An adaptive compensation processing method for aerospace large-size structural parts

Abstract

The present application relates to the technical field of structural part processing, and particularly relates to an adaptive compensation processing method for aerospace large-size structural parts. The method comprises the following steps: planning in-machine measurement points and paths by a curvature adaptive sampling method; performing measurement to obtain actual points and calculating error values of each point to obtain error field data; analyzing original numerical control machining tool path, extracting cutting part tool points as tool tip points, subtracting a radius from the tool tip points along the tool axis direction to obtain tool core points, and projecting the tool core points to a theoretical surface to obtain tool contact points; estimating error values of any tool contact point by a bilinear interpolation method based on the error field data; performing reverse compensation of the error value of any tool contact point to obtain compensated tool contact points, the tool axis is unchanged, the tool tip points of the cutting part are solved in a reverse process, and an adaptive machining tool path is formed. The present application is helpful to improve the precision and processing efficiency of aerospace large-size structural part processing.

Description

Technical Field

[0001] This invention relates to the field of structural component processing technology, and in particular to an adaptive compensation processing method for large-size aerospace structural components. Background Technology

[0002] In existing aerospace structural component manufacturing processes, these large, thin-walled parts are typically formed using CNC milling, chemical milling, or a combination of machining methods. If machining is performed using a single blank removal method, the gradual release of internal stress during material removal can easily lead to warping or localized collapse of the structural component, especially in areas with high slenderness ratios, dense ribs, or dramatic curvature changes. Such deformation exhibits significant nonlinear superposition characteristics, making it difficult to suppress through simple clamping or path optimization. While multi-clamping and step-by-step machining can disperse stress release to some extent, cumulative errors exist between different clamping references, further amplifying overall contour deviations and making it difficult to meet design requirements for final forming accuracy.

[0003] Furthermore, during machining, the coupling effect of factors such as cutting force, cutting heat, and machine tool dynamic stiffness can easily cause vibration and tool deflection in thin-walled areas, resulting in local wall thickness deviations or wavy errors. In addition, when performing five-axis simultaneous machining on complex curved surfaces, the tool axis posture changes frequently. If the path planning is unreasonable, overcutting or undercutting may occur at locations of abrupt curvature changes, affecting surface quality and aerodynamic shape accuracy. For critical load-bearing components such as central flanges or wing panels, localized insufficient thickness or stress concentration will directly affect structural strength and fatigue life, posing significant safety hazards. Summary of the Invention

[0004] Therefore, it is necessary for the present invention to provide an adaptive compensation machining method for large-size aerospace structural components to solve at least one of the above-mentioned technical problems.

[0005] To achieve the above objectives, an adaptive compensation machining method for large-size aerospace structural components includes the following steps: Step S1: Perform bidirectional curvature analysis on the target machining surface of the input aerospace large-size structural component model, and perform adaptive sampling based on the bidirectional curvature analysis results to plan the on-machine measurement points, thereby generating the measurement path; Step S2: Perform the measurement according to the measurement path to obtain the actual location, and calculate the error value between the on-machine measured location and the actual location to obtain the error field data; Step S3: Analyze the input original CNC machining toolpath, and perform tool axis reverse offset calculation based on the analysis results to determine the tool center point; project the tool center point onto the theoretical surface of the target machining surface to generate the tool contact point and tool contact point normal vector; Step S4: Perform bilinear interpolation based on the error field data to estimate the error value of any tool contact point; Step S5: Calculate the depth of cut by the error value of the reverse compensation tool contact point to obtain the compensation tool contact point; perform the reverse process of tool axis offset calculation based on the compensation tool contact point to solve for the adaptive tool tip point of the cutting part and generate the adaptive machining toolpath.

[0006] The beneficial effects of the technical solution provided by this invention are: This invention obtains the actual machining error of large-size aerospace structural parts through machine measurement, constructs the measurement point distribution based on the curvature adaptive sampling method, and extracts the tool contact information by analyzing the machining toolpath. Combined with the bilinear interpolation algorithm, a continuous error field is established, and the measurement error is mapped to the tool contact position. Under the condition of keeping the tool axis direction unchanged, the tool contact is compensated in the normal direction, thereby generating an adaptive compensation machining toolpath, realizing closed-loop adaptive machining of complex curved surface structural parts.

[0007] This invention effectively solves many technical problems in the machining of large-size aerospace structural components, such as the difficulty in modeling uneven error distribution, the difficulty in directly using on-machine measurement data for toolpath compensation, the inability of traditional compensation methods to adapt to complex surface changes, and the discontinuity of compensation toolpaths. It greatly improves the machining accuracy, on-machine compensation efficiency, and machining stability of complex curved surface structural components, reduces repeated clamping and manual correction processes, and helps to improve the high-precision and high-efficiency machining capabilities of large key aerospace structural components. It is of great significance to the construction of a digital closed-loop machining technology system for high-end aerospace equipment manufacturing. Attached Figure Description

[0008] Other features, objects, and advantages of the invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings: Figure 1 This is a schematic diagram of the steps of the adaptive compensation machining method for large-size aerospace structural components of the present invention; Figure 2 This is a schematic diagram illustrating the principle of the measurement path in one embodiment; Figure 3 This is a schematic diagram illustrating the relationship between the blade contact point and the blade position point in one embodiment; Figure 4 This is a schematic diagram of the adaptive machining toolpath in one embodiment; The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0009] The technical method of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0010] Furthermore, the accompanying drawings are merely illustrative of the invention and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and therefore repeated descriptions of them will be omitted. Some block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor methods and / or microcontroller methods.

