A robotic adaptive milling method and system for wing-rib type parts

By employing online measurement and registration, differentiated parameters, and real-time adjustment methods, the problems of warping deformation and clamping deviation in the machining of rib-type parts were solved, achieving high-precision and stable milling results.

CN123103442APending Publication Date: 2026-11-17NINGBO XINYU MOLD TECHNOLOGY CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611307318.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-08-27
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Existing robotic milling technology fails to effectively consider the warping deformation of rib blanks and clamping posture deviations, resulting in low machining accuracy, easy deformation and chatter in thin-walled areas, and a lack of regionally differentiated parameters and real-time adaptive adjustment capabilities.

Method used

The actual surface point cloud data is obtained by online measurement device and registered with theoretical digital model. The processing area is divided and differentiated parameters are matched. Vibration signals and load current are collected in real time for fuzzy inference adjustment. Combined with flexible clamping device, the support force is maintained in real time to form a closed-loop control of the whole process.

Benefits of technology

It improves machining accuracy and surface quality, suppresses deformation and chatter in thin-walled areas, and achieves efficient and high-precision machining of rib-type parts.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123103442A_ABST
    Figure CN123103442A_ABST
Patent Text Reader

Abstract

This invention discloses a robot adaptive milling method and system for wing-rib type parts, relating to the field of wing-rib type part machining technology, including the following steps: Step 1: Obtain the theoretical digital model of the wing-rib part to be machined, and plan the initial machining trajectory of robot milling based on the theoretical digital model. The initial machining trajectory includes tool posture information and feed path information. This robot adaptive milling method and system for wing-rib type parts uses an online measurement device to scan the actual wing-rib part after clamping, and accurately registers and compares the obtained actual surface point cloud data with the theoretical digital model to obtain the machining allowance distribution and actual pose deviation of each machining area. This makes the generation and adjustment of the machining trajectory based on the actual blank state rather than relying solely on the theoretical digital model, fundamentally eliminating overcutting and undercutting defects caused by blank warping deformation and clamping pose deviation.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to wing-rib type parts machining technology, specifically to a robot adaptive milling machining method and system for wing-rib type parts. Background Technology

[0002] Wing ribs are core load-bearing components in aircraft wing structures. Typical wing ribs are characterized by thin walls, complex curved surfaces, and high aspect ratios. Their machining accuracy directly affects the aerodynamic performance and assembly quality of the wing. Currently, wing ribs are mainly machined using large five-axis CNC machine tools. Some companies are beginning to explore the introduction of industrial robots to compensate for the limited working space and high cost of CNC machine tools.

[0003] However, in existing robotic milling technologies, the machining trajectory is generated offline based entirely on the theoretical digital model, without considering the warping deformation of the rib blank during forging or heat treatment, or the pose deviation introduced by clamping and positioning. When there is a deviation between the actual workpiece surface and the theoretical digital model, machining according to the theoretical trajectory will inevitably lead to overcutting in some areas and undercutting in others. At the same time, the machining allowance distribution in different areas of the rib part is usually uneven. The top surface of the rib, the web surface, the flange surface, and the R-angle transition area have different structural stiffness and machining requirements. However, the existing technology uses uniform cutting parameters for machining the entire area, which cannot match the optimal parameters according to the actual allowance of each area. This makes thin-walled areas prone to deformation and chatter, and it is difficult to guarantee machining accuracy and surface quality. Summary of the Invention

[0004] The purpose of this invention is to provide a robotic adaptive milling method and system for wing-rib type parts, so as to solve the problems of low machining accuracy and easy deformation and chatter in thin-walled areas caused by ignoring the actual surface deviation and regional allowance differences of the blank in the prior art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a robot adaptive milling method for wing-rib type parts, comprising the following steps:

[0006] Step 1: Obtain the theoretical digital model of the rib part to be processed, and plan the initial machining trajectory of the robot milling based on the theoretical digital model. The initial machining trajectory includes tool posture information and feed path information.

[0007] Step 2: Use an online measuring device to scan the actual wing rib part after clamping to obtain actual surface point cloud data. Register and compare the actual surface point cloud data with the theoretical digital model to calculate the machining allowance distribution and actual pose deviation of each machining area of ​​the wing rib part.

[0008] Step 3: Based on the machining allowance distribution, the machining area of ​​the rib part is divided into a uniform allowance area, a gradual allowance area, and a sudden allowance area. Rough milling parameters and finish milling parameters are matched for different areas respectively to generate differentiated milling subroutines corresponding to each area.

[0009] Step 4: During the milling process performed by the robot, the vibration signal of the robot end effector and the spindle load current signal are collected in real time. Based on the vibration signal and load current signal, it is determined whether the current machining state is in a stable cutting range.

[0010] Step 5: When the machining state deviates from the stable cutting range, the feed rate and spindle speed in the current milling machining parameters are adjusted using fuzzy reasoning method, the compensated machining parameters are generated and the robot machining instructions are updated until the machining state returns to the stable cutting range.

[0011] Step 6: After completing the processing of the current area, the online measuring device is used again to re-measure the surface. The re-measurement results are compared with the theoretical model. If there is a margin deviation, steps 3 to 5 are repeated for compensation processing until all processed areas meet the dimensional accuracy requirements.

[0012] Further, in step two, the specific method for registering and comparing the actual surface point cloud data with the theoretical digital model is as follows: the actual surface point cloud and the theoretical digital model are coarsely registered using the iterative nearest point algorithm, and then finely registered using the nearest point iterative algorithm to obtain the spatial pose transformation matrix of the actual rib part relative to the theoretical digital model. The normal deviation value of each processing area is calculated based on the pose transformation matrix as the actual pose deviation.

