Precise structural part trajectory planning method based on cutting path optimization

Abstract

The invention provides a precise structural part trajectory planning method based on cutting path optimization, and the method comprises the steps: constructing a voxelized permissible material removal rate space field according to a three-dimensional model and physical attributes of a to-be-machined workpiece; a growth type algorithm is adopted, under the constraint of a space field of the permissible material removal rate, a motion sequence of a cutter envelope body is generated, and each motion step needs to meet the requirement that the actual material removal rate of a covering voxel is not higher than a permissible value of the corresponding voxel; and outputting a processing track according to the motion sequence. According to the method, the voxelized permissible material removal rate space field is constructed and utilized, constraint growth type path planning is combined, the defect that in the prior art, trajectory planning is disjointed with the physical machining process is effectively overcome, and an advanced solution is provided for high-quality, efficient and stable machining of precise structural parts.

Description

Technical Field

[0001] This invention relates to the field of structural component machining, and in particular to a method for trajectory planning of precision structural components based on cutting path optimization. Background Technology

[0002] As consumer electronic devices such as laptops and tablets become thinner, lighter, and more integrated, the complexity and precision requirements of their internal components continue to increase. Precision metal components must meet dimensional tolerances of ±0.005mm and surface roughness requirements of less than Ra0.8μm. Currently, traditional CNC machining path planning relies heavily on empirical parameters, resulting in problems such as repetitive toolpaths and a high percentage of idle travel (approximately 20%–30%). This not only leads to low machining efficiency but also makes the workpiece prone to deformation due to fluctuations in cutting forces, maintaining a defect rate above 5%. This makes it difficult to meet the planned monthly production capacity of 1,500KPcs for metal parts.

[0003] In actual production, the machining trajectory consistency of different batches of precision structural parts is poor. Affected by factors such as material hardness and tool performance, the machining effect of the same trajectory fluctuates significantly (the surface roughness deviation can reach 0.3μm), leading to poor fit problems in subsequent assembly processes. At the same time, traditional trajectory planning is not adaptable to the machining of complex cavities and thin-walled structures, with long machining time and high tool wear (the life of a single tool is about 800 pieces). There is an urgent need to improve the scientificity and accuracy of trajectory planning through cutting path optimization technology. Summary of the Invention

[0004] In view of the above situation, the main objective of this invention is to propose a precision structural component trajectory planning method based on cutting path optimization in order to solve the above-mentioned technical problems.

[0005] This invention proposes a trajectory planning method for precision structural parts based on cutting path optimization, the method comprising the following steps: Based on the three-dimensional model and physical properties of the workpiece to be processed, a voxelized permissible material removal rate spatial field is constructed. Using a growth algorithm, a motion sequence of the tool envelope is generated under the constraint of the permissible material removal rate space field. Each motion step must satisfy that the actual material removal rate of its covered voxel is not higher than the permissible value of the corresponding voxel. Output the processing trajectory based on the motion sequence.

[0006] The method proposed in this invention effectively overcomes the drawback of the disconnect between trajectory planning and physical processing in the prior art by constructing and utilizing a "voxelated permitted material removal rate space field" and combining it with constrained growth path planning, providing an advanced solution for high-quality, efficient and stable processing of precision structural parts.

[0007] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. Existing trajectory planning technologies mainly rely on workpiece geometry, making it difficult to dynamically and accurately consider physical limits (such as cutting force, vibration, deformation, and chip removal) during the material removal process. This invention constructs a "voxelated permissible material removal rate spatial field" of the target workpiece, quantifying various physical properties such as workpiece material, tool performance, machine tool capability, and workpiece structural stiffness into "permissible removal rate" thresholds for each point in space. This provides precise physical constraint boundaries for path planning, thereby preventing problems such as cutting chatter, tool breakage, overcutting, or excessive deformation caused by excessively high local material removal rates from the source.

[0008] 2. Traditional methods typically use constant cutting parameters (such as feed rate) throughout the entire or a large portion of the path, or make simple adjustments based solely on geometric features, failing to accurately match the changes in the physical load-bearing capacity of the material removal space. This invention employs a "growth algorithm," ensuring at each step, under the constraint of the permissible removal rate space field, that the "actual material removal rate within the tool envelope area does not exceed the permissible value of the corresponding voxel." This allows the generated tool motion sequence to intelligently "sense" and adapt to the material removal capacity of different workpiece regions. Efficiency is improved in high-permissible-value regions, while automatic speed reduction or posture adjustment occurs in low-permissible-value regions (such as thin walls and corners), achieving a uniform and rational distribution of cutting load in time and space.

[0009] 3. Because the machining trajectory generated by this invention is strictly controlled by the permissible removal rate calculated based on physical properties, it can effectively suppress forced vibration and self-excited vibration during machining, reducing tool deformation and surface contour errors caused by cutting force fluctuations. Simultaneously, the uniform cutting load helps maintain stable tool wear, extending tool life and ensuring consistency in the machining process, ultimately achieving higher dimensional accuracy, geometric accuracy, and superior surface integrity.

[0010] 4. This method transforms the complex multiphysics coupling analysis (materials, tools, machine tools, structures) into a voxel space field construction problem, and automatically generates motion sequences that satisfy physical limits through a growth search under constraints. This reduces over-reliance on the experience of process engineers and provides a systematic, computable, and optimizable trajectory generation framework, which is particularly suitable for the automated programming of precision components with complex structures, poor rigidity, and high quality requirements (such as thin-walled aerospace parts, optical molds, etc.).