[0011] It should be understood that although the terms "first," "second," etc., may be used herein to describe various units, these units should not be limited by these terms. These terms are used merely to distinguish one unit from another. For example, without departing from the scope of the exemplary embodiments, a first unit may be referred to as a second unit, and similarly, a second unit may be referred to as a first unit. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0012] To achieve the above objectives, please refer to Figures 1 to 4 This invention provides an adaptive compensation machining method for large-size aerospace structural components, the method comprising the following steps: Step S1: Perform bidirectional curvature analysis on the target machining surface of the input aerospace large-size structural component model, and perform adaptive sampling based on the bidirectional curvature analysis results to plan the on-machine measurement points, thereby generating the measurement path; In one embodiment of the present invention, a bidirectional curvature analysis is first performed on the marked target machining surface in the model of a large-scale aerospace structural component. Based on the target machining surface, an adaptive sampling algorithm is used to accurately determine the position of the on-machine measurement points according to the local variation characteristics of the surface. Assuming that the target surface is a complex surface, the curvature is first calculated in both the U (main tool path direction) and V (cross tool path direction). If the curvature in the U direction is larger, it indicates that more measurement points are needed in this direction, so the sampling density is increased. Based on the distribution of the on-machine measurement points, a measurement path is divided, including the tool advance / retreat length, retraction length, probe length, and transition length. This measurement path will cover the entire machining surface, ensuring that surface errors at each position can be captured.

[0013] It is worth noting that the construction process of large-scale aerospace structural component models includes the following: First, the geometric data of the structural component needs to be acquired. Typically, this data can be obtained through a scanner or digitized from design drawings and CAD models. In this embodiment, CAD design drawings are chosen to create the initial model. Assuming the target structural component is a complex wing panel, the design drawings provide its dimensions, surface features, and preset machining tolerances. First, CAD software (such as CATIA or SolidWorks) is used to convert the two-dimensional design drawings of the structural component into a three-dimensional model. Based on the surface definition of the target machining surface and the curvature and thickness parameters indicated on the drawings, a smooth surface based on NURBS (Non-Uniform Rational B-Splines) is established to ensure the continuity and smoothness of the surface in all directions. Next, topology optimization is performed on the three-dimensional model to ensure that there are no redundant topological structures or geometric abrupt changes in the model. The surface segments are partitioned and filtered, eliminating unnecessary parts to make the surface of the target machining surface more concise.

[0014] Step S2: Perform the measurement according to the measurement path to obtain the actual location, and calculate the error value between the on-machine measured location and the actual location to obtain the error field data; In a further embodiment, the CNC machine tool begins to execute the measurement task according to the planned measurement path. Specifically, the measurement points start from a preset starting point along the measurement path and are measured sequentially one by one. Assuming the measurement path involves 100 points, on-machine measurement is performed to obtain the actual point positions. The error value of each point is calculated by projecting the Euclidean distance between the actual measured point and the theoretical measured point (i.e., the on-machine measured point) onto the normal of the theoretical point, thus obtaining the error field data. For example, at a measurement point (e.g., X=5.3mm, Y=12.7mm), the difference between the theoretical and actual values ​​is 0.02mm; this value is the error and is recorded as the error value for that point. Through error calculations at all measurement points, a complete error field is constructed, marking the specific error at each point.

[0015] Step S3: Analyze the input original CNC machining toolpath, and perform tool axis reverse offset calculation based on the analysis results to determine the tool center point; project the tool center point onto the theoretical surface of the target machining surface to generate the tool contact point and tool contact point normal vector; In a further embodiment, the original CNC machining toolpath generated based on the CAD model and design requirements refers to a set of instructions represented in G-code (such as ISO standard G-code) or other control languages. This toolpath file contains the path information of the tool performing machining tasks in the workspace, including each tool position point (i.e., tool position), the feed rate, depth of cut, cutting direction, and tool posture for each machining step. The main function of the CNC machining toolpath is to provide instructions to the machine tool control system, ensuring that the tool performs machining according to a predetermined trajectory. Parsing the original CNC machining toolpath involves analyzing these control instructions line by line and extracting relevant data for each machining path segment, such as tool position information (tool position point), tool radius, and tool posture. The tool position points of the cutting portion (assumed to be...) are extracted from the parsing results. Next, using the known tool radius... (Assuming) (10mm), and the tool axis direction vector Ak, to perform tool axis reverse offset calculation. For example, the tool axis direction vector The point is (0,0,1), which is the tip of the knife. If the value is (10, 15, 5), then the center of the knife is... The coordinates can be calculated as Next, place the center of the knife... Projecting the theoretical plane onto the target machining surface generates the tool contact point. And calculate the corresponding knife contact point normal vector. .

[0016] Step S4: Perform bilinear interpolation based on the error field data to estimate the error value of any tool contact point; In a further embodiment, bilinear interpolation is performed based on the obtained error field data to estimate the error value of any tool contact point Lk. Assume that the error values ​​for two adjacent measurement points have already been calculated, respectively... , ,and The parameters are When performing interpolation, the interpolation formula is used: , and conclude that The error value at that point. Continue interpolating in the intersection direction to finally obtain... The error value of a point, that is, the actual error of that point.

[0017] Step S5: Calculate the depth of cut by the error value of the reverse compensation tool contact point to obtain the compensation tool contact point; perform the reverse process of tool axis offset calculation based on the compensation tool contact point to solve for the adaptive tool tip point of the cutting part and generate the adaptive machining toolpath.

[0018] In this embodiment, based on the obtained error value of the tool contact point, the error magnitude of the tool contact point is compensated in reverse to calculate the compensated tool contact point. Assuming The error value is 0.03mm, and the normal vector... If it is (0,0,1), then the compensation tool contact is... The calculation formula is: = The compensated tool contact point obtained through this calculation is offset in the opposite direction of the original tool axis by one tool radius to obtain the adaptive tool tip point of the cutting part. Finally, the various tool tip points are continuously interpolated and connected according to the parameterized order of the original NC program (such as G-code time sequence or path segment numbering), and reconstructed into a continuous toolpath in the CNC system, thus forming the adaptive machining toolpath. This is based on the tool radius. The calculated compensation path accurately corrects errors in the original machining path, generating an adaptive machining toolpath while keeping the tool axis direction and non-cutting parts unchanged.