[0013] Further, in step three, the calculation method for the machining allowance distribution is as follows: calculate the distance between each measurement point in the actual surface point cloud and the corresponding point in the theoretical digital model along the normal direction of the rib part surface, and use the distance value as the machining allowance value of that point; generate a machining allowance distribution map based on the machining allowance values ​​of each point, define the area where the machining allowance change rate is less than the first threshold as the uniform allowance area, define the area where the machining allowance change rate is between the first threshold and the second threshold as the gradual allowance area, and define the area where the machining allowance change rate is greater than the second threshold as the abrupt allowance area.

[0014] Furthermore, in step three, for the allowance abrupt change zone, a machining subroutine containing multiple layer milling is generated. The cutting depth of each layer is adaptively determined according to the actual allowance value of the region, and the feed direction of each layer cutting path is consistent with the extension direction of the ribs in the region.

[0015] Furthermore, in step four, the vibration signal is acquired by a triaxial accelerometer installed on the robot end effector, and the spindle load current signal is acquired by a current transformer installed on the power supply line of the spindle drive motor; the stable cutting range is obtained in advance based on the material properties of the rib parts and the cutting parameters of the tool used.

[0016] Furthermore, in step five, the fuzzy inference method includes: using vibration amplitude deviation and load current deviation as input variables of the fuzzy controller, using feed speed correction and spindle speed correction as output variables of the fuzzy controller, and performing fuzzy inference and defuzzification processing through a pre-established fuzzy control rule base to obtain the compensated machining parameters.

[0017] Furthermore, in step six, if the retest results show that the actual material removal amount in a certain processing area has not reached the target removal amount, the compensation cutting parameters for that area are recalculated based on the remaining material value, and the cutting depth in the compensation cutting parameters is less than the cutting depth of the previous processing.

[0018] A robotic adaptive milling system for wing-rib type parts, comprising:

[0019] The robot body has at least six degrees of freedom of motion, and its end is equipped with a milling electric spindle and a cutting tool.

[0020] An online measuring device is fixedly installed at the end of the robot body or independently set next to the processing station, and is used to collect the surface point cloud data of the wing rib parts;

[0021] A flexible clamping device is provided at the processing station for positioning and clamping rib parts. The flexible clamping device has multiple support units with independently adjustable support height.

[0022] The signal acquisition module includes a vibration sensor mounted on the robot's end effector and a current sensor mounted on the power supply line of the spindle drive motor.

[0023] The control unit is communicatively connected to the robot body, the online measurement device, the flexible clamping device, and the signal acquisition module. The control unit includes a trajectory planning module, a margin analysis module, a parameter adaptive adjustment module, and a machining status monitoring module.

[0024] The trajectory planning module generates an initial machining trajectory based on the theoretical numerical model of the rib part. The allowance analysis module generates machining allowance distribution information based on the comparison results between the actual surface point cloud data collected by the online measuring device and the theoretical numerical model. The parameter adaptive adjustment module matches differentiated machining parameters based on the machining allowance distribution information and corrects the machining parameters in real time based on the feedback signal from the machining status monitoring module. The machining status monitoring module determines the current machining status based on the vibration signal and current signal collected by the signal acquisition module.

[0025] Furthermore, each support unit of the flexible clamping device is equipped with an independent drive mechanism and force sensor. Each support unit automatically adjusts its support height according to the theoretical profile of the rib part, and each support unit detects the support contact force in real time through the force sensor. When the support contact force of any support unit exceeds the preset threshold range, the support unit automatically adjusts its support height to maintain the support contact force within the preset threshold range.

[0026] Furthermore, the online measuring device is a laser 3D scanner or a structured light 3D measuring instrument, and the control unit also includes a measurement path planning module. The measurement path planning module generates a scanning path covering the entire processing area of ​​the rib part based on the theoretical digital model of the rib part, and enables the online measuring device to perform scanning measurement along the scanning path.

[0027] Compared with the prior art, the robot adaptive milling method and system for wing-rib type parts provided by the present invention have the following beneficial effects:

[0028] The actual rib part is scanned after clamping using an online measuring device. The acquired actual surface point cloud data is precisely registered and compared with the theoretical digital model to obtain the machining allowance distribution and actual pose deviation of each machining area. This allows the generation and adjustment of the machining trajectory to be based on the actual blank state rather than solely relying on the theoretical digital model, fundamentally eliminating over-cutting and under-cutting defects caused by blank warping and clamping pose deviations. Based on this, the machining area is divided into a uniform allowance zone, a gradual change zone, and abrupt change zone according to the allowance change rate, and differentiated machining parameters are matched. Layered milling is used in the allowance abrupt change zone, with the feed direction consistent with the rib extension direction. This avoids the deformation and chatter problems in thin-walled areas caused by uniform parameters across the entire area in existing technologies, improving machining accuracy and surface quality.

[0029] By acquiring vibration signals from the end effector and spindle load current signals in real time during robotic milling, quantitative monitoring methods accurately determine whether the current machining state is within the stable cutting range, replacing the subjective judgment method of relying entirely on operator visual and auditory perception in existing technologies. When the machining state deviates from the stable cutting range, a fuzzy inference method is used to automatically adjust the feed rate and spindle speed and update the machining commands in real time, realizing closed-loop real-time adaptive adjustment of machining parameters. The response speed and adjustment accuracy are significantly better than the manual emergency stop intervention method in existing technologies, effectively suppressing machining chatter and ensuring the stability and consistency of the machining process.