[0011] 5. By establishing a refined permissible removal rate space field, this invention sets a clear efficiency upper limit (permissible value) for each material removal action. The growth algorithm searches for feasible tool poses within this upper limit, essentially pursuing a local optimum of material removal efficiency within physical constraints. This method can systematically explore the processing potential of the process system while ensuring machining safety and quality, achieving the best balance between efficiency and quality. Compared to traditional methods that rely on conservative empirical parameters, the overall machining efficiency is expected to be improved.

[0012] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by means of embodiments of the invention. Attached Figure Description

[0013] Figure 1 This is a flowchart of the precision structural component trajectory planning method based on cutting path optimization proposed in this invention; Figure 2 This is a quantitative comparison chart of the cutting path idle stroke and repeated cutting of the present invention and the traditional solution; Figure 3 This is a comparison chart showing the correlation between cutting force fluctuation and workpiece deformation in this invention and a traditional solution; Figure 4 This is a comparison chart showing the adaptability evaluation and testing of the complex structure processing trajectory of the present invention and traditional solutions; Figure 5 This is a comparison chart showing the deviation detection of the consistency of multiple batch processing trajectories between the present invention and the traditional solution. Detailed Implementation

[0014] Embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0015] These and other aspects of the embodiments of the present invention will become clear from the following description and accompanying drawings. In these descriptions and drawings, some specific embodiments of the present invention are specifically disclosed to illustrate some ways of implementing the principles of the embodiments of the present invention; however, it should be understood that the scope of the embodiments of the present invention is not limited thereto.

[0016] Please see Figure 1 This embodiment provides a trajectory planning method for precision structural parts based on cutting path optimization. The method includes the following steps: Step 1: Based on the 3D model and physical properties of the workpiece to be processed, construct a voxelized permissible material removal rate spatial field; In a preferred embodiment of the present invention, a voxel-based permissible material removal rate spatial field is constructed based on the three-dimensional model and physical properties of the workpiece to be processed, specifically including the following steps: Obtain the 3D model of the blank and the 3D model of the target workpiece, and perform Boolean difference operation on the 3D model of the blank and the 3D model of the target workpiece to obtain the material body to be removed; Create a three-dimensional containment space that can completely enclose the blank 3D model; An adaptive precision hexahedral voxel partitioning is performed on the three-dimensional bounding space to obtain a blank voxel mesh framework containing voxel space coordinate indices. The adaptive precision hexahedral voxel partitioning rules are as follows: fine voxels (e.g., side length 0.02 mm) are used in the region at a set distance (e.g., 0.5 mm) from the surface of the target workpiece 3D model, and at features where the radius of curvature of the target workpiece 3D model is less than a set threshold (e.g., 2 mm); rough voxels (e.g., side length 0.1 mm) are used in the remaining regions. The positional relationship between each voxel in the blank voxel mesh frame with the blank 3D model, the target workpiece 3D model, and the material to be removed is calculated. Voxels whose center point is located inside the target workpiece 3D model are marked as retained, voxels whose center point is located inside the volume obtained by Boolean difference operation are marked as to be removed, and voxels whose center point is located outside the blank 3D model are marked as empty. This yields a voxelized model containing voxel space mesh coordinates and the initial state attributes of each voxel. Obtain process parameters and use them to assign static attributes to the voxel model to obtain an enhanced voxel model containing attribute values. The permissible instantaneous material removal rate is calculated for each voxel to be removed in the enhanced voxelization model, and the permissible material removal rate spatial field is obtained.

[0017] In this embodiment, by introducing the "adaptive precision hexahedral voxel partitioning" rule (using fine voxels near the surface and high curvature features, and coarse voxels in other areas), this invention effectively solves the problem of massive computational burden caused by high-precision full-field analysis. This scheme can significantly reduce the number of voxels in non-critical areas while ensuring the analysis accuracy of critical areas (such as the final forming surface and small features). This makes it possible to perform full-field physical property analysis and planning for precision structural parts with complex geometries with limited computing resources, avoiding the drawback of being forced to reduce overall accuracy or simplify the model due to excessive computation. Furthermore, by clearly distinguishing between the blank, workpiece, and material to be removed through "Boolean difference operation," and by marking each voxel's state (retained, to be removed, empty) based on the "positional relationship between the voxel center point and the geometry," this method transforms the continuous CAD model into a structured discrete data model. This transformation not only provides a unified and regular spatial index foundation for all subsequent calculations, but also ensures that physical properties (such as permissible removal rates) can be accurately associated with the corresponding material spatial locations, avoiding the problems of mismatch between geometric and physical information or fuzzy boundary treatment in traditional methods, thus improving the robustness and accuracy of the entire planning system.

[0018] In a preferred embodiment of the present invention, process parameters are obtained, and static attributes are assigned to the voxelized model using the process parameters to obtain an enhanced voxelized model containing attribute values. Specifically, the steps include: Process parameters include material grade, tool parameters, and machine tool parameters; For each voxel in the voxelization model marked as to be removed or retained, a static physical property factor is assigned according to the process parameters to obtain an enhanced voxelization model containing property values. The static physical property factors include material machinability factor, tool chip tolerance factor, and structural stiffness factor. The specific operation of assigning static physical property factors according to the process parameters is as follows: Based on the material grade, the built-in material property mapping table is consulted to obtain the target cutting specific energy; based on the machine tool parameters, the recommended cutting specific energy is obtained, and the ratio of the target cutting specific energy to the recommended cutting specific energy is mapped to a dimensionless coefficient to obtain the material machinability factor; Based on the tool parameters, the equivalent volume of the tool chip groove and the curvature of the chip flow channel are calculated and mapped to dimensionless coefficients to obtain the tool chip factor. For all “retained” voxels and their adjacent “to be removed” voxels, the initial stiffness coefficient is calculated based on their distance to the nearest workpiece fixed support point and the moment of inertia of the cross section, thus obtaining the structural stiffness factor.