[0019] Optionally, the bidirectional curvature analysis in step S1 includes: Curvature is extracted along the main tool path and the cross tool path of the target machining surface, and the normal curvature value of the corresponding direction is calculated based on the curvature. In this embodiment, a surface analysis tool (such as the curvature analysis function in Catia or UG) is used to extract the local curvature of the target machined surface. It is assumed that the target machined surface is a large structural component with complex curvature variations, such as a section of a curved surface on an aircraft wing. To extract the curvature, the curvature is extracted along the main tool path of the machined surface (assumed to be...). Direction) and cross-cutting direction (assuming it is) Calculate the curvature in each direction. In terms of direction, assuming the curvature of a certain part of the surface is 0.015, and... With a value of 0.005 in each direction, these curvature values ​​represent the degree of curvature of the surface in two directions. Curvature is used to determine the degree of curvature and smoothness of the machined surface at different locations. Based on this, the curvature in the above directions is used as the basic quantity for the directional projection of the local normal curvature, and mapped onto the normal reference coordinate system to obtain the normal curvature components in the corresponding directions.

[0020] It is worth noting that during the construction of the normal reference coordinate system, a certain point to be analyzed on the target machining surface is used as the reference point. First, the spatial coordinates of this point are extracted from the surface model of the large-scale aerospace structural component, and the first-order deflection of this point is calculated through the surface parametric equations to obtain the following results. Direction tangential vector Su and The tangential direction vector is Sv. Based on the cross product of Su and Sv, the unit normal vector N of this point is determined, and N is used as the z-axis direction of the local coordinate system. Then, Su is normalized and used as the initial x-axis direction, and corrected by orthogonalization of the normal vector to ensure that the x-axis is strictly perpendicular to the normal vector N. Subsequently, the y-axis direction is generated by the cross product of N and the x-axis, thus constructing a local normal reference coordinate system {Xf, Yf, N} that satisfies the right-hand coordinate system rule.

[0021] Using a preset feed influence coefficient as the weight, the normal curvature values ​​of the main feed direction and the cross feed direction are weighted and calculated to obtain the local curvature of the surface; In a further embodiment, it is assumed that The normal curvature in the direction is 0.025. The normal curvature in the direction is 0.010. Then, the normal curvature value is calculated by weighting according to the pre-set feed influence coefficient (assumed to be 0.8). Here, the main feed direction... The normal curvature value is assigned a weight of 0.8, while the cross direction The normal curvature value is then assigned a weight of 0.2, resulting in the weighted local curvature as: Local curvature (principal direction) = Local curvature (intersecting direction) = Then, by merging the weighted curvatures in these two directions, the overall local curvature is obtained as 0.022.

[0022] The target processing surface is divided into several regions to be measured, and the initial sampling density of the regions to be measured is set according to the area of ​​the regions to be measured and the preset measurement resolution. In a further embodiment, the target machined surface is divided into several measurement regions based on its geometric features (such as local curvature, flatness, etc.). The area of ​​each region depends on the range of curvature variation within that region. For example, a relatively flat region has a larger area, while a region with greater curvature variation has a smaller area. The measurement resolution (e.g., 10mm × 10mm) is set during the technical solution design phase and represents the maximum distance between each measurement point. For example, if the set measurement resolution is 10mm, then the maximum distance between each measurement point is 10mm. For each measurement region on the target machined surface, the initial sampling density of that region is first determined based on its area and resolution. For example, if the area of ​​a region is 50mm × 50mm and the measurement resolution is 10mm × 10mm, then the initial number of sampling points for that region will be: =25 points. The initial sampling density can be set based on the number of sampling points obtained.

[0023] The initial sampling density is adaptively adjusted based on the local curvature of the surface corresponding to each area to be measured, thereby generating in-machine measurement points.

[0024] In a further embodiment, the initial sampling density is adaptively adjusted based on the local curvature of the surface in each region to be measured to generate in-machine measurement points. For regions with large curvature changes, the sampling density is increased (e.g., 10mm × 10mm), while for regions with small curvature changes, the sampling density is decreased (e.g., 20mm × 20mm).

[0025] Most importantly, the methods for obtaining the main tool path direction and the cross tool path direction include: Based on the surface features of the target machining surface, determine the geometric features of the target machining surface; perform tool path simulation based on the geometric features to obtain the cutting path simulation data of the target machining surface; In this embodiment, when processing large curved surfaces such as aerospace wings, the target machining surface is first parametrically analyzed, and the surface is discretized into... A grid (e.g., spacing set to 2mm) is used. Then, continuity information is extracted from the surface, including the surface order (e.g., quadratic or cubic NURBS) and the range of local curvature variations (e.g., 0.005~0.03). The geometric feature description set is defined by the normal variation trend and other parameters. This information is input into the toolpath simulation environment, which is typically based on a CAM kernel (such as the NX CAM simulation engine), and the tool diameter (e.g., ...) is set. The ball end mill, feed rate (e.g., 800 mm / min), and step size (e.g., 0.5 mm) are used to generate subsequent cutting trajectory simulation data.

[0026] The cutting trajectory of the target machining surface is determined based on the cutting path simulation data, and the main direction in the cutting trajectory prediction result is taken as the main tool direction. In a further embodiment, after completing the path simulation, a sequence of continuous contact trajectory points of the tool throughout the entire machining process is extracted from the simulation results (e.g., sampling one tool position point every 0.2 mm). These points form a spatial trajectory band, and principal component analysis (PCA) is performed on the trajectory points to obtain the direction with the maximum variance as the main direction. For example, in the machining of an airfoil along its chord, the trajectory principal axis may fluctuate within a range of 30° (±2°) in the X direction. This direction is defined as the main tool feed direction to represent the overall tool feed trend, while filtering out local oscillation trajectories to avoid local curvature disturbances affecting the path direction determination.

[0027] The cross product of the tangential and normal vectors of the target machining surface in the main feed direction is calculated to determine the cross feed direction.

[0028] In a further embodiment, after determining the main tool path direction, the tangential vector of the corresponding point in that direction is taken from a local curved surface of the target machining surface. (e.g., unit vector) and the normal vector of that point (like The cross-path tool travel direction is obtained by calculating the cross product of vectors. The calculation result will be automatically normalized to ensure it is a unit direction vector. For example, we can obtain... This is used to indicate an auxiliary machining direction orthogonal to the main tool feed direction, and is used for subsequent error sampling or path refinement control.

