, axis load state data, workpiece weight distribution data, and calibration error data, it uses a unified state representation construction mechanism to perform basic error modeling tasks in the entire machining domain and outputs a unified state sample set and a basic error representation model. The spatial error partitioning module is used to perform spatial error partitioning modeling tasks based on a unified set of state samples and a basic error characterization model, using a partitioning mechanism that couples position, load and weight distribution. It outputs an error partition set, a local error model and a boundary set. The compensation trajectory preprocessing module is used to perform the compensation trajectory preprocessing task based on the error partition set, local error model and boundary set, and adopts the partition boundary transition compensation mechanism to output the first compensation trajectory. The compensation trajectory update module is used to perform the compensation trajectory update task based on the first compensation trajectory using an adaptive feedback correction mechanism, and output the second compensation trajectory. The instruction correction control module is used to perform CNC instruction correction control during the formal machining process based on the second compensation trajectory, and output a set of accuracy control results.

9. A precision control device for heavy-duty CNC machine tools, characterized in that, The heavy-duty CNC machine tool precision control device includes: a memory, a processor, and a heavy-duty CNC machine tool precision control program stored in the memory and executable on the processor. When the heavy-duty CNC machine tool precision control program is executed by the processor, it implements a heavy-duty CNC machine tool precision control method according to any one of claims 1 to 7.

10. A computer program product, characterized in that, The computer program product includes a heavy-duty CNC machine tool precision control program, which, when executed by a processor, implements a heavy-duty CNC machine tool precision control method according to any one of claims 1 to 7.

Images