[0019] In this embodiment, the three core technological knowledge categories—"material grade," "tool parameters," and "machine tool parameters"—are systematically transformed into three dimensionless physical property factors: "material machinability factor," "tool chip tolerance factor," and "structural stiffness factor." This converts qualitative and implicit experience into quantitative and explicit model inputs. It effectively solves the problem that traditional process planning heavily relies on the personal experience of process engineers to select cutting parameters.

[0020] This method generates a "structural stiffness factor" by calculating the distances and moments of inertia of the "retained" voxels and adjacent "to-remove" voxels to the nearest fixed support point. This is the first time that a quantitative assessment of the dynamically changing structural stiffness during workpiece machining has been achieved at the discrete voxel level. This is far more accurate than traditional methods that use the initial stiffness of the complete workpiece or simple geometric rules. It allows the planning system to "sense" in advance how the stiffness of different regions of the workpiece will evolve as material is gradually removed, thus allocating a more conservative (lower) permissible removal rate to weakly stiff regions (such as thin walls far from supports) during the path generation stage, thereby mitigating or reducing the risk of machining deformation at its source.

[0021] As a preferred embodiment of the present invention, the permissible instantaneous material removal rate is calculated for each voxel to be removed in the enhanced voxelization model to obtain the permissible material removal rate spatial field, specifically including the following steps: The baseline removal rate is calculated based on the tool parameters and machine tool parameters; For each voxel marked as to be removed in the enhanced voxelization model, the corresponding material machinability factor, tool chip tolerance factor, and structural stiffness factor are fused by multiplication, and the baseline removal rate is adjusted to obtain the permissible material removal rate for each voxel marked as to be removed; then each voxel marked as empty and marked as retained is marked as zero to obtain the removal rate field. Spatially smooth the removal rate field to obtain the permissible material removal rate spatial field.

[0022] In this implementation, existing technologies, when considering multiple process constraints (such as material hardness, chip removal space, and structural stiffness), typically employ linear weighted summation or the setting of independent safety factors. This essentially treats the constraints as a "parallel" relationship, ignoring the "series" or "coupling" characteristics of these constraints in reality. For example, low structural stiffness (low stiffness factor) in a region directly limits the maximum removal rate that region can withstand even when using free-machining materials and excellent tools. This invention fuses the material machinability factor, tool chip removal factor, and structural stiffness factor in a product manner, accurately capturing the "barrel effect"—the overall permissible level of the system is decisively limited by the weakest link (the smallest factor). The final permissible value generated by this nonlinear fusion method is highly consistent with the formation mechanism of extreme conditions in actual machining, making the planned path physically more realistic and safer. Its prediction accuracy exceeds the effect expected by those skilled in the art based on conventional linear thinking.

[0023] Furthermore, the permissible removal rate field calculated directly based on discrete voxels may exhibit drastic spatial steps (e.g., at the boundary between fine and rough voxels, or in regions of abrupt changes in rigidity). Directly constraining path growth with this original field could lead to frequent and abrupt changes in tool feed rate, causing machine tool vibration and surface quality deterioration. This invention introduces a general "spatial smoothing process" before generating the final spatial field. A "permissible gradient field" is constructed in a physically meaningful way. This gradient field allows the tool to naturally explore along the direction of gently changing permissible values ​​during path growth, automatically avoiding "cliff-like" permissible value boundaries, thus generating a tool motion sequence with continuous and smoothly changing feed rates. The advanced requirements for motion stability in machining dynamics are pre-encoded into the static constraint field, enabling even relatively simple growth algorithms to produce high-quality smooth paths. This method bypasses the traditional approach of performing complex smoothing optimization during the path generation stage, significantly reducing computational complexity.

[0024] Furthermore, this invention first determines a theoretical "baseline removal rate" based on the capabilities of the cutting tool and machine tool. Then, using this as a base, each voxel to be removed carries its unique adjustment coefficient, which is a fusion of local multiple physical properties. This allows the machining trajectory to maintain high efficiency on a macroscopic level (automatically adopting a high value close to the baseline in rigid, easily machinable areas), while automatically and accurately "avoiding" all physical weak points on a microscopic level (such as thin-walled roots, small curvature corners, potential areas of material inclusions, etc.), achieving an adaptive optimal balance between global efficiency and local safety. This ability to dynamically adjust the permissible upper limit at the voxel level produces a refined management effect of "voxel-specific measures," with an optimization level far exceeding that of traditional methods based on setting parameters in geometric regions.

[0025] Step 2: Using a growth algorithm, under the constraint of the permissible material removal rate space field, generate the motion sequence of the tool envelope, wherein each motion step must satisfy that the actual material removal rate of its covered voxel is not higher than the permissible value of the corresponding voxel. In a preferred embodiment of the present invention, a growth algorithm is used to generate the motion sequence of the tool envelope under the constraint of the permissible material removal rate spatial field, specifically including the following steps: Step a: Select the largest continuous voxel cluster to be removed from the permitted material removal rate spatial field, and select the voxel position with the highest permitted material removal rate from the largest continuous voxel cluster to be removed as the initial seed point. Step b: Place the 3D model of the tool at the seed point to generate the tool envelope of the cutting part; Step c: Using the direction adjacent to the voxel to be removed as the candidate direction, try to place new tool envelopes in sequence according to the allowable material removal rate; Step d: Calculate the actual instantaneous material removal rate of all "to be removed" voxels covered by the new tool envelope based on the tool feed rate, radial width, and axial depth of cut of the new tool envelope. Step f: Check each covered voxel to confirm whether the actual instantaneous material removal rate is less than or equal to the allowable material removal rate. If all conditions are met, the tool pose is recorded as a valid path unit. If a certain direction fails to be determined, a local strategy of speed reduction, decomposition cutting, or reversal is triggered until a new valid path unit is generated or the direction is abandoned. Based on the previous effective path unit, repeat steps c to f to obtain the motion sequence of the tool envelope.