[0029] Optionally, the methods for obtaining the tool feed influence coefficient include: Obtain program instruction data and machine tool execution logs for CNC machining processes; Extract the feed rate command value, path length, and tool axis posture change of each machining path segment from the program instruction data, and simultaneously extract the actual feed rate of the corresponding machining path segment from the machine tool execution log; In this embodiment, G-code instruction streams are exported from the CNC system (such as Siemens 840D) in the machining workshop, while the machine tool operation log is read synchronously (sampling period approximately 10ms). The instruction data records the set feed rate (e.g., F=1200mm / min), path length (e.g., 3~15mm), and tool axis attitude change angle (e.g., A / C axis change ±5°~±20°) for each machining path segment. The execution log records the actual feed rate (e.g., actually only reaching 980 mm / min). The two types of data are aligned one-to-one according to the path segment number to form a comparable data sequence.

[0030] The relative deviation between the actual feed rate and the feed rate command value of each machining path segment is taken as the feed deviation amount, and the actual feed rate is corrected by the feed deviation amount to obtain the effective feed rate. In a further embodiment, the difference between the actual feed rate and the commanded value is normalized. For example, if the command is 1200 mm / min and the actual feed rate is 980 mm / min, the deviation rate is approximately 0.183. This deviation is used as a correction factor to apply inversely to the actual feed rate to obtain an effective feed rate. For example, the weighted value of the two after correction is approximately 1020 mm / min, which is used to reduce the abnormal speed impact caused by instantaneous jitter or load fluctuations, making subsequent time consumption calculations more stable.

[0031] The unit path time consumption is calculated based on the effective feed rate and path length of each machining path segment, and the cutting contact stability of each path segment is evaluated in combination with the tool axis posture change, thereby constructing directional influence characteristic parameters in different directions. In a further embodiment, for each machining path segment, the effective feed rate is divided by the path length (e.g., ...). The unit path time consumption is approximately 0.00098 min. Combined with the tool axis attitude change (e.g., A-axis fluctuation of 8°, C-axis fluctuation of 12°), a stability index is constructed, using a comprehensive score in the form of attitude change / time consumption. For example, a score of 0.25 for a certain segment indicates relatively stable cutting contact. All path segments are categorized and statistically analyzed separately according to the main tool path direction and the intersection direction.

[0032] The directional influence characteristic parameters of the corresponding path segments of the main feed direction and the cross feed direction are statistically summarized to obtain the average action intensity value of the two directions; the feed influence coefficient is assigned based on the average action intensity value of the two directions.

[0033] In a further embodiment, the stability score and average time consumption of the main feed direction path segments and the cross feed direction path segments are summarized respectively. Then, dimensionless processing is performed, using a fusion of stability and time consumption. For example, the stability weight is 0.7, and the time influence weight is 0.3. The intensity of a single segment is then calculated as 0.7 × stability value + 0.3 × (1 - normalized time consumption). For example, if a segment has a stability of 0.8 and a normalized time of 0.6, the intensity is 0.7 × 0.8 + 0.3 × 0.4 = 0.68. The intensity of all path segments in the same direction (main feed direction or cross feed direction) is arithmetically averaged to obtain the average intensity value for that direction. For example, the average value after summing 20 path segments in the main feed direction is 0.73, and in the cross feed direction it is 0.58. After normalization, a weight allocation ratio is obtained, for example, 0.6 for the main feed direction and 0.4 for the cross feed direction, which is used to form the feed influence coefficient.

[0034] Optionally, the measurement path in step S1 includes an infeed section, a probe section, a retraction section, and a transition section; the generated measurement path includes: The machine tool's current tool status is obtained, and the feed segment is generated based on the positional relationship between the tool position and the target machining surface in the current tool status. The feed length is set according to the preset machine tool acceleration constraints. In this embodiment, at the start of actual on-machine measurement, the current tool pose data (e.g., position coordinates X=120.5mm, Y=80.2mm, Z=50.0mm, attitude angles A / C axis ±3°) is read in real time from the CNC system. The spatial distance between this position and the target machining surface is calculated using a normal distance determination model (based on the vector N of the local normal). If the current distance is approximately 15mm, the feed rate and feed length are limited according to the machine tool acceleration constraints (e.g., maximum acceleration 0.8g), generating a smooth feed trajectory so that the tool gradually approaches the target surface at approximately 300mm / min, thereby forming a feed segment with buffering characteristics and avoiding rigid impact.

[0035] It is worth noting that during the generation of the feed segment, the minimum normal distance (detection distance) from the current spatial position of the tool to the target machining surface is used as the initial value, for example, the initial distance is 12mm. This is combined with the machine tool's maximum allowable acceleration (e.g., 0.8g) and the maximum allowable rate of change of impact velocity (e.g., ...). A motion constraint boundary is constructed. Based on this boundary, the tool approach process is segmented into three parts: an acceleration segment, a constant-speed approximation segment, and a deceleration buffer segment. The acceleration segment accounts for approximately 20% (e.g., 2mm) of the total tool approach length, the constant-speed segment accounts for approximately 60% (e.g., 7mm), and the deceleration segment accounts for approximately 20% (e.g., 3mm). This results in a constrained spatial approximation path with a tool approach length of 12mm, allowing the tool to gradually approach the target machining surface without exceeding the dynamic impact threshold, thus ensuring stable entry conditions for subsequent probing stages.

[0036] When the normal distance between the tool and the target machining surface is less than or equal to the preset detection distance, it is determined that the tool has entered the detection section, and the normal projection distance between the tool and the target machining surface is taken as the detection length of the detection section. In a further embodiment, when the normal distance between the tool and the target machining surface is detected by the distance sensor (sampling period of approximately 5ms) to have dropped to a set threshold (e.g., 2mm) during tool operation, the detection phase is determined. At this time, the current tool position is projected onto the normal of the target surface to obtain a projected distance value, for example, 1.6mm. This value is directly used as the length of the detection segment to control the micro-contact action of the probe or the end of the tool, and to keep the feed rate reduced to 50mm / min to ensure measurement stability.

