An experimental platform and adjustment method for hole position compensation and posture detection

Abstract

The present invention relates to the field of precision enhancement of automated intelligent hole-making systems, and in particular to an experimental platform and adjustment method for hole position compensation and posture detection. The purpose of the present invention is to verify the experimental compensation method by building an experimental platform and decomposing the processing steps of the hole-making system. The experimental platform establishes corresponding models and system simulation workstations respectively. The adjustment method corrects the theoretical numerical model of the hole position by solving the hole position conversion model, eliminating outliers with the RANSAC algorithm and the Grubbs statistical method; solving the spatial posture direction of the hole position surface with the vector method, realizing posture correction with the Euler angle matrix decomposition, and finally performing quantitative and qualitative evaluation in the simulation platform. The present invention can avoid unnecessary errors caused by reasons such as error accumulation, and retains necessary error sources such as workpiece clamping. The offline program can be directly updated. The compensation accuracy of the hole-making error has been verified to meet the requirements, which is of great significance in the field of precision enhancement of automated intelligent hole-making systems.

Description

Technical Field

[0001] The present invention relates to the field of precision enhancement of automated intelligent hole-making systems, and in particular to an experimental platform and an adjustment method for hole position compensation and posture detection. Background Art

[0002] At present, the automated intelligent hole-making system has replaced the traditional manual hole-making method and has become the mainstream of the era. In particular, in the aviation field, holes are made in aircraft fuselages and wing box panels. Since most of the materials are CFRP titanium alloy composite materials, they are difficult to process and the processing quality is more difficult to guarantee. The use of automated intelligent hole-making can better highlight its advantages and achieve better hole-making quality and efficiency. In the field of hole position compensation and attitude detection, the existing test methods rely on large-scale processing equipment for hole-making processing or hole-making simulation, and their verification process is quite complicated. More importantly, the accumulation of errors in the test process will introduce unnecessary error terms, which can easily interfere with the results and increase the amount of calculation. Therefore, in response to the current status of the field of precision enhancement of automated intelligent hole-making systems, the present invention proposes an experimental platform for verifying hole position compensation and attitude detection. On the basis of retaining the necessary error sources and not affecting the processing flow of the automated intelligent hole-making system, the hole position compensation and attitude detection are verified separately through step decomposition. This invention leverages the flexible configuration of the experimental platform. On the one hand, the proposed hole position compensation method effectively filters out holes that significantly impact the results while maintaining high precision and robustness. On the other hand, the proposed normal detection method allows for flexible and adaptive adjustment of the workpiece surface posture, enabling rapid fitting of the normal deflection angle and a wide range of verification applications. The final compensation and detection results meet the threshold requirements for precision enhancement in automated intelligent hole-making systems, effectively verifying both hole position compensation and posture detection. Summary of the Invention

[0003] The purpose of the present invention is to provide an experimental platform and adjustment method for hole position compensation and posture detection, and to verify hole position compensation and posture detection respectively by decomposing the steps of the hole making process.

[0004] The present invention is achieved through the following technical solutions:

[0005] An experimental platform for hole position compensation and posture detection is described. Its main body consists of support legs, a bottom worktable, a right-angle connecting plate, a Y-axis guide rail, and a Y-axis auxiliary guide rail. The support legs, right-angle connecting plate, Y-axis guide rail, and Y-axis auxiliary guide rail are all connected to the bottom worktable via threaded connections, providing support for the entire platform. Based on this, the experimental platform for hole position compensation and posture detection described herein includes an experimental platform for hole position compensation on planar workpieces and an experimental platform for posture detection on curved workpieces.

[0006] The experimental platform for hole position compensation for planar workpieces includes an experimental platform body, a hole position compensation module, and a hole position detection module. The hole position compensation module includes a planar experimental workpiece, an X-axis guide rail, and a Z-axis guide rail. The X-axis guide rail is connected to the Y-axis guide rail and the Y-axis auxiliary guide rail through threads. The span between the Y-axis guide rail and the Y-axis auxiliary guide rail is determined by the effective stroke and limit distance of the X-axis guide rail. The Z-axis guide rail is connected to the X-axis guide rail through threads, and the end of the Z-axis guide rail can be installed with a load. The experimental platform has three degrees of freedom of movement in the X, Y, and Z directions, which are used to equate the sequential movement of different holes in the hole position compensation process of the hole making equipment. The hole position detection module includes a visual detection unit and a visual mounting plate. By photographing the reference hole at different positions, the actual position coordinates of the reference hole are obtained. The theoretical digital model is compared with the actual position coordinates in the upper computer software to calculate the error model, and the theoretical position of the hole to be processed is compensated, thereby realizing the verification of the hole position compensation algorithm.