[0026] In this embodiment, the complex global trajectory planning problem is transformed into a local greedy growth and real-time verification process guided by a physical constraint field. Compared to traditional path planning, which is mostly based on global derivation of geometric contours, this method uses the permissible removal rate field as a dynamic "soil." Starting from the optimal seed point (high permissible value), like crystal growth, each step explores new poses only in the adjacent voxels of the current tool position according to the order of permissible values. This not only greatly reduces the search space, but more importantly, each step ensures physical feasibility by calculating and verifying the actual instantaneous removal rate. When a certain direction fails, it is not simply abandoned, but a local strategy of deceleration, decomposition, or reversal is triggered for fine-tuning, which is equivalent to allowing the path to intelligently "deform" and pass through when encountering physical "hard constraints." This closed-loop mechanism of "growth-verification-adaptive adjustment" makes the final generated tool motion sequence not only globally approximate the optimal material removal efficiency, but also strictly satisfy multiphysics constraints for each unit at the microscopic level, achieving a unity of macroscopic efficiency and microscopic safety. The reliability and quality stability of its generated path far exceed traditional planning methods based on geometry or fixed parameters.

[0027] Step 3: Output the processing trajectory based on the motion sequence.

[0028] In a preferred embodiment of the present invention, the processing trajectory is output based on the motion sequence, specifically including the following steps: A graph model is constructed by treating each path point in the motion sequence as a node and using the idle travel time between nodes as the edge weight. The ant colony algorithm is used to solve the shortest global access path for the graph model, and the discrete path points in the shortest global access path are connected into an ordered tool movement master sequence. Based on the permissible material removal rate and tool parameters at each path point in the tool movement master sequence, the theoretical optimal feed rate of the tool movement master sequence is calculated in reverse, resulting in a tool movement master sequence with permissible material removal rate and theoretical optimal feed rate. The tool movement master sequence with permitted material removal rate and theoretical optimal feed rate is compiled into standard G-code instructions and inserted into machine tool control instructions to obtain the G-code program; The G-code program was simulated and verified to obtain the machining trajectory.

[0029] In this embodiment, traditional methods typically separate toolpath planning and feed rate setting into two independent stages, which can easily lead to conflicts in optimization objectives. This invention, however, constructs a graph model using path points in the motion sequence as nodes and idle travel time as edge weights, and employs an ant colony algorithm to solve for the shortest global access path. This essentially greatly reduces non-cutting time from a material handling efficiency perspective. Crucially, it then calculates the theoretically optimal feed rate based on the permissible material removal rate and tool parameters at each path point, thus accurately "compiling" the information from the physical constraint field into the final movement command. These two steps, one after the other, first ensure the optimal macroscopic sequence of tool movement, and then ensure the optimal motion parameters at each microscopic position. Through the unified physical field data correlation, they achieve a dual maximization of machining efficiency and process stability at the system level.

[0030] In a preferred embodiment of the present invention, the G-code program is simulated and verified to obtain the machining trajectory, specifically including the following steps: In a virtual environment, the physical simulation of the cutting process is performed according to the G-code program to obtain the cutting force waveform; A simplified finite element model of the workpiece is constructed using voxels marked as retained. The cutting force waveform is applied as a load to the simplified finite element model of the workpiece, and the workpiece deformation cloud map is calculated by linear statics solution. Based on the workpiece deformation cloud map, the key influencing areas are located, and one or more specially designed compensating path units are inserted into the original path sequence to actively counteract or pre-compensate for the upcoming deformation, and the G-code program is regenerated; the regenerated G-code program is simulated and iterated to obtain the final machining trajectory.

[0031] In this embodiment, the traditional "offline verification" simulation process is innovated into an intelligent control closed loop of "online prediction-active compensation," resulting in an unexpected technical effect of moving from passive error detection to active deformation suppression. Traditional simulation is only used to detect geometric errors such as collisions or overcuts, while this method obtains the cutting force waveform through physical simulation and applies it as a dynamic load to a simplified finite element model of the workpiece composed of "retained" voxels, achieving a forward-looking and accurate calculation of the workpiece deformation field during machining. Crucially, it does not stop at prediction but instead uses the deformation cloud map to reverse-locate the cutting action with the greatest impact and dynamically inserts specially designed "compensatory" path units into the original path sequence. Its core purpose is to proactively generate a "pre-compensated" cutting force that can offset or mitigate the deformation before the actual cutting action that causes deformation occurs. This method achieves "feedforward control" of the machining process, giving the toolpath itself the intelligent characteristics of actively suppressing and correcting workpiece deformation.