[0037] Once the tool completes the detection of a single on-machine measurement point, it is determined that the tool has entered the retraction section. The tool is controlled to retract away from the target machining surface in the opposite direction of the normal until the retraction length is reached according to the preset safety clearance parameter, so as to enter the transition section to perform path repositioning. In a further embodiment, after contact sampling is completed at a single measurement point, the system immediately switches to the retraction control logic, using the reverse direction of the target surface normal as the retraction direction, and executes reverse displacement control. For example, if the safety clearance parameter is set to 5mm, the tool will retract from the current measurement point along the normal to a position ≥5mm from the target surface, with the retraction speed controlled within 400mm / min to avoid residual vibration affecting the measurement accuracy of the next point.

[0038] The feed section, probe section, retraction section, and transition section corresponding to each on-machine measurement point are connected sequentially to form a measurement path.

[0039] In a further embodiment, the smooth approximation path of the feed segment, the normal micro-motion path of the probe segment, the safe withdrawal path of the retraction segment, and the subsequent transition path for moving to the next measurement point are sequentially connected in time and continuously reconstructed using CNC interpolation (interpolation step size of approximately 0.1 mm). This ultimately forms a continuous measurement trajectory without abrupt changes, ensuring that the tool maintains continuous posture and smooth movement when switching between different measurement points.

[0040] Figure 2 This is a schematic diagram illustrating the principle of the measurement path in one embodiment; such as Figure 2 As shown, the measurement path includes the feed segment, probe segment, retraction segment, and transition segment, corresponding to the feed length, retraction length, probe length, and transition length. In the diagram, the probe length is represented by the orange line segment, which is the normal projection distance between the tool and the target machining surface measurement point after the tool enters the probe segment. The probe segment is triggered when the normal distance between the tool and the target surface drops to a preset threshold. This length is used to control the micro-contact action of the probe or tool tip to ensure measurement stability. The feed length is represented by the solid green line on the right, which is the total displacement length of the tool as it gradually approaches the target inclined plane from the end of the transition segment until it enters the probe segment. It is based on the minimum normal distance from the tool to the target surface, combined with motion constraints such as machine tool acceleration and impact velocity change rate, and is generated in segments to achieve smooth feed without rigid impact. The dashed green line on the right is the normal reference line for the feed length, which is perpendicular to the inclined plane (target machining surface) on the right and is used to define the direction of the feed motion and the feed length. Measurement reference; the retraction length is the solid green line on the left, which is the displacement length of the tool from the measuring point position to the starting point of the transition section along the opposite direction of the plane normal after completing the plane measuring point detection; the tool retraction length is the dashed green line on the left, which is the safe retraction distance reserved after the tool completes the retraction section (which can also be understood as the total retraction distance from the measuring point to the final safe position, including the retraction length), which is set by the preset safety clearance parameter to ensure that there is no risk of collision between the tool and the workpiece when the tool enters the transition section; the transition length is the yellow line segment, which is the spatial displacement length of the tool from the current safe position to the starting position of the tool infeed section of the next measuring point after the tool completes the retraction of a measuring point, along the safe path of the non-machined surface. It is used to realize the path repositioning between different measuring points. With CNC interpolation, it can ensure the motion continuity and posture smoothness of the entire measurement trajectory.

[0041] Optionally, the following steps are included before performing path relocation: The on-machine measurement points are sorted according to the curvature distribution and spatial adjacency of the target machining surface, and a spatial relocation path for the transition segment is generated between adjacent on-machine measurement points. In this embodiment, during the actual execution of the measurement plan, all on-machine measurement points are first written into the same spatial point set, and then combined with the local curvature value of the target machining surface at each point (e.g., 0.003~0.025). The points are prioritized based on their range, and the Euclidean distance between them is calculated using a k-nearest neighbor search structure (k=6) to determine their spatial adjacency. The prioritization rule prioritizes the continuity of curvature changes, followed by ensuring the shortest spatial connection, thus avoiding abrupt jumps in the measurement path. Transitional connection trajectories are then generated between adjacent points, with each transition path typically controlled to a length of 5-15 mm. Cubic spline interpolation is used to smooth the trajectories, ensuring the tool movement meets speed continuity constraints (e.g., maximum speed change ≤ 150 mm / min).

[0042] Once the tool reaches the retraction length, it is determined that the tool has entered the transition section, and the tool is controlled to perform path repositioning according to the spatial repositioning path.

[0043] In a further embodiment, when the tool completes a single-point measurement and retracts to a preset safety clearance position (e.g., 5mm from the workpiece surface), it is determined that it has entered a path transition state, and the generated spatial connection sequence is activated. At this time, the tool no longer performs measurement actions, but moves according to the pre-sorted point links. Continuous path repositioning is performed by the interpolation controller at a transition speed of approximately 300mm / min, keeping the tool posture consistent with the previous measurement direction. This avoids the accumulation of positioning errors caused by sudden changes in direction, thereby completing a smooth switch between different measurement points.

[0044] Optionally, the calculation of the reverse offset of the tool axis in step S3 includes: Extract the tool position point, tool radius, and tool axis direction of the cutting part from the analysis results, where the tool position point of the cutting part is taken as the tool tip point; In this embodiment, when analyzing the original CNC machining toolpath, the cutting segment tool points involved in material removal are selected from the G-code trajectory. For example, spatial point sequences are extracted at a sampling interval of 0.2 mm, and the corresponding tool parameters (such as the ball end mill radius R=5 mm) and tool axis direction vector (calculated from the A / C axis attitude, for example (0.2,0.1,0.97)) are read simultaneously. These cutting segment tool points are directly regarded as the tool tip positions where the tool contacts the workpiece surface, and are used to represent the actual cutting contact geometry.

[0045] The tool tip is offset in the opposite direction along the tool axis by one tool radius to obtain the tool center point.

[0046] In a further embodiment, using the tool tip as a reference, a spatial offset calculation is performed along the tool axis using a reverse unit vector, and the tool's geometric center position is deduced by using a tool radius of 5mm as the offset distance. For example, if the current tool tip is (120.0, 80.0, 50.0), then the tool center point (119.1, 79.6, 45.3) is obtained by moving 5mm in the reverse direction along the tool axis. This point represents the true position of the tool's rotation center in space, providing a unified geometric reference for subsequent tool contact point projection calculations.