[0007] The experimental platform for posture detection of curved workpieces includes an experimental platform body and a normal detection module. The difference between the experimental platform body of this part and the experimental platform body of the experimental platform for hole position compensation of planar workpieces is that the Y-axis guide rail and the Y-axis auxiliary guide rail in the experimental platform body described in this part mainly realize the feeding action of the device, facing the feeding movement of the equivalent hole-making device in the single-hole preparation process, and the span between the two is determined by the distance between the threaded connection holes at both ends of the bottom of the clamping seat. The normal detection module includes a curved experimental workpiece, a touch table, a movable joint, a signal receiving board, a distance detection unit, and a clamping seat. The touch table at the front end of the normal detection module contacts the curved experimental workpiece, and the signal receiving board will adaptively make angle adjustments. The data collected by the three distance detection units can be processed in the upper computer software to obtain the normal deflection angle, thereby realizing the verification of the posture detection.

[0008] Another aspect of the present invention provides a method for hole position compensation for a planar workpiece and posture detection for a curved workpiece. To better implement the present invention, a method for hole position compensation for a planar workpiece is further established, comprising the following steps:

[0009] S1: Build a twin model of a platform for hole position compensation for planar workpieces and a simulation workstation for a hole position compensation system for aircraft wing boxes. Perform uneven surface hole division on the planar experimental workpiece used in the experimental platform for hole position compensation. A larger number of holes are used as reference holes for fitting and solving the conversion model; a smaller number of holes are used as verification holes to verify the accuracy of the conversion model. The total number of experimental holes and the number of reference holes on the planar experimental workpiece are N and N1, respectively.

[0010] S2: Detect the plane experimental workpiece through the visual inspection unit to obtain the corresponding actual position coordinates of the reference hole. Derive the theoretical numerical model of the reference hole in the twin model of the verification platform, and use the Rodrigues matrix to calculate the conversion model between the theoretical numerical model and the actual coordinates of the reference hole. The conversion model has Use the remaining (N-N1) verification holes to A conversion model was validated;

[0011] S3: If the number of verification holes that exceed the tolerance in the experimental group in step S2 is less than n, the deviation between the actual coordinates and the theoretical coordinates of the verification holes is compared by the RANSAC algorithm. The holes that exceed the threshold will be removed as outliers, and the holes within the threshold will be treated as inliers and saved in the set R. T middle;

[0012] S4: If the number of verification wells that exceed the tolerance in the experimental group in step S2 is not less than n, use the Grubbs statistical method to detect outliers based on the influence factor of the reference wells, remove the reference well with the largest influence factor, and save the remaining reference wells to the set R B ;

[0013] S5: R T and R B The two hole sets are intersected, and the resulting set R is used to recalculate the final transformation model T.

[0014] S6: Correct the position deviation of the verification hole according to the final conversion model T obtained in the above step S5, determine its correction accuracy, repeat the above steps S2 to S5 until the correction accuracy meets the threshold requirement, and further correct the theoretical digital model of the prefabricated hole exported from the MFG file in the hole position compensation system simulation workstation.

[0015] In order to better implement the present invention, further, a method for posture detection of a curved workpiece is established, comprising the following steps:

[0016] R1: Establish a mathematical model for the platform that performs attitude detection on curved workpieces and a simulation workstation for the normal detection system for the local part of the aircraft fuselage. Use the mathematical model to solve the spatial attitude direction of each hole surface verified by the experiment, and use it as the theoretical value before normal detection;

[0017] R2: The host computer drives the normal detection device to perform feed motion until the contact table presses the curved surface experimental workpiece, and the signal receiving board adaptively adjusts the angle. At this time, the three distance detection units can detect three displacement values ​​h1, h2, and h3 with different distances;

[0018] R3: Solve the normal deflection angle θ using the vector method proposed in the present invention. Combined with step R1 above, obtain the theoretical value of the posture before normal detection and the posture correction model. Repeat steps R2 and R3 above until the angle error falls within the allowable error range.

