A device and method for testing the surface magnetic field distribution of a permanent magnet

Abstract

The application relates to the technical field of magnetic material detection, and discloses a permanent magnet surface magnetic field distribution testing device and method, which comprises a workbench, a staggered separation feeding mechanism, a magnetic flux detection mechanism and a discharging mechanism arranged in sequence, and a carrying mechanism arranged above. The magnetic flux detection mechanism is provided with a shell with a sliding cover, and a standard reference magnet is fixed in the shell; and the carrying mechanism mechanical hand is provided with a measurement execution end integrated with a Hall piece and a laser displacement sensor. During testing, the sliding cover is closed to form a shielding space; reference data is extracted based on the standard reference magnet; the laser displacement sensor is used to measure the distance to solve the workpiece surface inclination angle, dynamically compensate the mechanical hand posture and height, and make the Hall piece keep a set lift-off distance scanning, and the surface magnetic distribution data is generated after algorithm compensation. The application excludes external magnetic field interference, realizes profile adaptive following of the probe, and improves the measurement precision and consistency.

Description

Technical Field

[0001] This invention relates to the field of magnetic material testing technology, and in particular to a device and method for testing the magnetic field distribution on the surface of a permanent magnet. Background Technology

[0002] Permanent magnets are widely used in electronic devices such as motors and sensors, and the uniformity of their surface magnetic field distribution directly affects the performance of the end product. Currently, the surface magnetic field distribution of permanent magnets is typically measured using a multi-axis motion platform equipped with a Hall probe.

[0003] In actual testing, the production workshop or testing environment commonly contains the Earth's magnetic field and stray magnetic fields generated by the operation of surrounding electrical equipment. Because existing surface magnetic field testing equipment often uses an open-bench structure, external environmental magnetic fields directly enter the testing area and are superimposed on the measurement signal, causing distortion in the magnetic field data acquired by the probe. Simultaneously, due to limitations in manufacturing processes, permanent magnet surfaces inevitably have contour undulations, minor deformations, or placement tilt tolerances. Existing scanning mechanisms typically drive the probe along a preset absolute horizontal trajectory, failing to sense and adapt to changes in the physical morphology of the workpiece surface. Since Hall effect probes are extremely sensitive to lift-off distance and detection angle, unevenness on the workpiece surface causes the actual distance between the probe and the surface to continuously change, introducing significant measurement errors.

[0004] In addition, the hardware of the test system is prone to zero-point drift after long-term operation, and existing equipment usually lacks dynamic closed-loop calibration and data compensation mechanisms for ambient temperature fluctuations and slight differences in probe height during the test process, resulting in poor measurement accuracy of the final magnetic distribution data and poor consistency between products. Summary of the Invention

[0005] The purpose of this invention is to provide a device and method for testing the magnetic field distribution on the surface of permanent magnets, which solves the problems of existing testing equipment having difficulty shielding external environmental magnetic field interference when scanning the magnetic field on the surface of permanent magnets, and the detection probe being unable to dynamically and adaptively adjust its posture according to the workpiece surface contour, resulting in low accuracy and consistency of magnetic field distribution data measurement.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] The first aspect of the present invention provides a permanent magnet surface magnetic field distribution testing device, including a frame and a worktable. A touch control screen is installed on one outer wall of the worktable. A staggered separation feeding mechanism, a conveying mechanism, a magnetic flux detection mechanism and a unloading mechanism are arranged sequentially on the upper surface of the worktable along the conveying direction of the workpiece.

[0008] The transport mechanism is equipped with a measurement execution end;

[0009] The magnetic flux detection mechanism includes a housing and a support plate. The support plate inside the housing is provided with a workpiece support position and a calibration area. A standard reference magnet is fixedly installed in the calibration area, and an openable high-permeability shielding sheet is provided above the standard reference magnet.

[0010] The testing device also includes an automated loading and unloading module that is communicatively connected to the misaligned separation loading mechanism, the conveying mechanism and the unloading mechanism, an electromagnetic shielding control module that is communicatively connected to the magnetic flux detection mechanism, and an adaptive testing module that is communicatively connected to the touch control screen and the measurement execution end respectively.