[0030] By re-measuring the surface using an online measuring device after machining the current area, and automatically comparing the re-measuring results with the theoretical digital model, and automatically performing compensatory machining when there is a deviation in the allowance, a closed-loop control of the entire process of "measurement-machining-re-measuring-compensation" is formed. This eliminates the inefficiency and positioning accuracy loss caused by offline inspection, manual analysis, and re-clamping and repair in existing technologies. At the same time, the flexible clamping device dynamically maintains the support contact force within a preset threshold range throughout the machining process through independent drive of each support unit and real-time feedback from force sensors. This overcomes the clamping instability problem caused by changes in workpiece stiffness due to material removal in existing mechanical clamping methods, further ensuring the efficient and high-precision machining of ribbed thin-walled parts. Attached Figure Description

[0031] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0032] Figure 1 This is a flowchart of a robot adaptive milling method for wing-rib type parts according to the present invention. Detailed Implementation

[0033] To enable those skilled in the art to better understand the technical solution of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings.

[0034] Example 1

[0035] This embodiment provides a complete implementation process for a robotic adaptive milling method for wing rib-type parts. In this embodiment, the wing rib to be machined is a mid-section wing rib of a civil aircraft wing, made of 7075-T6 aluminum alloy. The blank is an integral forging, with an external outline dimension of approximately 1200mm × 800mm × 150mm. It has multiple T-shaped cross-section ribs and a web area with a thickness ranging from 3mm to 8mm, belonging to a typical complex thin-walled structural component.

[0036] In this embodiment, step one is first performed: obtaining the theoretical digital model of the rib part to be machined. This theoretical digital model is converted from the three-dimensional design model of the rib part and is in the STEP or IGES standard format. After importing the theoretical digital model into the control unit, the trajectory planning module generates the initial machining trajectory based on the geometric features of the rib part. The planning process of the initial machining trajectory adopts a feature-based programming method, that is, first identifying the machining features in the rib part digital model, such as the top surface of the rib, the side surface of the rib, the upper surface of the web, the lower surface of the web, and the flange surface, and then matching the corresponding machining strategy and tool path mode for each feature. In this embodiment, a 20mm diameter solid carbide end mill is used in the rough milling stage, and a 12mm diameter solid carbide end mill is used in the finish milling stage. The initial machining trajectory includes a reciprocating feed path along the length of the rib and tool attitude information corresponding to the normal direction of each machined surface. The tool attitude information is stored in the machining command in Euler angle form.

[0037] Step two is then executed: After the wing rib part blank is positioned and clamped by the flexible clamping device, the control unit drives the online measuring device to scan the actual wing rib part surface. In this embodiment, the online measuring device is a laser 3D scanner, mounted on an independent bracket beside the end of the robot body, with a scanning accuracy of ±0.05mm. The measurement path planning module generates a scanning path covering the entire area to be processed based on the theoretical digital model of the wing rib part. The scanning path is arranged in an S-shape along the length of the wing rib, with an overlap rate of 15% between adjacent scanning zones. The laser 3D scanner scans along this scanning path, acquiring actual surface point cloud data, obtaining approximately 1.2 million measurement points. The acquired actual surface point cloud data is imported into the allowance analysis module of the control unit. Coarse registration is performed using an iterative nearest point algorithm, followed by fine registration using an iterative nearest point algorithm, aligning the actual surface point cloud with the theoretical digital model. After registration, the distance from each measurement point to the corresponding point in the theoretical digital model is calculated along the normal direction of the wing rib surface; this distance represents the machining allowance value for that point. Simultaneously, based on the pose transformation matrix obtained from registration, the six degrees of freedom deviation of the rib part relative to the theoretical position in the actual clamping state are calculated as the actual pose deviation.

[0038] In this embodiment, through registration and comparison, it was found that the rib part has a warping deformation of about 2mm in the middle area along the length direction, the machining allowance of the edge area at one end is about 4mm to 5mm, and the machining allowance of the edge area at the other end is about 2mm to 3mm, showing obvious uneven distribution.

[0039] Next, step three is executed: the allowance analysis module generates a machining allowance distribution map based on the calculated machining allowance values ​​at each point. A first threshold is set at 0.5 mm / m (i.e., the allowance variation per meter along the rib surface does not exceed 0.5 mm), and a second threshold is set at 2 mm / m. Areas with a machining allowance variation rate less than 0.5 mm / m are designated as uniform allowance zones, mainly including the flange surfaces at both ends of the rib; areas with a machining allowance variation rate between 0.5 mm / m and 2 mm / m are designated as gradual allowance zones, mainly including certain sections of the web area; areas with a machining allowance variation rate greater than 2 mm / m are designated as abrupt allowance zones, mainly appearing in the transition area between the rib root and the web, and in the intermediate areas with significant warping deformation.

[0040] For the three types of regions mentioned above, the parameter adaptive adjustment module matches differentiated machining parameters. For the uniform allowance region, constant milling parameters are used: rough milling depth of cut is 3mm, feed rate is 1500mm / min, and spindle speed is 8000r / min. For the gradually changing allowance region, the machining area is divided into several sub-intervals along the allowance change direction. The cutting depth of each sub-interval is adaptively determined based on the average allowance value of that sub-interval, and the feed rate is appropriately increased as the allowance decreases. For the abruptly changing allowance region, a machining subroutine containing multiple layered milling operations is generated. The number of layers is determined based on the ratio of the maximum allowance value to the maximum allowable depth of cut in a single operation. The cutting depth of each layer is adaptively allocated based on the actual allowance value of the region, and the feed direction of each cutting path is set to be consistent with the extension direction of the ribs in that region to reduce the extrusion deformation of the rib sidewalls by the milling force.