[0019] R4: According to the decomposition principle of robot inverse kinematics proposed in the present invention, the normal deflection angle θ is decomposed into angles α and γ rotating along two orthogonal coordinate axes;

[0020] R5: Use the orthogonal rotation angles α and γ obtained in the above step R4 to correct the theoretical numerical model of the posture before the normal detection of the prefabricated hole in the normal detection system simulation workstation.

[0021] In order to better implement the present invention, further, a simulation analysis of the above error compensation results is performed in a virtual platform, which is characterized by the following steps:

[0022] T1: According to the prefabricated hole theoretical numerical model correction method provided in step S6 and step R5 above, a simulation sequence and path editor are set in the virtual platform;

[0023] T2: Build a simulation workstation for the hole compensation system for the aircraft wing box and a simulation workstation for the normal detection system for the local aircraft fuselage. Define the mechanisms of the drilling end effector and robotic equipment. Define the three operating states of the drilling end effector: clamping seat feed, single hole preparation feed, and retraction. Define the TCP tool of the robotic equipment as the drilling end effector. Define the components of the aircraft wing box and fuselage, and project each hole position corresponding point as a weld point onto the corresponding workpiece surface.

[0024] T3: In the operation mode, a new composite operation is created, the operating state of the drilling end effector device is defined in the new equipment operation, and the process points and prefabricated hole points of the hole position compensation process and the normal detection process are defined in the new welding operation;

[0025] T4: The actual coordinate values ​​and actual posture values ​​after hole position compensation and normal detection are added to the simulation process, and compared with the parameters obtained in the process before hole position compensation and normal detection. When the scene is zoomed in to the highest magnification, it can be seen that there are certain differences before and after hole position compensation and normal detection. By adding a data table to the path editing module, the position and posture before and after hole position compensation and normal detection are quantitatively compared, which verifies the rationality and practicality of this method in the field of precision enhancement of automated intelligent hole-making systems.

[0026] The present invention has the following beneficial effects compared to the prior art:

[0027] 1. The experimental platform for hole position compensation and posture detection described in the present invention can avoid unnecessary errors caused by the superposition of complex transformation matrix relationships, while retaining necessary error sources such as workpiece clamping, and can realize the verification of hole position compensation and posture detection;

[0028] 2. The method described in the present invention for hole position compensation for planar workpieces and posture detection for curved workpieces has a compensation accuracy of 0.13mm for position errors and a correction accuracy of 0.17° for posture errors, both of which meet the hole making accuracy threshold requirements.

[0029] 3. The experimental platform for hole position compensation and posture detection described in the present invention, after verification through the adjustment method for hole position compensation and posture detection described in the present invention, can directly correct and update the offline program of the automated intelligent hole making system, which is of great significance in the field of automated hole making. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 This is a flowchart of an experimental platform for verifying hole position compensation and posture detection proposed by the present invention and serves as an abstract figure;

[0031] Figure 2 This is a twin model diagram of the experimental platform for hole position compensation for planar workpieces established by the present invention;

[0032] Figure 3 This is a twin model diagram of the experimental platform for posture detection of curved surface workpieces established by the present invention;

[0033] Figure 4 This is a schematic diagram of the solution of the normal deflection angle and TCP transformation posture of the present invention;

[0034] Among them: 1. Support feet, 2. Bottom workbench, 3. Right-angle connecting plate, 4. Flat experimental workpiece, 5. Visual inspection unit, 6. Visual mounting plate, 7. Z-axis guide rail, 8. Y-axis guide rail, 9. Y-axis auxiliary guide rail, 10. X-axis guide rail, 11. Curved experimental workpiece, 12. Touch table, 13. Movable joint, 14. Signal receiving board, 15. Distance detection unit, 16. Clamping seat. DETAILED DESCRIPTION

[0035] The present invention is further described below with reference to the accompanying drawings and examples. The present invention includes but is not limited to the following examples.

[0036] Example 1

[0037] An experimental platform for hole position compensation and posture detection, comprising a support leg 1, a bottom worktable 2, a right-angle connecting plate 3, a Y-axis guide rail 8, and a Y-axis auxiliary guide rail 9. The support leg 1, right-angle connecting plate 3, Y-axis guide rail 8, and Y-axis auxiliary guide rail 9 are all connected to the bottom worktable 2 via threaded connections, providing support for the entire platform. Based on this, the experimental platform for hole position compensation and posture detection described herein includes an experimental platform for hole position compensation on planar workpieces and an experimental platform for posture detection on curved workpieces.