[0011] The adaptive testing module is used to process spatial coordinate data, calculate motion compensation, and generate the final measurement results.

[0012] The adaptive testing module works in conjunction with various mechanisms to integrate automatic feeding, online self-calibration, attitude compensation scanning, and sorting and unloading, thereby improving the equipment's continuous operation capability and measurement accuracy.

[0013] Furthermore, the conveying mechanism includes: a gantry frame fixedly mounted on the worktable; and a motor connected to an X-axis linear transmission assembly mounted on the gantry frame, for driving robot arm 1, robot arm 2, and robot arm 3 to move horizontally.

[0014] The No. 1 robotic arm and the No. 3 robotic arm are respectively driven to lift by corresponding cylinders. The lifting end of the No. 1 robotic arm is detachably connected to a suction cup through a buckle one, and the lifting end of the No. 3 robotic arm is detachably connected to a suction cup two through a buckle two.

[0015] The second manipulator includes a Y-axis linear transmission component connected to the moving end of the X-axis linear transmission component, a Z-axis drive mechanism and an attitude adjustment component disposed at the moving end of the Y-axis linear transmission component, and the attitude adjustment component includes an A-axis drive mechanism and a B-axis drive mechanism arranged orthogonally.

[0016] A laser displacement sensor and a Hall plate are fixedly installed side by side at the end of the attitude adjustment assembly. The measuring optical axis of the laser displacement sensor and the measuring central axis of the Hall plate are kept parallel by a known offset distance, and there is a pre-calibrated hardware installation height difference between the two in vertical height.

[0017] A temperature sensor is mounted close to the Hall plate, and its sensing end is fixedly connected to the substrate surface of the Hall plate by thermally conductive adhesive.

[0018] Furthermore, the top of the housing is provided with an opening for the workpiece and the measuring end of the No. 2 robotic arm to enter and exit, and the magnetic flux detection mechanism also includes a sliding cover that is slidably connected to the guide structure on the top of the housing;

[0019] A two-dimensional overlapping magnetically conductive slide plate is stacked on top of the sliding cover and supported by a slide frame connected to the X-axis and Y-axis moving ends of the No. 2 robotic arm. The lifting end of the No. 2 robotic arm extends into the interior of the outer shell through the central hole on the two-dimensional overlapping magnetically conductive slide plate.

[0020] A drive cylinder fixed to the side of the housing for driving the sliding cover to open or close the opening; and a sealing strip disposed around the opening and at the overlap position between the sliding cover and the housing.

[0021] Furthermore, the staggered separation feeding mechanism includes: a base plate horizontally fixedly connected above the workbench; a feeding slide vertically installed on the front end of the upper surface of the base plate; a fine-tuning screw that is drivenly connected to the adjustable side plate of the feeding slide; and a hydraulic cylinder fixedly installed on the base plate.

[0022] A push rod fixedly connected to the power output end of a hydraulic cylinder, the front end of which is horizontally aligned with the discharge port at the bottom of the feeding slide; and a baffle assembly located at the bottom of the feeding slide, including a lateral blocking cylinder and a non-magnetic positioning block connected to its output end.

[0023] Furthermore, the feeding mechanism includes: a support plate; a main body fixed to the side of the support plate; a drive module mounted on the main body; a positioning pin for positioning the material clamp; an ejector pin connected to the moving end of the drive module; and a slide rail with an outlet facing the defective product tray.

[0024] A second aspect of the present invention provides a method for testing the magnetic field distribution on the surface of a permanent magnet, applied to the aforementioned testing apparatus, comprising the following steps:

[0025] The adaptive testing module drives the No. 2 robotic arm to move above the standard reference magnet to sample the reference data and calculate the dynamic correction coefficient of the Hall plate.

[0026] The automated loading and unloading module controls the misaligned separation loading mechanism to separate the workpiece and push it into the gripping position;

[0027] The automated loading and unloading module uses a suction cup to pick up the workpiece and transfer it to the workpiece bearing position on the carrier plate;

[0028] The electromagnetic shielding control module controls the drive cylinder to move the sliding cover to close and fit the sealing strip, forming an electromagnetic shielding space.