[0032] In a preferred embodiment of the present invention, based on the workpiece deformation cloud map, the key influencing area is located, and one or more specially designed "compensatory" path units are inserted into the original path sequence to actively offset or pre-compensate for the impending deformation, and the G-code program is regenerated; the regenerated G-code program is iteratively simulated and verified to obtain the final machining trajectory, specifically including the following steps: Find the node with the largest deformation in the workpiece deformation cloud map to obtain the node with the largest deformation. Centered on the node with the maximum deformation, search for all voxels whose Euclidean distance from the node with the maximum deformation is within a set range on the simplified finite element model of the workpiece, and define the region formed by the voxels within the set range as the current deformation influence domain. Obtain the first tool position corresponding to the instantaneous cutting force load that causes the maximum deformation in the current deformation influence domain on the simulation time axis, and the set of voxels to be removed that it covers; Find the position of the tool immediately preceding the position of the first tool in the motion sequence, and use the point between the first tool position and the position of the tool immediately preceding the position as the compensation insertion point. Based on the simplified finite element model of the workpiece and the load direction relative to the position of the first tool, calculate the magnitude and direction of the equivalent compensation force vector required to generate the target reverse pre-deformation in the current deformation influence domain. At the location immediately adjacent to the previous tool position, with the geometric center of the current deformation influence domain as the direction reference, search for an adjacent voxel region with a low allowable removal rate value in the allowable material removal rate space field according to the direction of the required equivalent compensation force vector. Place the tool in the adjacent voxel region with a low allowable removal rate value, and adjust the tool pose so that during cutting, the tool-workpiece interaction force is opposite in direction to the equivalent compensation force vector in the principal component. In the adjacent voxel region with a low allowable removal rate value, a feed rate with an actual material removal rate lower than the allowable material removal rate of the corresponding voxel and a cutting force component sufficient to generate the required equivalent compensation force vector is configured. The tool position in the adjacent voxel region with a low allowable removal rate value and the path point of the feed rate are used as feedforward compensation path units. The feedforward compensation path unit is inserted into the compensation insertion point, and the theoretical optimal feed rate of the first tool position and subsequent path points is adaptively adjusted to obtain the updated motion sequence of the tool envelope. G-code is generated using the updated tool envelope motion sequence, and iterative simulation is performed again until the maximum deformation of the workpiece is lower than the preset threshold, in order to generate the final machining trajectory.

[0033] In this embodiment, traditional compensation methods are mostly based on geometric offset or global speed reduction, while this invention is based on a precise mapping between deformation contour maps and instantaneous cutting forces to reverse-locate the specific tool position and its load that triggers the maximum deformation. Before the "problematic cut," instead of simply modifying the original path, a tiny, specially designed compensatory cutting action is actively "opened" in a region with a low permissible removal rate, guided by the physical field. This action precisely adjusts the tool pose and feed rate so that the resulting cutting force is opposite in direction and controllable in magnitude to the impending "harmful" deformation force in its principal component, forming a "feedforward mechanical cancellation loop." This precise mechanical intervention, tightly coupled to the root cause of the problem in both time and space, is like implanting "immune cells" that actively cancel deformation during machining. With minimal path modification (inserting short paths only in low permissible areas), it achieves targeted suppression of specific deformation modes, with control precision and efficiency far exceeding traditional compensation methods based on experience or overall stiffness models.

[0034] This experiment was designed to systematically verify the comprehensive effectiveness of the present invention in terms of processing efficiency, process stability, geometric adaptability, and process consistency.

[0035] The experimental subject is a comprehensive test piece of 7075 aluminum alloy containing typical precision structural features. These features must include: vertical thin walls of varying thicknesses (0.8 mm, 1.5 mm), deep cavities, bosses, freeform surfaces, and acute-angle features with a radius of curvature less than 2 mm. This design aims to comprehensively challenge the adaptability of trajectory planning to complex geometry and rigidity variations.

[0036] Using the same 3D model and commercially available CAM software commonly used in the industry, experienced process engineers compiled a machining program of "traditional contour roughing + residual milling corner clearing" as a reference.

[0037] The experimental equipment consisted of a five-axis vertical machining center, a spindle power / torque sensor, and non-contact displacement sensors installed on key parts of the workpiece to monitor real-time deformation.

[0038] The main testing areas are quantitative testing of cutting path idle travel and repeated cutting, correlation testing of cutting force fluctuation and workpiece deformation, adaptability evaluation of machining trajectory for complex structures, and deviation testing of consistency of machining trajectories in multiple batches.

[0039] The quantitative detection method for cutting path idle travel and repeated cutting is as follows: The G-code program generated by this invention and the traditional program of the control group are run respectively. The movement trajectory and time of the tool center point in non-cutting states (rapid traverse, approach / retraction) are accurately recorded through the CNC system's log function.

[0040] The quantitative indicators are: Total idle travel length and time percentage: Calculate the percentage of total idle travel length to total path length, and the percentage of idle travel time to total processing time.

[0041] Voxel analysis of repeated cutting areas: After machining, the actual model of the machined workpiece (obtained by scanning) is compared with the target theoretical model using voxelized Boolean methods. The number and spatial distribution of "to be removed" voxels that have been cut more than once are counted.

[0042] The results are shown in Table 1: Table 1:

[0043] pass Figure 2As shown in Table 1, the idle travel time of this invention accounts for 12.5%–25.6%, while that of traditional CAM is 18.3%–45.2%. This is because this invention intelligently connects discrete toolpath units through global path optimization (ant colony algorithm to solve the shortest access path), reducing the idle travel time by an average of about 42%. The advantage is most obvious when machining complex aerospace structural parts (sample 5), where the idle travel time ratio drops from 45.2% to 25.6%, which means that nearly half of the non-cutting time is saved.

[0044] Regarding the proportion of repeatedly cut voxels, the present invention achieves 0.8%–5.2%, while traditional CAM achieves 5.2%–32.7%. This is because the present invention relies on a growth algorithm and real-time removal rate verification to ensure that each voxel to be removed is theoretically removed only once. The proportion of repeated cutting is reduced by an average of 85%, completely solving the problem of repeated cutting of "residual material" caused by fixed step distances in traditional circumferential cutting or equal-height strategies, directly reducing tool wear and ineffective machining.

[0045] The verification results show that this invention optimizes both the macroscopic path sequence and the microscopic material removal, thereby minimizing ineffective working hours and improving machine tool utilization while ensuring the integrity of the processing.