[0038] The experimental platform for hole position compensation for planar workpieces includes an experimental platform body, a hole position compensation module, and a hole position detection module. The hole position compensation module includes a planar experimental workpiece 4, a Z-axis guide rail 7, and an X-axis guide rail 10. The X-axis guide rail 10 is connected to the Y-axis guide rail 8 and the Y-axis auxiliary guide rail 9 by threads. The span between the Y-axis guide rail 8 and the Y-axis auxiliary guide rail 9 is determined by the effective stroke and limit distance of the X-axis guide rail 10. The Z-axis guide rail 7 is connected to the X-axis guide rail 10 by threads, and the end of the Z-axis guide rail 7 can be installed with a load. The experimental platform has three degrees of freedom of movement in the X, Y, and Z directions, which are used to equate the sequential movement between different holes of the hole-making equipment during the hole position compensation process. The hole position detection module includes a visual detection unit 5 and a visual mounting plate 6. By detecting the reference hole at different positions, the actual position coordinates of the reference hole are obtained. The theoretical digital model is compared with the actual position coordinates in the host computer software to calculate the error model, and the theoretical position of the hole to be processed is compensated, thereby realizing the verification of the hole position compensation algorithm.

[0039] Another aspect of the present invention provides a method for hole position compensation for a planar workpiece. To better implement the present invention, a method for hole position compensation for a planar workpiece is established in combination with an example. The method is characterized by comprising the following steps:

[0040] S1: Build a platform twin model for hole compensation of planar workpieces and a hole compensation system simulation workstation for aircraft wing boxes;

[0041] S2: The platform twin model is built as follows Figure 2As shown. Its main body is a drive module, and the four-axis three-way gantry structure guide rail table is connected to the bottom workbench 2 by bolts. The outer dimensions of the bottom workbench 2 are 900mm×900mm×13mm, and the four corners are supported by the four support legs 1. The planar experimental workpiece 4 is vertically fixedly connected to the bottom workbench 2 by the two right-angle connecting plates 3. The planar experimental workpiece 4 has an outer dimension of 400mm×400mm×13mm, a surface flatness of 0.1mm, 15×15 holes, a hole diameter of 4mm, and a hole spacing of 25mm. The flatness of the planar experimental workpiece 4 is 0.1mm, and the verticality of the right-angle connecting plate 3 is 0.1°, both of which meet the accuracy threshold for method verification of the present invention. The load end of the Z-axis guide rail 7 is installed with the visual mounting plate 6, and the visual detection unit 5 is installed on the visual mounting plate 6. The visual detection unit 5 is installed horizontally and is used to detect the holes on the planar experimental workpiece 4. Taking into account the motion range of the four-axis three-way gantry structure guide rail platform and the optimal field of view of the visual inspection unit 5, and the short-distance motion of the four-axis three-way gantry structure guide rail platform is not convenient for investigating the motion error term, the central area range of the plane experimental workpiece is selected to be 7×7 holes, and one hole is selected as an experimental hole every two holes, that is, 3×3 holes are used as experimental holes. When conducting the hole position error compensation test, three holes are randomly selected as reference holes, and the remaining six holes are used as verification holes. The conversion model between the actual coordinates and the theoretical digital model is calculated using the reference holes, and the conversion model is verified using the remaining six holes.

[0042] S3: Construct a conversion model between the actual coordinates of the reference hole and the theoretical numerical model, and modify the theoretical value of the prefabricated hole in the hole position compensation system simulation workstation for the aircraft wing box according to the conversion model. As can be seen from the above step S2, nine experimental holes are selected and numbered. The platform twin model verified by the hole position compensation algorithm exports the MFG file, which contains the theoretical numerical model of the nine experimental holes in the world coordinate system, and records the theoretical coordinates of the nine experimental holes in the set T E ;

[0043] S4: In contrast, by building an experimental platform for verifying the hole compensation algorithm for planar workpieces, the actual coordinates of the nine experimental holes in the world coordinate system can be obtained. The specific steps are:

[0044] SX1: Photograph the calibration plate, calibrate the parameters of the visual inspection unit in the HALCON software, and construct a transformation matrix from the visual inspection unit coordinate system to the world coordinate system based on the calibrated parameters. The transformation relationship from the pixel coordinate system to the visual inspection unit coordinate system can be expressed by the following formula:

[0045]

[0046] Where K is the internal parameter matrix of the visual detection unit;

[0047] SX2: Through the host computer control software, debug the four-axis three-way gantry structure guide rail platform so that the field of view of the visual inspection unit is at a more appropriate position of the first experimental hole;

[0048] SX3: Through the host computer control software, the guide rails of each axis are driven to move according to the spacing of the experimental holes until all nine experimental holes are photographed. After the coordinate system conversion operation in the host computer control software, the actual coordinates of the nine experimental holes in the world coordinate system are recorded in the set A. E middle.