[0029] The adaptive testing module acquires the three-dimensional coordinate data of the workpiece surface, calculates the normal vector and distance deviation of the workpiece surface, and controls the No. 2 robot to perform spatial posture and height compensation.

[0030] The adaptive test module drives the Hall plate to perform magnetic field scanning at a set lift-off distance along a preset three-dimensional trajectory, reads the original sampling data, substitutes it into the dynamic correction coefficient, performs multiplication, and generates surface magnetic distribution data.

[0031] The automated loading and unloading module controls the No. 3 robotic arm to transfer the tested workpieces to the unloading mechanism for sorting and unloading.

[0032] The above testing method calculates a dynamic correction coefficient using a standard reference magnet before scanning to eliminate the zero-point drift error of the sensor. During scanning, the algorithm vector is solved based on the three-dimensional point cloud coordinates of the workpiece surface, and the deflection angle and Z-axis height of the Hall plate are adjusted synchronously accordingly to ensure that the measuring end is always perpendicular and equidistant from the surface of the measured point, thereby reducing the magnetic flux density measurement deviation caused by probe posture deviation.

[0033] Further, calculating the dynamic correction coefficient includes: obtaining the original sampled value of the reference magnetic field on the surface of the standard reference magnet; and retrieving the pre-written theoretical reference magnetic flux density value.

[0034] When the absolute value of the original sampled value is determined to be greater than or equal to the preset zero-point noise threshold, the theoretical reference magnetic induction intensity value is used as the target reference quantity, and the original sampled value is used as the current feedback quantity. The dynamic correction coefficient is calculated by dividing the two values.

[0035] When the calculated dynamic correction coefficient falls within the preset effective tolerance range, it is written into the system dynamic compensation register, and the current temperature is recorded as the calibration reference temperature.

[0036] Furthermore, the calculation of the deviation between the normal vector and the distance includes: extracting multiple consecutive distance measurement data using a laser displacement sensor, removing extreme values ​​that exceed three times the standard deviation, and calculating the arithmetic mean, which is then used as the final effective distance value;

[0037] Read the absolute height of the Z-axis encoder of the current No. 2 robot, retrieve the pre-stored hardware installation height difference, call the coordinate calculation model to calculate the height coordinates of the point on the workpiece surface, and generate three effective non-collinear spatial point cloud coordinates.

[0038] Two coplanar space vectors are constructed in this way, and the original normal vector is calculated by the cross product of three-dimensional vectors. The sign of the Z-axis component of the original normal vector is checked, and each component of the original normal vector is divided by the modulus to obtain the unit space normal vector. Clamp protection is applied to each component, and the roll angle and pitch angle are calculated by inverse trigonometric functions.

[0039] Furthermore, attitude and height compensation includes: calling the kinematic matrix compensation algorithm, calculating the attitude and height compensation value based on the fixed structural offset distance parameters and the current roll and pitch angles; and calculating the final target coordinates on the Z-axis based on the height coordinates, attitude and height compensation value, and target lift-off distance, and driving the Hall plate to move downwards.

[0040] The dynamic height deviation is obtained by subtracting the current relative physical distance from the target lift-off distance; when it is within the safe threshold, it is input to the PID closed-loop controller to calculate the dynamic compensation pulse command for the Z-axis.

[0041] The local normal vector at the current scanning node is obtained to control the A-axis and B-axis to adjust the direction of the Hall plate, while the Z-axis is controlled to correct the lift-off distance between the Hall plate and the workpiece surface.

[0042] Furthermore, generating the surface magnetic distribution data includes multiplying the latched original magnetic flux density value by a dynamic correction coefficient to obtain the online calibrated magnetic flux density value;

[0043] The current operating temperature is obtained through the built-in temperature measurement and conditioning circuit. When it is within the effective range, the difference between the temperature and the calibrated reference temperature is used to obtain the temperature deviation. The online calibration magnetic flux density value is compensated by combining the pre-programmed temperature comprehensive compensation coefficient and the temperature-compensated magnetic flux density is calculated. The residual height deviation is extracted. When it is within the tolerance threshold, the lift-off distance-magnetic induction intensity compensation model is called to perform height compensation on the current magnetic flux density value.