[0041] In this embodiment, the parameter adaptive adjustment module generates corresponding machining program segments for each region, which are then merged to form a complete robot machining program.

[0042] Then, step four is executed: the robot begins milling. During the machining process, a triaxial accelerometer mounted on the robot's end effector collects vibration signals in real time, and a current transformer mounted on the spindle drive motor's power supply line collects the spindle load current signal in real time. The collected signals are filtered, amplified, and then transmitted to the machining status monitoring module. The machining status monitoring module compares the real-time collected vibration amplitude with a pre-calibrated stable cutting range threshold. In this embodiment, the calibration method for the stable cutting range is as follows: using the same tool and material as the machining, a variable parameter cutting test is performed on the workpiece, and the vibration signals and load current signals under different combinations of cutting parameters are recorded. The signal range corresponding to the machining state where the surface roughness meets the requirements and there are no chatter marks is calibrated as the stable cutting range.

[0043] When machining the middle area of ​​the rib, the vibration signal amplitude exceeds the upper limit of the stable cutting range. At this point, step five is executed: the parameter adaptive adjustment module starts the fuzzy inference program. The fuzzy controller uses the vibration amplitude deviation and load current deviation as input variables, and the feed rate correction and spindle speed correction as output variables. The fuzzy control rule base contains 25 inference rules, such as "if the vibration deviation is large and the current deviation is small, then the feed rate correction is large negative and the spindle speed correction is small positive." Through fuzzy inference and center-of-gravity defuzzification, the specific correction amount is obtained. In this embodiment, the vibration amplitude exceeds the threshold of 12%, and after fuzzy inference, the output correction amount is a 15% reduction in feed rate and a 5% increase in spindle speed. The robot controller updates the machining instructions accordingly.

[0044] After the machining parameters are adjusted, machining continues. Once the machining status monitoring module confirms that the vibration signal has returned to the stable cutting range, the current machining parameters are maintained to complete the machining of that area.

[0045] Finally, in step six: after completing the rough and finish milling of all areas, the laser 3D scanner is restarted to re-measure the surface of the machined part. Comparing the re-measurement results with the theoretical model reveals a localized area of ​​approximately 0.15mm undercut in the web near the root of the rib. For this localized area, the compensation cutting parameters are recalculated based on the remaining allowance, setting the depth of cut to 0.15mm and reducing the feed rate to 800mm / min to ensure machining stability. Compensation machining is then performed. After compensation machining, the dimensions of this area are measured again, and the accuracy requirement of ±0.1mm is met, completing the entire machining process.

[0046] Example 2

[0047] This embodiment provides a robot adaptive milling method for wing-rib type parts. The main difference between this method and the first embodiment lies in the specific implementation of the registration and comparison method between the actual surface point cloud data and the theoretical digital model in step two.

[0048] In this embodiment, the wing rib part to be processed is a wing rib plate of a certain type of UAV, made of 2024-T3 aluminum alloy, with relatively small dimensions of approximately 500mm × 300mm × 60mm. Since the web thickness of this wing rib part is only 2mm, it is an ultra-thin-walled structure and extremely sensitive to deformation during processing. Therefore, this embodiment has specifically optimized the registration and comparison method in step two.

[0049] Specifically, the online measurement device in this embodiment uses a structured light 3D measuring instrument, which has a fast measurement speed and a large field of view in a single measurement, making it suitable for rapid full-view scanning of small-sized parts. After acquiring the actual surface point cloud, in the coarse registration stage, principal component analysis is first used to calculate the principal direction vectors for the actual surface point cloud and the theoretical digital model respectively. The initial pose estimation of the point cloud and the digital model is achieved by aligning the principal directions, thereby reducing the number of calculation iterations in the subsequent fine registration. In the fine registration stage, the nearest point iteration algorithm with weighted factors is used. The weighting factors are set according to the sensitivity of each region of the rib part to the machining accuracy. Specifically, the top surface of the rib rib and the edge surface are the final assembly surfaces, and their registration accuracy is given a higher weight, with a weighting coefficient of 1.5; the web surface is a non-assembly surface, and its registration accuracy is given a lower weight, with a weighting coefficient of 0.8.

[0050] Using the weighted registration method described above, the registration accuracy of key assembly surfaces is prioritized while ensuring the overall registration accuracy meets requirements. After registration, the normal deviation value of each machining area is calculated. In this embodiment, due to the extremely thin wall thickness of the part, a bulge deformation of approximately 1.5 mm occurs in the web area after clamping, while the deformation in the flange area is less than 0.2 mm due to its larger cross-sectional thickness. The registration method of this embodiment can accurately identify the deformation differences in different areas, providing accurate basic data for matching subsequent differentiated machining parameters.

[0051] The remaining steps in this embodiment are the same as in Embodiment 1, and will not be repeated here.

[0052] Example 3

[0053] This embodiment provides an implementation process for a robot adaptive milling method for wing-rib type parts, focusing on the specific implementation scheme of machining status monitoring and adaptive adjustment in steps four and five.

[0054] In this embodiment, the wing rib part to be processed is the central wing rib of a certain type of large transport aircraft. The material is 7050-T7451 aluminum alloy. It has large dimensions, approximately 2500mm in length and 600mm in height, and has multiple I-shaped ribs with varying cross-sections. Due to the large processing stroke of this wing rib part, the stiffness of the robot varies significantly under different processing configurations, making processing stability a particularly prominent issue.