[0049] Table 1 Actual pixel coordinates and sizes of reference holes

[0050]

[0051] S5: From the above step S2, we know that the conversion model of the actual coordinates of the reference hole and the theoretical digital model has Select three holes as reference holes and the remaining six holes as verification holes. Use the Rodrigues matrix to solve the conversion model between the actual coordinates of the reference holes and the theoretical numerical model. The specific steps are as follows:

[0052] SY1: Find the center of mass C A 、C B . Find the theoretical coordinate set P of the reference hole A and the actual coordinate set P of the reference hole B The centroid of these two datasets is as follows:

[0053]

[0054] SY2: Singular value decomposition transformation matrix H, calculate the rotation matrix R and translation matrix T. Translate the two data sets to the origin of the same coordinate system, and then calculate the optimal rotation matrix R,

[0055]

[0056]

[0057] The transformation matrix P is constructed from the rotation matrix R and the translation matrix T, then:

[0058]

[0059] S6: Based on the conversion matrix P obtained in step S5, the conversion model is verified using the remaining six holes, and two levels of accuracy are determined:

[0060] SZ1: If the number of verification holes that exceed the tolerance is less than 3, the larger error comes from the verification holes themselves. The deviation between the actual coordinates and the theoretical coordinates of the verification holes is compared by the RANSAC algorithm. The holes that exceed the threshold are removed as outliers, and the holes within the threshold are regarded as inliers and saved in the set R. T The formula is as follows:

[0061]

[0062]

[0063] Among them, Q A To verify the theoretical coordinates of the hole, Q B To verify the actual coordinates of the hole, Q' B The actual coordinates of the verification hole are obtained by inversely calculating the transformation matrix P;

[0064] SZ2: If the number of verification holes that are out of tolerance in this group is not less than 3, the larger error comes from the three reference holes used to construct the transformation matrix P. By using the Grubbs statistical method, outliers are detected based on the influence factors of the reference holes, and the reference hole with the largest influence factor is removed and the remaining reference holes are saved in the set R B .

[0065] SZ3: According to the above steps SZ1 and SZ2, the deviation between reference hole 1 and reference hole 8 is large, which has a serious impact on the results. Therefore, reference hole 1 and reference hole 8 are eliminated, and the remaining 7 holes are reordered and experiments and conversion matrix calculations are carried out.

[0066] Table 2 Actual coordinates of the reference hole in the product coordinate system

[0067]

[0068] S7: R T and R B The intersection of the two hole sets is taken, and the final conversion model T is recalculated using the intersection set R. The offline program is derived from the hole position compensation system simulation workstation for the aircraft wing box. The offline program contains the prefabricated hole theoretical model. The prefabricated hole theoretical model is compensated according to the final conversion model T, and the position compensation accuracy reaches 0.13mm. The conversion matrix T is

[0069]

[0070] Example 2

[0071] The experimental platform for posture detection of curved workpieces includes an experimental platform body and a normal detection module. The difference between the experimental platform body of this part and the experimental platform body of the experimental platform for hole position compensation of planar workpieces is that the Y-axis guide rail 8 and the Y-axis auxiliary guide rail 9 in the experimental platform body described in this part mainly realize the feeding action of the device, facing the feeding movement of the equivalent hole-making device in the single-hole preparation process, and the span between the two is determined by the distance between the threaded connection holes at both ends of the bottom of the clamping seat 16. The normal detection module includes a curved experimental workpiece 11, a touch table 12, a movable joint 13, a signal receiving board 14, a distance detection unit 15, and a clamping seat 16. The touch table 12 at the front end of the normal detection module contacts the curved experimental workpiece 11, and the signal receiving board 14 will adaptively make angle adjustments. The data collected by the three distance detection units 15 can be processed in the upper computer software to obtain the normal deflection angle, thereby realizing the verification of the posture detection.