[0044] Calculate the first, second, and third powers of the height-compensated magnetic flux density and multiply them by the corresponding nonlinear weighting coefficients. Sum the product results with the zero-order bias constant to reconstruct the final standard magnetic flux density and store it in the three-dimensional coordinate mapping matrix.

[0045] In summary, the present invention has at least one of the following beneficial technical effects:

[0046] 1. This invention creates a relatively enclosed electromagnetic shielding space at the testing station by setting up a shell with a sliding cover, a two-dimensional overlapping magnetic sliding plate, and a composite sealing strip, which blocks the interference of stray magnetic fields from the external environment on the scanning process; at the same time, it works with a standard reference magnet fixed on the carrier plate to extract reference data to eliminate system zero drift and ensure the physical accuracy of the magnetic field sampling data.

[0047] 2. This invention integrates a Hall effect sensor and a laser displacement sensor side-by-side at the end effector of a robotic arm. By acquiring the point cloud coordinates of the workpiece surface through laser ranging and calculating the normal vector and tilt Euler angles, the robotic arm can be dynamically compensated for its attitude and height. This method enables the Hall effect sensor to adapt to the workpiece surface contour and maintain a stable lift-off distance during scanning, thus solving the detection errors caused by the unevenness of the permanent magnet surface.

[0048] 3. In the data processing stage, this invention introduces a multi-step algorithm process that includes online benchmark calibration, working temperature deviation correction, residual height compensation, and nonlinear weight reconstruction. This eliminates the coupling interference of ambient temperature fluctuations and slight differences in physical spacing on the magnetic field signal from the software level, further improving the consistency and reliability of the final output surface magnetic distribution data. Attached Figure Description

[0049] Figure 1 This is a schematic diagram of the external structure of the entire machine of the present invention;

[0050] Figure 2 This is a schematic diagram of the internal structure of the entire machine of the present invention;

[0051] Figure 3 This is a schematic diagram of the internal structure of the staggered separation feeding mechanism of the present invention;

[0052] Figure 4 This is a schematic diagram of the internal structure of the conveying mechanism of the present invention;

[0053] Figure 5 This is a schematic diagram of the internal structure of the magnetic flux detection mechanism of the present invention;

[0054] Figure 6 This is a schematic diagram of the internal structure of the feeding mechanism of the present invention;

[0055] Figure 7 This is a flowchart of the method of the present invention.

[0056] Among them, 1. rack; 2. cabinet door; 3. casters; 4. alarm light;

[0057] 5. Offset separation feeding mechanism; 51. Base plate; 52. Hydraulic cylinder; 53. Push rod; 54. Feeding slide; 55. Workpiece; 56. Fine adjustment screw;

[0058] 6. Handling mechanism; 61. Gantry frame; 62. Motor; 63. Robotic arm No. 1; 631. Buckle 1; 632. Suction cup 1; 64. Robotic arm No. 2; 641. Hall effect sensor; 642. Laser displacement sensor; 65. Robotic arm No. 3; 651. Buckle 2; 652. Suction cup 2; 66. Cylinder;

[0059] 7. Magnetic flux detection mechanism; 71. Housing; 72. Drive cylinder; 73. Sealing strip; 74. Standard reference magnet; 75. Sliding cover;

[0060] 8. Feeding mechanism; 81. Support plate; 82. Main body; 83. Drive module; 84. Positioning pin; 85. Ejector pin; 86. Slide rail; 87. Defective product tray;

[0061] 9. Work surface; 10. Touch control screen. Detailed Implementation

[0062] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0063] See attached document Figure 1 To be continued Figure 6 The present invention provides a device for testing the magnetic field distribution on the surface of a permanent magnet, which may include: a frame 1, a worktable 9, and an electromechanical actuator and control module disposed on the worktable 9.