[0055] In step four of this embodiment, in addition to using a triaxial accelerometer to collect vibration signals, angle encoders are also installed at each joint of the robot to collect the angle values ​​of each joint in real time. The machining status monitoring module calculates the stiffness characteristic value of the robot's current machining configuration based on the angle values ​​of each joint, and makes a comprehensive judgment by combining the vibration signal and the load current signal. Specifically, the judgment criterion for the stable cutting range is not fixed, but dynamically adjusted according to the stiffness characteristic value of the current robot configuration. The natural frequency distribution map under different configurations is obtained in advance through robot modal testing. The natural frequency at each machining position is compared with the excitation frequency under the current cutting parameters. When the excitation frequency is close to the natural frequency, the threshold of the stable cutting range is narrowed to improve the sensitivity to vibration signals.

[0056] In the fuzzy inference method of step five, this embodiment uses the stiffness characteristic value of the robot's current configuration as the third input variable of the fuzzy controller, forming a three-input, two-output fuzzy control rule base. The establishment of the fuzzy control rule base is based on expert knowledge of robot milling processes and orthogonal cutting test data. When the robot is in a low-stiffness configuration, even if the vibration signal amplitude does not exceed the conventional threshold, the fuzzy controller will actively reduce the feed rate to avoid potential chatter. When the robot is in a high-stiffness configuration, the machining efficiency can be appropriately increased while ensuring machining quality.

[0057] In this embodiment, when the robot moves to the end region of the wing rib, the robot's wrist joint is in a large-angle deflection configuration, resulting in low system stiffness. At this time, the machining status monitoring module dynamically narrows the stable cutting interval threshold based on the configuration stiffness characteristic value and actively adjusts the feed rate from 1500 mm / min to 1100 mm / min, effectively avoiding chatter in this region. After machining, the surface roughness Ra value reaches 0.8 μm, meeting the design requirements.

[0058] Example 4

[0059] This embodiment provides a robotic adaptive milling system for wing-rib type parts. The main difference between this system and the system configuration in Embodiment 1 lies in the detailed structure and working method of the flexible clamping device.

[0060] The flexible clamping device in this embodiment includes 10 independently adjustable support units mounted on a machining platform. Each support unit is arranged in two rows of five units, with five units in each row, along the length of the rib part. Each support unit is equipped with a servo motor-driven lead screw lifting mechanism and a force sensor mounted on the top of the support head.

[0061] During the part clamping stage, the control unit first calculates the ideal support height of each support unit based on the theoretical digital model of the rib part, ensuring that the top surfaces of each support head together form a support surface that matches the theoretical lower surface of the rib part. The servo motor of each support unit drives the lead screw lifting mechanism to adjust the support head to the predetermined height based on this ideal height value. Then, the rib part blank is placed on each support head, and the force sensor on each support head detects the support contact force in real time.

[0062] During processing, as material is continuously removed, the stiffness distribution of the rib components changes, and localized areas may experience deformation due to stress release. When the support contact force detected by any support unit exceeds a preset threshold range, the support unit automatically initiates a height adjustment program, fine-tuning the support head height via a servo motor to bring the support contact force back within the preset threshold range. In this embodiment, the preset threshold range is 50N to 200N. When the force sensor of a support unit detects that the contact force drops below 30N, it indicates that the component in that area has detached from the support head due to deformation, and the support head automatically rises by 0.5mm. When the contact force rises above 250N, it indicates that the area is excessively compressed due to deformation, and the support head automatically descends by 0.3mm. Through this active adaptive adjustment of the support force, the stable support state of the rib components is maintained throughout the processing, effectively suppressing vibration of the thin-walled structure during processing.

[0063] Furthermore, the flexible clamping device in this embodiment also includes a clamping force control module. Each support unit is equipped with a pneumatic clamping mechanism, which uses a larger clamping force during the roughing stage and appropriately reduces the clamping force during the finishing stage to minimize the impact of workpiece deformation caused by the clamping force on the finishing accuracy.

[0064] Example 5

[0065] This embodiment provides a robot adaptive milling method for wing-rib type parts, which focuses on the optimization of the compensation machining strategy in step six.

[0066] In this embodiment, the wing rib part to be processed is an integral rib plate of a certain type of fighter jet wing, made of titanium alloy TC4, which has high strength and low thermal conductivity, and is considered a difficult-to-machine material. The machining accuracy requirements for this wing rib part are high, with a web thickness tolerance of ±0.08mm and a rib height tolerance of ±0.1mm.

[0067] In step six, the surface re-measurement stage, the same laser 3D scanner used in step two was used to perform a full surface re-measurement of the machined part. The re-measurement results showed that there were three areas with varying degrees of undercutting: one area in the middle of the web with a remaining allowance of 0.18mm, one area at the root of the rib with a remaining allowance of 0.12mm, and one area on the end face of the flange with a remaining allowance of 0.08mm.

[0068] For these three regions, this embodiment employs differentiated compensation machining strategies. For the central web region with the largest remaining allowance, since the web thickness in this region is close to the final size and the thin-wall stiffness is low, a smaller cutting depth (set to 0.12 mm) is used during compensation machining, and climb milling is employed to reduce thin-wall deformation caused by milling forces. For the R-corner region at the root of the rib, since this region has a curved transition structure, a ball end mill is used during compensation machining, and the tool axis is kept aligned with the normal direction of the machined surface to prevent overcutting adjacent surfaces. For the flange end face region with a remaining allowance of only 0.08 mm, the tolerance requirement for this region is ±0.1 mm. Polishing is used directly to remove the remaining allowance without further cutting to avoid accuracy loss due to re-clamping and positioning.