[0046] The method for detecting the correlation between cutting force fluctuation and workpiece deformation is as follows: when machining the thin-walled area of ​​the above-mentioned test piece, the spindle cutting force signal (XYZ three directions) and the dynamic deformation displacement signal of the specified point at the top of the thin wall are collected simultaneously.

[0047] The quantitative indicators are: Cutting force fluctuation coefficient: The ratio of the standard deviation to the mean of the cutting force signal.

[0048] Force-deformation correlation coefficient and phase difference: Calculate the time-domain correlation coefficient between the cutting force component (especially the radial force) and the deformation displacement signal, and analyze its phase relationship.

[0049] Maximum instantaneous deformation: Records the maximum absolute value of deformation at the monitoring points in the thin-walled area throughout the entire processing.

[0050] The results are shown in Table 2: Table 2

[0051] pass Figure 3As shown in Table 2, for the cutting force fluctuation coefficient, the present invention has a coefficient of 0.10~0.22, while traditional CAM has a coefficient of 0.25~0.48. This is because the allowable removal rate spatial field constraint limits the load on each voxel of the tool to within a safe threshold, resulting in an extremely stable cutting force signal. The fluctuation coefficient is reduced by an average of 61%, indicating that the machining process has changed from "impactful" to "stable," fundamentally avoiding the risk of chatter. For the maximum instantaneous deformation, the present invention has a deformation of 3.2~15.2 μm, while traditional CAM has a deformation of 15.8~42.5 μm. This data shows that in thin-walled and other weakly rigid regions, the deformation of the present invention is reduced by an average of 68%. This is because the structural stiffness factor has been assigned a low allowable value for weak regions during the planning stage, and a feedforward compensation mechanism has been designed subsequently. The deformation in the "compensation region" in the chart is only 3.2 μm, which intuitively demonstrates that the inserted compensation path unit actively cancels out most of the deformation.

[0052] Regarding the force-deformation correlation coefficient, the value in this invention is 0.35~0.70, while that in traditional CAM is 0.72~0.91. The lower the coefficient, the greater the influence of active control factors other than cutting force on the deformation. The significant reduction in the correlation coefficient of this invention proves that the feedforward compensation path generates a "corrective force" that is opposite to the cutting force, breaking the inherent strong correlation between deformation and cutting force, and realizing active control.

[0053] The verification results show that this invention has achieved a leap from "passive bearing" to "active prediction and compensation", reducing the instability of processing dynamics to an extremely low level, and is particularly suitable for processing easily deformable precision parts.

[0054] The adaptive evaluation and testing method for complex structure processing trajectory is as follows: after the sample processing is completed, the actual contour point cloud data of all feature areas (thin-walled side, deep cavity bottom and sidewall, curved surface, acute angle) are obtained using a CMM or blue light scanner.

[0055] The quantitative indicators are: Overall profile error (RMS value): Calculates the root mean square error between all measured points and the theoretical model.

[0056] Key feature dimensional accuracy: Measure the actual thickness of the thin wall, the depth and width of the deep cavity, the radius of the acute angle, etc., and compare them with the design values.

[0057] Surface morphology and vibration analysis: The freeform surface region was scanned with a white light interferometer to analyze the surface roughness (Sa) and the presence of periodic vibrations.

[0058] The results are shown in Table 3: Table 3:

[0059] pass Figure 4 As shown in Table 3, for the contour error RMS, the present invention has an RMS of 10.8–20.3 μm, while traditional CAM has an RMS of 25.4–50.7 μm. This data demonstrates that the present invention improves contour accuracy by an average of 58% across all features. This is attributed to the adaptive voxel partitioning providing a foundation for fine analysis in high-curvature regions, and the physical property fusion calculation automatically reducing the permissible removal rate in weak areas (such as thin walls and sharp angles), thus automatically employing a more conservative and precise strategy when generating paths.

[0060] For the surface roughness Sa, the present invention achieves 0.78~1.15 μm, while traditional CAM achieves 1.42~2.85 μm. This data shows that the present invention improves surface quality by an average of 53%. Stable cutting force is a prerequisite for obtaining a good surface. Simultaneously, the growth algorithm explores along the smoothed permissible field, naturally generating a toolpath with continuously varying feed rates, avoiding chatter marks caused by sudden speed changes.

[0061] Regarding the vibration ripple amplitude, the present invention achieves 0.9~2.8 μm, while traditional CAM achieves 3.5~9.6 μm. This data shows that the vibration ripple amplitude is reduced by approximately 70%, which is the direct cause of the improved surface roughness. This comprehensively demonstrates the effectiveness of the present invention in both suppressing vibration (through smooth cutting forces) and ensuring motion smoothness (through path smoothing).

[0062] The verification results show that this invention, through refined spatial physical field modeling, enables the processing trajectory to be "adapted to local conditions," ensuring the processing quality of various features simultaneously on workpieces with high geometric complexity and large rigidity differences.

[0063] The deviation detection method for consistency of machining trajectories in multiple batches is as follows: Without changing any input parameters (3D model, process parameter library), the software system of this invention is used to restart and independently run three complete trajectory planning processes, generating three sets of G-code programs (A, B, C). On the same machine tool, using the same new tool, three identical test blanks are machined using these three sets of programs respectively.

[0064] The quantitative indicators are: Consistency of planning results: Compare the total processing time, total path length, and key steps (such as seed point selection and compensation path insertion points) of the three programs to see if they are the same.

[0065] Process consistency: Record the peak value and fluctuation of cutting force at the same characteristic position (such as the same position when cutting thin wall for the first time) in three batches of processing.

[0066] Consistency of results: The contour error of three batches of finished products at the same feature point was measured, and the mean and standard deviation were calculated.