[0047] Figure 3 This is a schematic diagram illustrating the relationship between the blade contact point and the blade position point in one embodiment; as shown... Figure 3 As shown, the theoretical surface is a continuous target machining surface obtained by extracting surface segments of the target area from the CAD model of a large-scale aerospace structural component, and reconstructing them through continuity detection and topology fusion. It serves as the geometric reference for machining and on-machine measurement, corresponding to the blue solid line at the bottom of the figure. The toolpath is the tool planning motion trajectory obtained by analyzing the original CNC machining toolpath, corresponding to the blue dashed line parallel to the theoretical surface in the figure. The tool tip point is the tool tip point extracted from the CNC cutting section, which is the reference point for the CNC system to control the tool movement, corresponding to the red point at the bottom of the ball end mill in the figure, located on the toolpath. The tool contact point is the point where the ball end mill's spherical surface actually makes cutting contact with the theoretical machining surface, corresponding to the red point where the tool and the theoretical surface are in contact in the figure. The tool center point is the geometric rotation center of the ball end mill, obtained by offsetting the tool tip point in the opposite direction along the tool axis (vertically upward in the figure) by one tool radius, corresponding to the red point at the center of the ball end mill in the figure, providing a unified geometric reference for subsequent tool contact point projection calculation and machining error analysis.

[0048] Most importantly, the formula for calculating the coordinates of the blade center point is as follows: ; In the formula, Let be the coordinate vector of the knife center point. Let be the position coordinate vector of the knife tip. For the tool radius, This is the tool axis direction vector.

[0049] Optionally, the method for obtaining the theoretical surface of the target machining surface in step S3 includes: The surface topology of the large-scale aerospace structural components is analyzed from the model, and the model is partitioned and filtered according to the preset target machining surface to extract the set of surface segments corresponding to the target machining surface. In this embodiment, when processing the 3D model of aerospace structural components, the surface topology information is first read from the CAD model (such as STEP or IGES format), and the entire model is decomposed into facet units with adjacency relationships. For example, a surface mesh is generated with a spatial sampling accuracy of 0.5mm, and the normal, curvature, and boundary connection relationship of each facet are recorded. Then, spatial filtering is performed according to the preset processing area (for example, a 0.8m×0.3m area on the upper surface of an airfoil), retaining only surface segments that intersect with the target processing area or are less than 2mm away, thus forming an initial set of surface segments.

[0050] The set of surface segments is subjected to continuity detection and boundary consistency constraint processing to eliminate geometric abrupt changes and redundant topological structures at the splicing points of surface segments, thereby obtaining an optimized set of surface segments; In a further embodiment, after obtaining the fragment set, the geometric continuity between adjacent fragments is checked, for example, by calculating the normal angle (threshold set to ≤5°) and the boundary curve deviation (controlled within...). Within a certain range, splicing regions exceeding a threshold are reconstructed and corrected, and small segments with duplicate topology or degenerate boundaries (areas smaller than a certain value) are removed. After this processing, a set of optimized surface segments with good geometric continuity is obtained, making the overall surface transition smoother and more consistent.

[0051] The optimized set of surface segments is topologically fused to serve as the theoretical surface of the target processing surface. The theoretical surface is then reparameterized to establish the surface function description relationship.

[0052] In a further embodiment, these optimized surface segments are fused according to topological adjacency relationships, and a unified parameter domain reconstruction method (e.g., local...) is used. The parameters are uniformly mapped to the global parameter domain (0~1 range), and a continuous surface function expression is generated by cubic NURBS surface reconstruction, so that the entire target machining surface is transformed from a discrete fragment form into a computable continuous theoretical surface model, providing a unified geometric basis for subsequent tool contact projection and error analysis.

[0053] Optionally, the bilinear interpolation performed in step S4 includes: Linear interpolation is performed on the error field data to obtain the grid error field; In this embodiment, when processing error field data, the error points obtained from discrete measurements (e.g., sampling interval of 2~3mm) are first reconstructed according to the uv parameter grid. The missing values ​​are filled in between adjacent measurement points by one-dimensional linear interpolation. For example, the corresponding error value of 0.02 is inserted between adjacent 0.01 and 0.03 parameter positions in the u direction, thereby forming a regularized initial grid error distribution, so that each grid node has basic error information.

[0054] The grid error field is completed according to the distribution density of on-machine measurement points on the target machining surface in the main tool path and the cross tool path; In a further embodiment, the mesh is density-completed based on the actual distribution of the on-machine measurement points. For example, when the spacing between measurement points in the main direction of the tool travel is large (about 5 mm) and the spacing in the intersecting direction is dense (about 2 mm), virtual interpolation nodes are added in the main direction to supplement the original sparse region to a parameter step size of 0.02, which is close to that of the dense region. The error value is corrected by local weighted averaging (neighborhood weight 0.6 / 0.4), so that the mesh error field has consistent spatial resolution in both directions.

[0055] In the completed grid error field, a combined interpolation is performed on any tool contact point in the main tool travel direction and the cross tool travel direction to obtain the error value of each tool contact point.

[0056] In a further embodiment, after completion, for any tool contact point (such as a surface parameter point calculated from the tool path)... First, a linear interpolation is performed on two adjacent grid nodes in the main tool path direction to obtain the intermediate error value in the main direction. Then, using this result as the base node, a linear interpolation calculation is performed again on the two grid lines corresponding to the cross tool path direction to finally obtain the spatial error value corresponding to the tool contact point, thus realizing error estimation based on bidirectional coupling.

[0057] Optionally, performing combined interpolation in the main tool path and the cross tool path includes: In the mesh error field after completion, the error values ​​of adjacent mesh nodes of any tool contact point in the main tool travel direction are interpolated to obtain the error value of the intermediate node along the main tool travel direction. In this embodiment, when actually calculating the tool contact point error, the two adjacent nodes of the tool contact point in the main tool travel direction are first selected in the completed regular mesh, such as the parameter position. =0.32 and =0.34, corresponding to error values ​​of 0.018mm and 0.022mm respectively. By using linear interpolation, an intermediate error value of approximately 0.020mm is generated at the 0.33 position, which is used to characterize the continuous change trend of this point in the main direction.