[0069] After completing the above-mentioned differential compensation processing, a retest was conducted, and the dimensional deviations of all processed areas were within the tolerance range. This embodiment also compared the effect of using a single-parameter compensation processing method. The results show that the processing pass rate using the differential compensation strategy of this embodiment is 98%, while the processing pass rate using the single-parameter compensation method is only 82%, verifying the effectiveness of the differential compensation strategy of this embodiment.

[0070] Comparison Example

[0071] I. Background of the Comparative Case

[0072] In the aerospace manufacturing field, the milling of wing rib parts has long relied primarily on large five-axis CNC machine tools. With the development of industrial robot technology, some manufacturing companies have begun to explore the introduction of robots into the machining of wing rib parts to compensate for the limited working space and high cost of CNC machine tools. The comparative examples described below represent common robotic milling machining solutions in the industry and reflect the prior art level prior to the date of this invention application.

[0073] II. System Composition of the Comparative Example

[0074] The robotic milling system of the comparative example includes the following components:

[0075] (a) Robot body and machining execution unit

[0076] The system employs a six-axis industrial robot, with a milling electric spindle and tool holder mounted at the robot's end effector. The robot has a large workspace, covering the entire machining range of the ribbed parts. The milling electric spindle's speed is manually preset via a frequency converter. Before machining, the operator sets a fixed speed value on the control panel based on the tool diameter and workpiece material, and this value remains unchanged during machining.

[0077] (ii) Workpiece clamping unit

[0078] Traditional mechanical clamps are used to clamp and position the rib parts, including multiple locating blocks and pressure plates mounted on the machining platform. The operator manually adjusts the positions of the locating blocks according to the dimensions of the rib parts, and uses the pressure plates to hold the parts firmly on the machining platform. After clamping, a dial indicator or micrometer is used to manually check and correct the levelness and position of the parts.

[0079] (III) Offline Programming and Trajectory Generation Unit

[0080] The system is equipped with offline programming software. Operators import the theoretical CAD model of the rib part into the offline programming software, which automatically generates the tool path based on the model. The generation of the machining path is entirely based on the geometric information of the theoretical model, including the position coordinates of the toolpath points and the direction of the tool axis. The generated machining path is saved in G-code or robot-specific instruction format and imported into the robot controller once before machining.

[0081] (iv) Measurement Unit

[0082] The system is equipped with a coordinate measuring machine or a handheld 3D scanner, independent of the robot, for dimensional inspection of the rib parts before and after machining. The measurement and machining operations are performed in separate timeframes. After the parts are clamped, the measuring personnel first perform surface inspection, and then the measurement data is imported into the computer for manual comparison and analysis. Based on the comparison results, the process engineers manually adjust the machining parameters or modify the machining trajectory and regenerate the machining program.

[0083] III. Processing flow of the comparative example

[0084] Step 1: Workpiece clamping and alignment

[0085] The operator hoists the rib blank onto the mechanical fixture on the machining platform, manually adjusts the position of each pressure plate, and then tightens the clamping bolts to secure the workpiece. A dial indicator is then used to select several test points on the workpiece's upper surface to measure the workpiece's positional deviation relative to the machine tool coordinate system. The positional deviation is controlled within the allowable range by tapping the workpiece or adding shims. This alignment process typically takes 30 to 60 minutes and is highly dependent on the operator's experience.

[0086] Step 2: Offline Programming and Trajectory Generation

[0087] The process engineers open the theoretical digital model of the rib part in the offline programming software, set the origin of the machining coordinate system, and select the tool type and cutting parameters (including depth of cut, feed rate, spindle speed, etc.). The software automatically generates roughing and finishing trajectories. All machining areas use uniform cutting parameters, without distinguishing between different areas' allowance differences. After the generated machining program is imported into the robot controller, the operator performs a no-load run to verify that there is no collision interference on the trajectory before preparing for formal machining.

[0088] Step 3: Pre-processing surface measurement (optional)

[0089] For blanks with significant deformation, some operators will use a handheld 3D scanner to scan the clamped workpiece before formal machining to obtain the actual surface point cloud. The point cloud data is then imported into third-party reverse engineering software and manually registered and compared with the theoretical digital model. Process engineers determine which areas require adjustments to the machining allowance based on the comparison results. If adjustments are needed, the process engineers manually modify the machining trajectory or cutting parameters in the offline programming software, regenerate the machining program, and re-import it into the robot controller. Because this process is time-consuming, many operators skip this step in actual production and directly machine according to the theoretical trajectory.

[0090] Step 4: Milling Machining

[0091] The robot is started to execute the machining program. During machining, the operator visually observes and listens to determine if the machining status is normal. If abnormal vibration, noise, or cutting sparks are detected, the operator manually presses the emergency stop button to pause machining, adjusts the feed rate multiplier knob or spindle speed based on experience, and then restarts machining. Throughout the entire machining process, the system does not automatically collect vibration signals or load current signals, and does not perform any form of real-time status monitoring or automatic parameter adjustment.

[0092] Step 5: Post-processing inspection and rework

[0093] After machining, the rib parts are unloaded from the fixture and transferred to the coordinate measuring machine (CMM) inspection area. Measuring personnel measure key dimensions point by point according to the inspection plan, generating an inspection report. If overcut or undercut defects are found, process engineers analyze and judge based on the defect location and severity, deciding whether rework is necessary and the rework plan. For undercut areas, manual hand-held pneumatic tools are typically used for grinding, or the workpiece is re-clamped and partially machined on a robot. Because the clamping and positioning accuracy during rework is difficult to guarantee, the reworked parts often still fail to achieve ideal accuracy, resulting in a high scrap rate.