[0072] Another aspect of the present invention provides a method for posture detection of a curved workpiece. To better implement the present invention, a method for posture detection of a curved workpiece is established in conjunction with an example. The method is characterized by comprising the following steps:

[0073] R1: Build a mathematical model for a platform that performs attitude detection on curved workpieces and a simulation workstation for a normal detection system for a local part of an aircraft fuselage. Use this mathematical model to solve the spatial attitude direction of each experimentally verified hole surface, which serves as the theoretical value before normal detection.

[0074] R2: The mathematical model is as follows Figure 3 As shown. The present invention selects the Y-axis guide rail 8 and the Y-axis auxiliary guide rail 9 as the mobile platform of the clamping seat 16, and installs the clamping seat 16 on the slider of the Y-axis guide rail 8 and the Y-axis auxiliary guide rail 9 by bolt connection. A positioning edge is designed on one side of the bottom mounting surface of the clamping seat 16 to ensure the positioning accuracy during installation. The specifications of the curved surface experimental workpiece 11 are 300mm×260mm×10mm, and the curvature of the curved surface part is The number of holes is 6×6, the hole diameter is 12mm, the vertical hole spacing is 40mm, and the horizontal hole angle is 5°. Three distance detection units are installed on the pressing seat 16. The plane formed by the signal emission points of the three distance detection units is parallel to the signal receiving board 14. The installation layout of the three distance detection units adopts the intersection of the tool axis, i.e., the axis of the inner hole of the pressing seat 16 and its signal emission points, as the center of the circle, and is evenly distributed at 60 degrees on the pressing seat 16. The signal receiving board 14 is installed in the inner hole of the pressing seat 16 through the movable joint 13, so that the floating signal receiving board unit can adaptively adjust the angle when the pressing seat 16 is fed to the touch table 12 to press the curved experimental workpiece 11. The outer circumference diameter of the touch table 12 is 20 mm, which is convenient for pressing the holes on the curved experimental workpiece 11 in a small range. Since the core part of the posture detection verification method involved in the present invention is that the three distance detection units transmit detection signals to the signal receiving board 14, it is necessary to ensure that the surface of the signal receiving board 14 receiving the signal has a high precision.

[0075] R3: Use the vector method to solve the normal deflection angle θ. The host computer control software drives the pressing seat 16 to perform a feed motion until the contact table 12 presses the curved experimental workpiece 11. The signal receiving board 14 will adaptively adjust the angle. At this time, the three distance detection units 15 can detect three different distance displacement values ​​h1, h2, and h3. The host computer control software solves the normal deflection angle θ according to the vector method proposed by the present invention.

[0076] like Figure 4 , all laser beams are emitted parallel to the tool axis, where ΔA'B'C' represents the triangular area formed by the three distance detection unit signal emission points, and ΔABC represents the triangular area formed by the three signal points on the plane of the signal receiving plate after adapting to the workpiece surface. A Cartesian coordinate system is established with point O, the center of ΔA'B'C'. The Z-axis direction is perpendicular to ΔABC and points to ΔA'B'C'. The X-axis direction is parallel to C'B', then the Y-axis direction can be determined by the right-hand rule. In the coordinate system {O'}, O'H is perpendicular to ΔA'B'C', which represents the axial direction of the tool at this time. MH is perpendicular to OF, NH is perpendicular to OG, and MH coincides with ON after ΔA'B'C' rotates around the Y axis by an angle of γ. On this basis, NH coincides with OM after rotating around the X axis by an angle of α. Then the normal direction of ΔA'B'C' coincides with ΔABC at this time. The solution formula for the normal deflection angle θ is as follows:

[0077]

[0078] Wherein (l1, l2, l3) represents the normal detection vector coordinates of the signal receiving board 14 after adaptive adjustment, h1, h2, h3 represent the distance values ​​measured by the three distance detection units, and a represents the installation spacing of the three distance detection units;

[0079] R4: According to the decomposition principle of robot inverse kinematics, the normal deflection angle θ is equivalent to the angles α and γ rotating around two orthogonal coordinate axes:

[0080]

[0081] Among them, E(x0, y0, z0) represents the posture of the device end before normal detection, and E(x, y, z) represents the posture of the device end after normal detection.

[0082] The above matrix is ​​already known, and we need to solve (x, y, z) in E(x, y, z) as follows:

[0083] RX1: When cosy>0,

[0084]

[0085] RX2: When cosy < 0,

[0086]

[0087] RX3: When cosy=0,

[0088] If m z =-1, then we have

[0089]

[0090] If m z =1, then we have

[0091]

[0092] R5: Use the orthogonal rotation angles α and γ obtained in step R4 above to correct the theoretical numerical model of the prefabricated hole posture in the normal detection system simulation workstation. The average value of the correction angle is 0.17°.