[0067] The results are shown in Table 4: Table 4:

[0068] pass Figure 5 As shown in Table 4, for the portion of total processing time, the coefficient of variation for this invention is 0.35%, while that for traditional CAM is 4.73%. This data demonstrates that the extremely low time fluctuation (<0.5%) proves the high determinism of the method in this invention. The planning results do not depend on random initial values ​​or human intervention, ensuring precise and controllable production cycle time.

[0069] For the critical dimension (thickness) error portion, the coefficient of variation of this invention is 4.27%, while the coefficient of variation of traditional CAM is 11.65%. This data shows that the significant improvement in the consistency of critical dimensions of workpieces using the method of this invention stems from the consistency of the machining process (cutting force, deformation). This ensures that the dimensions of all parts in mass production remain stable within a very small deviation band.

[0070] For the peak cutting force, the coefficient of variation of this invention is 0.79%, while that of traditional CAM is 5.21%. The consistency of the machining process signal directly reflects the high repeatability of tool load conditions, which is crucial for ensuring predictable tool life and stable process windows. This data shows that the present invention is significantly lower than traditional CAM, thus demonstrating superior process stability.

[0071] Regarding seed point selection consistency, the coefficient of variation in this invention is 100%, while that of traditional CAM is 40%. This metric directly verifies the robustness of the growth algorithm logic in this invention. This invention starts each planning process from the same physically optimal location, while traditional methods or methods based on stochastic optimization exhibit greater uncertainty.

[0072] The verification results show that this invention solidifies process knowledge into algorithms and models, eliminating the experience-based adjustments that vary from person to person in traditional methods, and realizing the "digital solidification" of processing technology, providing a highly reliable technical foundation for digital production and intelligent manufacturing.

[0073] It should be understood that although the steps in the flowcharts of the various embodiments of the present invention are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the various embodiments may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least a portion of the sub-steps or stages of other steps.

[0074] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0075] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0076] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention. Therefore, the scope of protection of this patent should be determined by the appended claims.

Claims

1. A method for trajectory planning of precision structural parts based on cutting path optimization, characterized in that, The method includes the following steps: Based on the three-dimensional model and physical properties of the workpiece to be processed, a voxelized permissible material removal rate spatial field is constructed. Using a growth algorithm, a motion sequence of the tool envelope is generated under the constraint of the permissible material removal rate space field. Each motion step must satisfy that the actual material removal rate of its covered voxel is not higher than the permissible value of the corresponding voxel. Output the processing trajectory based on the motion sequence.

2. The precision structural component trajectory planning method based on cutting path optimization according to claim 1, characterized in that, Based on the 3D model and physical properties of the workpiece to be processed, a voxelized permissible material removal rate spatial field is constructed, which includes the following steps: Obtain the 3D model of the blank and the 3D model of the target workpiece, and perform Boolean difference operation on the 3D model of the blank and the 3D model of the target workpiece to obtain the material body to be removed; Create a three-dimensional containment space that can completely enclose the blank 3D model; The three-dimensional bounding space is divided into hexahedral voxels with adaptive precision to obtain a blank voxel mesh framework containing voxel space coordinate indices. The adaptive precision hexahedral voxel division rule is: fine voxels are used in the region at a set distance from the surface of the target workpiece three-dimensional model and at the features of the target workpiece three-dimensional model with a radius of curvature less than a set threshold. The remaining areas are roughness voxels; The positional relationship between each voxel in the blank voxel mesh frame with the blank 3D model, the target workpiece 3D model, and the material to be removed is calculated. Voxels whose center point is located inside the target workpiece 3D model are marked as retained, voxels whose center point is located inside the volume obtained by Boolean difference operation are marked as to be removed, and voxels whose center point is located outside the blank 3D model are marked as empty. This yields a voxelized model containing voxel space mesh coordinates and the initial state attributes of each voxel. Obtain process parameters and use them to assign static attributes to the voxel model to obtain an enhanced voxel model containing attribute values. The permissible instantaneous material removal rate is calculated for each voxel to be removed in the enhanced voxelization model, and the permissible material removal rate spatial field is obtained.

3. The precision structural component trajectory planning method based on cutting path optimization according to claim 2, characterized in that, Obtain process parameters, and use these parameters to assign static attributes to the voxelized model to obtain an enhanced voxelized model containing attribute values. The specific steps include: Process parameters include material grade, tool parameters, and machine tool parameters; For each voxel in the voxelization model marked as to be removed or retained, a static physical property factor is assigned according to the process parameters to obtain an enhanced voxelization model containing property values. The static physical property factors include material machinability factor, tool chip tolerance factor, and structural stiffness factor. The specific operation of assigning static physical property factors according to the process parameters is as follows: Based on the material grade, the built-in material property mapping table is consulted to obtain the target cutting specific energy; based on the machine tool parameters, the recommended cutting specific energy is obtained, and the ratio of the target cutting specific energy to the recommended cutting specific energy is mapped to a dimensionless coefficient to obtain the material machinability factor; Based on the tool parameters, the equivalent volume of the tool chip groove and the curvature of the chip flow channel are calculated and mapped to dimensionless coefficients to obtain the tool chip factor. For all "retained" voxels and their adjacent "to be removed" voxels, the initial stiffness coefficient is calculated based on their distance to the nearest workpiece fixed support point and the moment of inertia of the cross section, thus obtaining the structural stiffness factor.

4. The precision structural component trajectory planning method based on cutting path optimization according to claim 3, characterized in that, For each voxel to be removed in the enhanced voxelization model, the permissible instantaneous material removal rate is calculated to obtain the permissible material removal rate spatial field. The specific steps include the following: The baseline removal rate is calculated based on the tool parameters and machine tool parameters; For each voxel marked as to be removed in the enhanced voxelization model, the corresponding material machinability factor, tool chip tolerance factor, and structural stiffness factor are fused by multiplication, and the baseline removal rate is adjusted to obtain the permissible material removal rate for each voxel marked as to be removed; then each voxel marked as empty and marked as retained is marked as zero to obtain the removal rate field. Spatially smooth the removal rate field to obtain the permissible material removal rate spatial field.