[0058] The error value of the intermediate node and the node error values ​​on the two grid lines corresponding to the cross-cutting direction in the grid error field after the completion process are used together as the basic node for bilinear interpolation. Linear interpolation is then performed in the cross-cutting direction to obtain the error value of the tool contact in the cross-cutting direction.

[0059] In a further embodiment, this intermediate error value, along with the data of the two corresponding grid lines in the cross-feed direction, is used as the basic input, for example in... Direction selection =0.40 and Two grid lines with a value of 0.42 and node errors of 0.015 mm and 0.019 mm respectively are used. These, together with the intermediate value of 0.020 mm obtained in the previous step, form a local interpolation unit, giving the error in this area a two-dimensional spatial correlation. Finally, a linear interpolation process is performed in the cross-cutting direction, for example, in... =0.41 position, based on the interpolation of 0.015mm and 0.019mm, 0.017mm is obtained. Combined with the aforementioned main direction intermediate value constraint, the final error value of the tool contact point is determined under two-dimensional directional coupling conditions, thereby avoiding the error offset problem caused by single-direction interpolation.

[0060] In another embodiment, performing bilinear interpolation includes: The error field data is interpolated once to form a grid of parameter distribution at the sampling points, and then arbitrary tool contact points are obtained based on bilinear interpolation. error value One of the interpolation mesh completion processes can be performed according to the surface. Use the point with the most directional points to complete the list. place The direction has the most measurement points, with a total of [number]. The parameters of the measurement point are ,in , in turn in the rest Online computing The error value is used to achieve mesh completion, and the formula is: ; in and for Recent front and back values, and This represents the error value for the position.

[0061] In the completed mesh error field, calculate the error value of any tool contact point, assuming the surface parameters of the tool contact point on the theoretical surface are... For the tool contact points within the mesh, first... The difference is calculated using the following formula: ; ; Then based on The two points of interpolation are at Obtain the direction by secondary interpolation The error value at the point of contact, i.e., the error value at the tool contact point, is calculated using the following formula: ; If the tool contact is outside the grid, you need to find the two points closest to the tool contact point. direction or A linear interpolation can be performed in the direction.

[0062] Optionally, the inverse process of performing the tool axis reverse offset calculation in step S5 includes: Under the premise that the tool axis direction remains unchanged, the compensation tool contact point is used as the contact reference point between the tool and the theoretical surface; Using the contact reference point as a reference, the tool as a whole is regarded as a rigid geometry. It is translated equidistantly in the direction of the tool axis according to the radius of a single tool, while maintaining the consistency of the contact normal between the tool and the theoretical surface, so as to obtain the adaptive tool tip point of the corresponding cutting part.

[0063] In this embodiment, the corrected compensation tool contact point is first used as the new contact reference point, for example, the spatial coordinates of this point are (125.2, 78.6, 43.0), and the normal vector N of the processing theoretical surface is recorded simultaneously (e.g., unit vector 0.15, -0.08, 0.98). At this time, the tool axis direction remains unchanged, for example, it is still (0.1, -0.05, 0.99), to ensure that the continuity of the machining posture does not change. Then, the tool is regarded as a rigid geometric model (e.g., a Φ10mm ball end mill or fillet end mill), and the contact reference point is used as the starting point of the overall movement. An equidistant offset is made in the positive direction of the tool axis, and the offset is taken as a tool radius R = 5mm, so that the center of the tool is translated a fixed distance in this direction, without changing the tool posture structure, to ensure that the spatial relationship between the tool and the workpiece only changes in translation. During this process, the angle between the tool axis direction and the normal of the theoretical surface is corrected simultaneously to maintain consistency constraints (e.g., the angle is controlled within ±2°) to avoid the amplification of contact error caused by tool tilting. Finally, through this rigid translation operation, the contact reference point is converted into the corresponding tool tip position, thereby completing the reverse recovery from the error compensation position to the actual machining tool geometry center.

[0064] Figure 4 This is a schematic diagram of the adaptive machining toolpath in one embodiment; as shown. Figure 4As shown, the theoretical surface is represented by the dark blue dashed line at the top, which is the target machining design reference surface extracted from the CAD model of a large-scale aerospace structural component and obtained through topological fusion reconstruction. The actual machining surface is represented by the dark blue solid line in the middle, which is the curved surface actually formed after the workpiece has undergone preliminary machining. Since the machining error is located below the theoretical surface, the normal distance between the two is the machining error value to be compensated. The cutting tool is represented by the blue cylindrical ball end mill at the top, which is the tool for performing cutting and is currently in contact with the actual machining surface. The adaptive machining toolpath (tool contact point) is represented by the red solid line at the bottom, which is the corrected contact reference trajectory obtained after reverse depth of cut compensation of the original theoretical tool contact point based on the error value between the actual machining surface and the theoretical surface. It fits the error distribution of the actual machining surface. Using this as a reference, while keeping the tool axis direction unchanged, the corresponding adaptive tool tip point can be calculated by offsetting it along the positive direction of the tool axis by a tool radius. The finally generated adaptive machining toolpath can accurately correct the error of the original machining path, making the final machining result close to the theoretical surface.

[0065] Therefore, the embodiments should be considered as exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of the equivalents of the application are intended to be included within the invention.

[0066] The above description is merely a specific embodiment of the present invention, enabling those skilled in the art to understand or implement the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features of the invention herein.

Claims

1. An adaptive compensation machining method for large-size aerospace structural components, characterized in that, Includes the following steps: Step S1: Perform bidirectional curvature analysis on the target machining surface of the input aerospace large-size structural component model, and perform adaptive sampling based on the bidirectional curvature analysis results to plan the on-machine measurement points, thereby generating the measurement path; Step S2: Perform the measurement according to the measurement path to obtain the actual location, and calculate the error value between the on-machine measured location and the actual location to obtain the error field data; Step S3: Analyze the input original CNC machining toolpath, and perform tool axis reverse offset calculation based on the analysis results to determine the tool center point; project the tool center point onto the theoretical surface of the target machining surface to generate the tool contact point and tool contact point normal vector; Step S4: Perform bilinear interpolation based on the error field data to estimate the error value of any tool contact point; Step S5: Calculate the depth of cut by the error value of the reverse compensation tool contact point to obtain the compensation tool contact point; perform the reverse process of tool axis offset calculation based on the compensation tool contact point to solve for the adaptive tool tip point of the cutting part and generate the adaptive machining toolpath.