[0093] Table 3 Actual normal vector and correction angle of prefabricated hole posture

[0094]

[0095] In order to better implement the present invention, further, the above results are simulated and analyzed in a virtual platform in combination with examples, which is characterized by the following steps:

[0096] T1: According to the prefabricated hole theoretical numerical model correction method provided in step S7 and step R5 above, a simulation sequence and path editor are set in the virtual platform;

[0097] T2: Build a simulation workstation for the hole compensation system for the aircraft wing box and a simulation workstation for the normal detection system for the local aircraft fuselage. Define the mechanisms of the drilling end effector and robotic equipment, define the three operating states of the drilling end effector: clamping seat feed, single hole preparation feed, and retraction, and define the TCP tool of the robotic equipment as the drilling end effector. Define the components of the aircraft wing box and fuselage, and project each hole position corresponding point as a weld point onto the corresponding workpiece surface.

[0098] T3: In the operation mode, a new composite operation is created, the operating state of the drilling end effector device is defined in the new equipment operation, and the process points and prefabricated hole points of the hole position compensation process and the normal detection process are defined in the new welding operation;

[0099] T4: The actual coordinate values ​​and actual posture values ​​after hole position compensation and normal detection are added to the simulation process, and compared with the parameters obtained in the process before hole position compensation and normal detection. When the scene is zoomed in to the highest magnification, it can be seen that there are certain differences before and after hole position compensation and normal detection. By adding a data table to the path editing module, the position and posture before and after hole position compensation and normal detection are quantitatively compared, which verifies the rationality and practicality of this method in the field of precision enhancement of automated intelligent hole-making systems.

[0100] Through the above steps, it can be concluded that the experimental platform for verifying hole position compensation and posture detection described in the present invention can avoid unnecessary errors caused by the superposition of complex transformation matrix relationships, retain necessary error sources such as workpiece clamping, and improve error compensation efficiency. After verification by specific implementation cases, based on the experimental platform and method proposed by the present invention, the compensation accuracy of position error reached 0.13mm, and the correction of posture error reached 0.17°, both meeting the hole-making accuracy threshold requirements and enabling verification of the proposed hole position compensation and posture detection.

[0101] The above are merely preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. All technical solutions that fall within the scope of protection of the present invention are within the scope of protection of the present invention. Any modifications and improvements made within the spirit and principles of the present invention shall be deemed to be within the scope of protection of the present invention.

Claims

1. An experimental platform and adjustment method for hole position compensation and posture detection, wherein the main body of the experimental platform is composed of a support leg (1), a bottom workbench (2), a right-angle connecting plate (3), a Y-axis guide rail (8) and a Y-axis auxiliary guide rail (9); the experimental platform for hole position compensation and posture detection respectively relates to an experimental platform for hole position compensation for a planar workpiece and an experimental platform for posture detection for a curved workpiece; The experimental platform for performing hole position compensation on a planar workpiece is characterized by: The invention comprises an experimental platform body, a hole position detection module and a hole position compensation module. The hole position detection module comprises a visual detection unit (5) and a visual mounting plate (6) for obtaining the actual coordinates of the hole position; the hole position compensation module comprises a plane experimental workpiece (4), a Z-axis guide rail (7) and an X-axis guide rail (10). The end of the Z-axis guide rail (7) can be mounted with a load. The experimental platform has the degrees of freedom of movement in three directions of X, Y and Z and is used to achieve the sequential movement between different holes of the equivalent hole making equipment during the hole position compensation process. By processing the theoretical numerical model of the hole position and the actual coordinates, the theoretical numerical model is corrected by using a conversion matrix to achieve compensation of the hole position. The experimental platform for performing posture detection on a curved surface workpiece comprises an experimental platform body and a normal detection module. The normal detection module comprises a curved surface experimental workpiece (11), a touch table (12), a movable joint (13), a signal receiving board (14), a distance detection unit (15), and a pressing seat (16). The touch table (12) at the front end of the normal detection module contacts the curved surface experimental workpiece (11), and the signal receiving board (14) adaptively adjusts the angle. The data collected by the three distance detection units (15) are processed in the upper computer software to obtain the normal deflection angle, thereby realizing the verification of the posture detection.