5. The precision structural component trajectory planning method based on cutting path optimization according to claim 4, characterized in that, A growth algorithm is used to generate the motion sequence of the tool envelope under the constraint of the allowable material removal rate space field. The specific steps include the following: Step a: Select the largest continuous voxel cluster to be removed from the permitted material removal rate spatial field, and select the voxel position with the highest permitted material removal rate from the largest continuous voxel cluster to be removed as the initial seed point. Step b: Place the 3D model of the tool at the seed point to generate the tool envelope of the cutting part; Step c: Using the direction adjacent to the voxel to be removed as the candidate direction, try to place new tool envelopes in sequence according to the allowable material removal rate; Step d: Calculate the actual instantaneous material removal rate of all "to be removed" voxels covered by the new tool envelope based on the tool feed rate, radial width, and axial depth of cut of the new tool envelope. Step f: Check each covered voxel to confirm whether the actual instantaneous material removal rate is less than or equal to the allowable material removal rate. If all conditions are met, the tool pose is recorded as a valid path unit. If a certain direction fails to be determined, a local strategy of speed reduction, decomposition cutting, or reversal is triggered until a new valid path unit is generated or the direction is abandoned. Based on the previous effective path unit, repeat steps c to f to obtain the motion sequence of the tool envelope.

6. The precision structural component trajectory planning method based on cutting path optimization according to claim 5, characterized in that, Based on the motion sequence, the processing trajectory is output, which includes the following steps: A graph model is constructed by treating each path point in the motion sequence as a node and using the idle travel time between nodes as the edge weight. The ant colony algorithm is used to solve the shortest global access path for the graph model, and the discrete path points in the shortest global access path are connected into an ordered tool movement master sequence. Based on the permissible material removal rate and tool parameters at each path point in the tool movement master sequence, the theoretical optimal feed rate of the tool movement master sequence is calculated in reverse, resulting in a tool movement master sequence with permissible material removal rate and theoretical optimal feed rate. The tool movement master sequence with permitted material removal rate and theoretical optimal feed rate is compiled into standard G-code instructions and inserted into machine tool control instructions to obtain the G-code program; The G-code program was simulated and verified to obtain the machining trajectory.

7. The precision structural component trajectory planning method based on cutting path optimization according to claim 6, characterized in that, The G-code program is simulated and verified to obtain the machining trajectory. The specific steps include the following: In a virtual environment, the physical simulation of the cutting process is performed according to the G-code program to obtain the cutting force waveform; A simplified finite element model of the workpiece is constructed using voxels marked as retained. The cutting force waveform is applied as a load to the simplified finite element model of the workpiece, and the workpiece deformation cloud map is calculated by linear statics solution. Based on the workpiece deformation cloud map, locate the key influencing areas, and insert one or more specially designed compensating path units into the original path sequence to actively offset or pre-compensate the upcoming deformation, and regenerate the G-code program. The regenerated G-code program is simulated and iterated to obtain the final machining trajectory.

8. The precision structural component trajectory planning method based on cutting path optimization according to claim 7, characterized in that, Based on the workpiece deformation cloud map, the key influencing areas are located, and one or more specially designed "compensatory" path units are inserted into the original path sequence to actively counteract or pre-compensate for impending deformation. The G-code program is then regenerated. Iterative simulation verification is performed on the regenerated G-code program to obtain the final machining trajectory. The specific steps include: Find the node with the largest deformation in the workpiece deformation cloud map to obtain the node with the largest deformation. Centered on the node with the maximum deformation, search for all voxels whose Euclidean distance from the node with the maximum deformation is within a set range on the simplified finite element model of the workpiece, and define the region formed by the voxels within the set range as the current deformation influence domain. Obtain the first tool position corresponding to the instantaneous cutting force load that causes the maximum deformation in the current deformation influence domain on the simulation time axis, and the set of voxels to be removed that it covers; Find the position of the tool immediately preceding the position of the first tool in the motion sequence, and use the point between the first tool position and the position of the tool immediately preceding the position as the compensation insertion point. Based on the simplified finite element model of the workpiece and the load direction relative to the position of the first tool, calculate the magnitude and direction of the equivalent compensation force vector required to generate the target reverse pre-deformation in the current deformation influence domain. At the location immediately adjacent to the previous tool position, with the geometric center of the current deformation influence domain as the direction reference, search for an adjacent voxel region with a low allowable removal rate value in the allowable material removal rate space field according to the direction of the required equivalent compensation force vector. Place the tool in the adjacent voxel region with a low allowable removal rate value, and adjust the tool pose so that during cutting, the tool-workpiece interaction force is opposite in direction to the equivalent compensation force vector in the principal component. In the adjacent voxel region with a low allowable removal rate value, a feed rate with an actual material removal rate lower than the allowable material removal rate of the corresponding voxel and a cutting force component sufficient to generate the required equivalent compensation force vector is configured. The tool position in the adjacent voxel region with a low allowable removal rate value and the path point of the feed rate are used as feedforward compensation path units. The feedforward compensation path unit is inserted into the compensation insertion point, and the theoretical optimal feed rate of the first tool position and subsequent path points is adaptively adjusted to obtain the updated motion sequence of the tool envelope. G-code is generated using the updated tool envelope motion sequence, and iterative simulation is performed again until the maximum deformation of the workpiece is lower than the preset threshold, in order to generate the final machining trajectory.

Images