2. The adaptive compensation machining method for large-size aerospace structural components according to claim 1, characterized in that, Step S1, the bidirectional curvature analysis, includes: Curvature is extracted along the main tool path and the cross tool path of the target machining surface, and the normal curvature value of the corresponding direction is calculated based on the curvature. Using a preset feed influence coefficient as the weight, the normal curvature values ​​of the main feed direction and the cross feed direction are weighted and calculated to obtain the local curvature of the surface; The target processing surface is divided into several regions to be measured, and the initial sampling density of the regions to be measured is set according to the area of ​​the regions to be measured and the preset measurement resolution. The initial sampling density is adaptively adjusted based on the local curvature of the surface corresponding to each area to be measured, thereby generating in-machine measurement points.

3. The adaptive compensation machining method for large-size aerospace structural components according to claim 2, characterized in that, Methods for obtaining the feed rate influence coefficient include: Obtain program instruction data and machine tool execution logs for CNC machining processes; Extract the feed rate command value, path length, and tool axis posture change of each machining path segment from the program instruction data, and simultaneously extract the actual feed rate of the corresponding machining path segment from the machine tool execution log; The relative deviation between the actual feed rate and the feed rate command value of each machining path segment is taken as the feed deviation amount, and the actual feed rate is corrected by the feed deviation amount to obtain the effective feed rate. The unit path time consumption is calculated based on the effective feed rate and path length of each machining path segment, and the cutting contact stability of each path segment is evaluated in combination with the tool axis posture change, thereby constructing directional influence characteristic parameters in different directions. The directional influence characteristic parameters of the corresponding path segments of the main feed direction and the cross feed direction are statistically summarized to obtain the average action intensity value of the two directions; the feed influence coefficient is assigned based on the average action intensity value of the two directions.

4. The adaptive compensation machining method for large-size aerospace structural components according to claim 1, characterized in that, The measurement path in step S1 includes the feed section, probe section, retraction section, and transition section; the generated measurement path includes: The machine tool's current tool status is obtained, and the feed segment is generated based on the positional relationship between the tool position and the target machining surface in the current tool status. The feed length is set according to the preset machine tool acceleration constraints. When the normal distance between the tool and the target machining surface is less than or equal to the preset detection distance, it is determined that the tool has entered the detection section, and the normal projection distance between the tool and the target machining surface is taken as the detection length of the detection section. Once the tool completes the detection of a single on-machine measurement point, it is determined that the tool has entered the retraction section. The tool is controlled to retract away from the target machining surface in the opposite direction of the normal until the retraction length is reached according to the preset safety clearance parameter, so as to enter the transition section to perform path repositioning. The feed section, probe section, retraction section, and transition section corresponding to each on-machine measurement point are connected sequentially to form a measurement path.

5. The adaptive compensation machining method for large-size aerospace structural components according to claim 4, characterized in that, Before path relocation, the following also applies: The on-machine measurement points are sorted according to the curvature distribution and spatial adjacency of the target machining surface, and a spatial relocation path for the transition segment is generated between adjacent on-machine measurement points. Once the tool reaches the retraction length, it is determined that the tool has entered the transition section, and the tool is controlled to perform path repositioning according to the spatial repositioning path.

6. The adaptive compensation machining method for large-size aerospace structural components according to claim 1, characterized in that, Step S3, which involves performing the reverse offset calculation of the tool axis, includes: Extract the tool position point, tool radius, and tool axis direction of the cutting part from the analysis results, where the tool position point of the cutting part is taken as the tool tip point; The tool tip is offset in the opposite direction along the tool axis by one tool radius to obtain the tool center point.

7. The adaptive compensation machining method for large-size aerospace structural components according to claim 1, characterized in that, The method for obtaining the theoretical surface of the target machining surface in step S3 includes: The surface topology of the large-scale aerospace structural components is analyzed from the model, and the model is partitioned and filtered according to the preset target machining surface to extract the set of surface segments corresponding to the target machining surface. The set of surface segments is subjected to continuity detection and boundary consistency constraint processing to eliminate geometric abrupt changes and redundant topological structures at the splicing points of surface segments, thereby obtaining an optimized set of surface segments; The optimized set of surface segments is topologically fused to serve as the theoretical surface of the target processing surface. The theoretical surface is then reparameterized to establish the surface function description relationship.

8. The adaptive compensation machining method for large-size aerospace structural components according to claim 1, characterized in that, Step S4 involves bilinear interpolation, which includes: Linear interpolation is performed on the error field data to obtain the grid error field; The grid error field is completed according to the distribution density of on-machine measurement points on the target machining surface in the main tool path and the cross tool path; In the completed grid error field, a combined interpolation is performed on any tool contact point in the main tool travel direction and the cross tool travel direction to obtain the error value of each tool contact point.

9. The adaptive compensation machining method for large-size aerospace structural components according to claim 8, characterized in that, Performing combined interpolation in the main tool path and the cross tool path includes: In the mesh error field after completion, the error values ​​of adjacent mesh nodes of any tool contact point in the main tool travel direction are interpolated to obtain the error value of the intermediate node along the main tool travel direction. The error value of the intermediate node and the node error values ​​on the two grid lines corresponding to the cross-cutting direction in the grid error field after the completion process are used together as the basic node for bilinear interpolation. Linear interpolation is then performed in the cross-cutting direction to obtain the error value of the tool contact in the cross-cutting direction.

10. The adaptive compensation machining method for large-size aerospace structural components according to claim 1, characterized in that, Step S5 involves performing the reverse process of calculating the tool axis reverse offset, which includes: Under the premise that the tool axis direction remains unchanged, the compensation tool contact point is used as the contact reference point between the tool and the theoretical surface; Using the contact reference point as a reference, the tool as a whole is regarded as a rigid geometry. It is translated equidistantly in the direction of the tool axis according to the radius of a single tool, while maintaining the consistency of the contact normal between the tool and the theoretical surface, so as to obtain the adaptive tool tip point of the corresponding cutting part.

Images