2. The experimental platform and adjustment method for hole position compensation and posture detection according to claim 1, characterized in that The method for performing hole position compensation on a planar workpiece includes the following steps: S1: Debugging the guide rail movement so that the visual inspection unit (5) has an optimal field of view, selecting one hole every two holes as an experimental hole, and when performing a hole position error compensation test, randomly selecting three holes from the N holes as reference holes, and the remaining (N-3) holes as verification holes, using the reference holes to calculate the conversion model between the actual coordinates and the theoretical digital model, and using the remaining (N-3) holes to verify the conversion model; S2: Construct a conversion model between the actual coordinates of the reference hole and the theoretical numerical model, and modify the theoretical value of the prefabricated hole in the hole position compensation system simulation workstation for the aircraft wing box according to the conversion model. As can be seen from the above step S1, N experimental holes are selected and numbered, and the MFG file is exported through the platform twin model of the hole position compensation, which contains the theoretical numerical model of the N experimental holes in the world coordinate system. The theoretical coordinates of the N experimental holes are recorded in the set T E ; S3: By building an experimental platform for hole position compensation for planar workpieces, the actual coordinates of N experimental holes in the world coordinate system can be obtained; S4: From the above step S1, we know that the conversion model of the actual coordinates of the reference hole and the theoretical digital model has Select three holes as reference holes and the remaining (N-3) holes as verification holes, and solve the conversion model P between the actual coordinates of the reference holes and the theoretical numerical model; S5: Based on the transformation matrix P obtained in step S4, the remaining (N-3) holes are used to verify the transformation model and perform two-level accuracy judgment; S6: R T and R B The intersection of the two hole sets is taken, and the set R after the intersection is used to recalculate the final conversion model T. An offline program is exported from the hole position compensation system simulation workstation for the aircraft wing box. The offline program includes a prefabricated hole theoretical model, and the prefabricated hole theoretical model is compensated according to the final conversion model T.

3. The experimental platform and adjustment method for hole position compensation and posture detection according to claim 2, characterized in that The method for solving the conversion model between the actual coordinates of the reference hole and the theoretical digital model described in step S4 of the method for hole position compensation for a planar workpiece has the following specific steps: SX1: Find the theoretical coordinate set P of the reference hole A and the actual coordinate set P of the reference hole B The centroid C of these two data sets A 、C B ; SX2: Translate the two data sets to the origin of the same coordinate system, and then find the optimal rotation matrix R and translation matrix T based on singular value decomposition, and construct the transformation matrix P from the rotation matrix R and translation matrix T.

4. The experimental platform and adjustment method for hole position compensation and posture detection according to claim 2, characterized in that The method for determining the accuracy of the conversion model at two levels described in step S5 of the method for hole position compensation for a planar workpiece comprises the following specific steps: SY1: If the number of verification holes that are out of tolerance is less than 3, the larger error comes from the verification holes themselves. The deviation between the actual coordinates and the theoretical coordinates of the holes is compared and verified by the RANSAC algorithm. The holes exceeding the threshold are removed as outliers, and the holes within the threshold are regarded as inliers and saved in the set R. T middle; SY2: If the number of verification holes that are out of tolerance is not less than 3, the larger error comes from the three reference holes used to construct the transformation matrix P. The outliers are detected based on the influence factor of the reference holes using the Grubbs statistical method. The reference holes with the largest influence factor are removed and the remaining reference holes are saved in the set R. B .

5. The experimental platform and adjustment method for hole position compensation and posture detection according to claim 1, characterized in that A method for posture detection of curved workpieces. The following steps are involved: R1: drives the pressing seat (16) to perform feeding motion until the contact table (12) presses the curved surface experimental workpiece (11), and the signal receiving board (14) will adaptively adjust the angle. At this time, the three distance detection units (15) can detect three displacement values ​​h1, h2, and h3 with different distances; R2: Calculate the normal deflection angle θ using the vector method proposed in the present invention. Combine the theoretical value of the posture before normal detection with the posture correction model and repeat steps R1 and R2 until the angle error falls within the allowable error range. R3: According to the decomposition principle of robot inverse kinematics, the normal deflection angle θ is equivalently decomposed into the angles α and γ rotating around two orthogonal coordinate axes; R4: Use the orthogonal rotation angles α and γ obtained in the above step R3 to correct the theoretical numerical model of the posture before the normal detection of the prefabricated hole in the normal detection system.

Images