[0094] IV. Technical defects in the comparative examples

[0095] The above comparative examples represent the general technical solutions for robotic milling of wing-ribbed parts prior to the application date of this invention, which have the following main technical defects:

[0096] 1. The machining trajectory lacks adaptability to the actual state of the blank.

[0097] The machining trajectory of the comparative example is generated entirely based on the theoretical digital model, without considering the warping deformation of the rib blank during forging or heat treatment, or the positional deviation introduced by clamping and positioning. When there is a deviation between the actual workpiece profile and the theoretical digital model, machining according to the theoretical trajectory will inevitably lead to overcutting in some areas and undercutting in others, making it difficult to guarantee the consistency of machining accuracy.

[0098] 2. The processing parameters are too simple and lack regional differentiation strategies.

[0099] Different regions of a rib component (rib top surface, web surface, flange surface, R-angle transition area, etc.) have different structural stiffness and machining requirements, and the machining allowance distribution in each region is usually uneven. Using uniform cutting parameters for machining the entire region, as in the comparative example, fails to match the optimal machining parameters according to the actual allowance and structural characteristics of each region, resulting in low machining efficiency and a tendency for deformation or chatter in weak areas.

[0100] 3. Lack of real-time status monitoring and adaptive adjustment capabilities during processing.

[0101] The machining process in the comparative example is an "open-loop" control, meaning that the machining program is executed once according to the preset trajectory and parameters and then ends. The system does not collect status information such as vibration and cutting force during the machining process, and it does not have the ability to adjust machining parameters in real time according to the actual cutting conditions. When situations such as accelerated tool wear or changes in workpiece stiffness leading to chatter occur during machining, the system cannot respond automatically and can only rely on manual judgment and intervention by the operator, resulting in slow response speed and low reliability.

[0102] 4. The measurement and processing stages are disconnected, resulting in low rework efficiency.

[0103] In the comparative example, the three stages of surface measurement, machining, and re-measurement are independent of each other, and are completed on offline programming software, robot systems, and coordinate measuring machines, respectively. Data transfer relies on manual import and conversion, and measurement results cannot be fed back to the machining system in real time to form a closed-loop control. Defects found after machining need to be manually analyzed by process engineers to formulate rework plans. The quality of rework is greatly affected by human factors, and the workpiece needs to be reclamped during the rework process, making it difficult to guarantee positioning accuracy.

[0104] 5. The clamping method lacks self-adaptive capability.

[0105] The comparative example uses a mechanical clamping plate fixture, where the support position and clamping force are manually set, and the support layout and clamping force distribution cannot be automatically adjusted according to the actual deformation state of the workpiece. For thin-walled rib parts, an unreasonable clamping scheme is itself one of the important causes of machining deformation.

[0106] The following table compares the effects of a robot adaptive milling method and system for wing-rib type parts (Examples 1 to 5) with a control example:

[0107]

[0108]

[0109]

[0110]

[0111] Compared with the comparative example, the embodiments of the present invention achieve a substantial leap from "open-loop experience-based machining" to "closed-loop adaptive intelligent machining". Before machining, each embodiment obtains the actual surface through online measurement and accurately registers it with the theoretical digital model, eliminating the influence of blank deformation and clamping deviation, and solving the inherent defects of overcutting and undercutting in the comparative example. In particular, Embodiment 2 introduces key surface weight registration for ultra-thin-walled parts to prioritize the accuracy of the assembly surface. In terms of machining parameter setting, each embodiment breaks through the uniform parameter mode of the comparative example, divides the region according to the allowance change rate and matches differentiated parameters, and adopts layered milling for the allowance change zone with the feed direction consistent with the extension of the rib, which improves efficiency and suppresses thin-wall deformation and chatter. During machining, each embodiment realizes quantitative monitoring of the machining state by real-time acquisition of vibration and load current signals, replacing the manual experience judgment of the comparative example; when the state is abnormal, the feed speed and spindle speed are automatically adjusted in real time through fuzzy inference, and the response speed and accuracy far exceed the manual intervention of the comparative example. Embodiment 3 further incorporates the robot configuration stiffness characteristic value into fuzzy inference to accurately suppress vibration at different positions. After processing, each embodiment forms a closed-loop process through online retesting and automatic compensation, eliminating the inefficiency and accuracy loss caused by offline inspection, manual analysis, and re-clamping for rework in the comparative example. Embodiment 5 performs differentiated compensation for the remaining allowance in different areas to avoid over-cutting during rework. Furthermore, the flexible clamping device in Embodiment 4 dynamically maintains stable contact force through independent drive of each support unit and real-time feedback from force sensors, overcoming the vibration problem caused by changes in workpiece stiffness under fixed clamping in the comparative example. In summary, each embodiment achieves substantial improvements over the comparative example across the entire process chain, jointly ensuring efficient and high-precision processing of rib-type parts.

[0112] The foregoing has only described certain exemplary embodiments of the present invention by way of illustration. Undoubtedly, those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the foregoing drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. A robotic adaptive milling method for wing-rib type parts, characterized in that, Includes the following steps: Step 1: Obtain the theoretical digital model of the rib part to be processed, and plan the initial machining trajectory of the robot milling based on the theoretical digital model. The initial machining trajectory includes tool posture information and feed path information. Step 2: Use an online measuring device to scan the actual wing rib part after clamping to obtain actual surface point cloud data. Register and compare the actual surface point cloud data with the theoretical digital model to calculate the machining allowance distribution and actual pose deviation of each machining area of ​​the wing rib part. Step 3: Based on the machining allowance distribution, the machining area of ​​the rib part is divided into a uniform allowance area, a gradual allowance area, and a sudden allowance area. Rough milling parameters and finish milling parameters are matched for different areas respectively to generate differentiated milling subroutines corresponding to each area. Step 4: During the milling process performed by the robot, the vibration signal of the robot end effector and the spindle load current signal are collected in real time. Based on the vibration signal and load current signal, it is determined whether the current machining state is in a stable cutting range. Step 5: When the machining state deviates from the stable cutting range, the feed rate and spindle speed in the current milling machining parameters are adjusted using fuzzy reasoning method, the compensated machining parameters are generated and the robot machining instructions are updated until the machining state returns to the stable cutting range. Step 6: After completing the processing of the current area, the online measuring device is used again to re-measure the surface. The re-measurement results are compared with the theoretical model. If there is a margin deviation, steps 3 to 5 are repeated for compensation processing until all processed areas meet the dimensional accuracy requirements.

2. The robot adaptive milling method for wing-rib type parts according to claim 1, characterized in that, In step two, the specific method for registering and comparing the actual surface point cloud data with the theoretical digital model is as follows: the actual surface point cloud and the theoretical digital model are coarsely registered using the iterative nearest point algorithm, and then finely registered using the nearest point iterative algorithm to obtain the spatial pose transformation matrix of the actual rib part relative to the theoretical digital model. The normal deviation value of each processing area is calculated based on the pose transformation matrix as the actual pose deviation.

3. The robot adaptive milling method for wing-rib type parts according to claim 1, characterized in that, In step three, the calculation method for the machining allowance distribution is as follows: calculate the distance between each measurement point in the actual surface point cloud and the corresponding point in the theoretical digital model along the normal direction of the rib part surface, and use the distance value as the machining allowance value of that point; generate a machining allowance distribution map based on the machining allowance values ​​of each point, define the area where the machining allowance change rate is less than the first threshold as the uniform allowance area, define the area where the machining allowance change rate is between the first threshold and the second threshold as the gradual allowance area, and define the area where the machining allowance change rate is greater than the second threshold as the abrupt allowance area.

4. The robot adaptive milling method for wing-rib type parts according to claim 1, characterized in that, In step three, for the area of ​​abrupt change in allowance, a machining subroutine containing multiple layers of milling is generated. The cutting depth of each layer is adaptively determined according to the actual allowance value of the area, and the feed direction of each layer cutting path is consistent with the extension direction of the ribs in the area.

5. The robot adaptive milling method for wing-rib type parts according to claim 1, characterized in that, In step four, the vibration signal is acquired by a triaxial accelerometer installed on the robot end effector, and the spindle load current signal is acquired by a current transformer installed on the power supply line of the spindle drive motor. The stable cutting range is obtained in advance based on the material properties of the rib part and the cutting parameters of the tool used.

6. The robot adaptive milling method for wing-rib type parts according to claim 1, characterized in that, In step five, the fuzzy inference method includes: using vibration amplitude deviation and load current deviation as input variables of the fuzzy controller, and feed speed correction and spindle speed correction as output variables of the fuzzy controller, performing fuzzy inference and defuzzification processing through a pre-established fuzzy control rule base to obtain the compensated machining parameters.

7. The robot adaptive milling method for wing-rib type parts according to claim 1, characterized in that, In step six, if the retest results show that the actual material removal amount in a certain processing area has not reached the target removal amount, the compensation cutting parameters for that area are recalculated based on the remaining amount. The cutting depth in the compensation cutting parameters is less than the cutting depth of the previous processing.

8. A robotic adaptive milling system for wing-rib type parts, applicable to the robotic adaptive milling method for wing-rib type parts as described in any one of claims 1-7, characterized in that, include: The robot body has at least six degrees of freedom of motion, and its end is equipped with a milling electric spindle and a cutting tool. An online measuring device is fixedly installed at the end of the robot body or independently set next to the processing station, and is used to collect the surface point cloud data of the wing rib parts; A flexible clamping device is provided at the processing station for positioning and clamping rib parts. The flexible clamping device has multiple support units with independently adjustable support height. The signal acquisition module includes a vibration sensor mounted on the robot's end effector and a current sensor mounted on the power supply line of the spindle drive motor. The control unit is communicatively connected to the robot body, the online measurement device, the flexible clamping device, and the signal acquisition module. The control unit includes a trajectory planning module, a margin analysis module, a parameter adaptive adjustment module, and a machining status monitoring module. The trajectory planning module generates an initial machining trajectory based on the theoretical numerical model of the rib part. The allowance analysis module generates machining allowance distribution information based on the comparison results between the actual surface point cloud data collected by the online measuring device and the theoretical numerical model. The parameter adaptive adjustment module matches differentiated machining parameters based on the machining allowance distribution information and corrects the machining parameters in real time based on the feedback signal from the machining status monitoring module. The machining status monitoring module determines the current machining status based on the vibration signal and current signal collected by the signal acquisition module.

9. A robotic adaptive milling system for wing-rib type parts according to claim 8, characterized in that, Each support unit of the flexible clamping device is equipped with an independent drive mechanism and force sensor. Each support unit automatically adjusts its support height according to the theoretical profile of the rib part, and each support unit detects the support contact force in real time through the force sensor. When the support contact force of any support unit exceeds the preset threshold range, the support unit automatically adjusts its support height to maintain the support contact force within the preset threshold range.

10. A robotic adaptive milling system for wing-rib type parts according to claim 8, characterized in that, The online measuring device is a laser 3D scanner or a structured light 3D measuring instrument. The control unit also includes a measurement path planning module, which generates a scanning path covering the entire processing area of ​​the rib part based on the theoretical digital model of the rib part, and causes the online measuring device to perform scanning measurement along the scanning path.