[0064] The bottom of the frame 1 is equipped with casters 3, and the outer side of the frame 1 is equipped with a cabinet door 2 and an alarm light 4. The workbench 9 is fixed to the top of the frame 1, and a touch control screen 10 is installed on one outer wall of the workbench 9.

[0065] The upper surface of the worktable 9 is arranged in sequence with a staggered separation feeding mechanism 5, a conveying mechanism 6, a magnetic flux detection mechanism 7, and a unloading mechanism 8 along the conveying direction of the workpiece 55.

[0066] The staggered separation feeding mechanism 5 is located at the upstream feeding end. The magnetic flux detection mechanism 7 is located at the middle testing station of the worktable 9. The unloading mechanism 8 is located at the downstream receiving end. The conveying mechanism 6 is mounted above the staggered separation feeding mechanism 5, the magnetic flux detection mechanism 7, and the unloading mechanism 8.

[0067] In terms of electrical control architecture, the testing device includes an automated loading and unloading module, an electromagnetic shielding control module, and an adaptive testing module.

[0068] The automated loading and unloading module is connected to the staggered separation loading mechanism 5, the conveying mechanism 6, and the unloading mechanism 8, and is responsible for controlling the power execution and motion coordination of each mechanical structure.

[0069] The electromagnetic shielding control module is connected to the magnetic flux detection mechanism 7 and is responsible for controlling the opening and closing of the electromagnetic shielding cavity at the test station.

[0070] The adaptive testing module is connected to the measurement execution end in the touch control screen 10 and the conveying mechanism 6 respectively, and is responsible for processing spatial coordinate data, calculating motion compensation amount and generating final measurement results.

[0071] See attached document Figure 7 This invention provides a method for testing the magnetic field distribution on the surface of a permanent magnet, comprising the following steps:

[0072] S100, after the device is powered on or the set calibration cycle is reached, the adaptive test module executes the self-calibration program, drives the No. 2 robot arm 64 inside the transport mechanism 6 to move to the standard reference magnet 74 inside the magnetic flux detection mechanism 7 to sample the reference data, and calculates the dynamic correction coefficient of the Hall plate 641 to complete the calibration of the probe sensitivity.

[0073] S200, the automated loading and unloading module controls the operation of the staggered separation loading mechanism 5 to separate the batch of workpieces 55 one by one, and push the separated workpieces 55 into the preset gripping position.

[0074] S300, the automated loading and unloading module controls the transport mechanism 6 to move horizontally. The No. 1 robot arm 63 inside the transport mechanism 6 moves downward and uses the suction cup 632 to pick up the workpiece 55 at the misaligned loading mechanism 5 and transfer the workpiece 55 to the workpiece bearing position on the bearing plate inside the magnetic flux detection mechanism 7.

[0075] S400, after the workpiece 55 is in place at the workpiece bearing position on the bearing plate, the electromagnetic shielding control module controls the drive cylinder 72 to move, driving the sliding cover 75 to move horizontally along the top of the outer shell 71 to the closed position, and making the sliding cover 75 fit with the sealing strip 73 to form a relatively closed electromagnetic shielding space.

[0076] Subsequently, the adaptive testing module acquires the three-dimensional coordinate data of the surface of workpiece 55, calculates the normal vector and distance deviation of the surface of workpiece 55, and controls the No. 2 robot arm 64 to perform spatial posture and height compensation.

[0077] S500, the adaptive test module drives the Hall plate 641 at the end of the No. 2 robot 64 to perform magnetic field scanning along the surface of the workpiece 55 at a set lifting distance according to the preset three-dimensional trajectory, read the original sampling data, substitute it into the dynamic correction coefficient calculated in step S100 for multiplication, calculate and generate the final surface magnetic distribution data.

[0078] S600, after data acquisition is completed, robot arm 2 64 exits the test position, and the electromagnetic shielding control module controls the drive cylinder 72 to drive the sliding cover 75 to move horizontally to the open position; the automated loading and unloading module controls robot arm 3 65 to descend, and uses suction cup 2 652 to pick up the tested workpiece 55 and transfer it to the unloading mechanism 8; qualified workpieces are loaded into the material clamp by the unloading mechanism 8 in sequence, and unqualified workpieces fall into the defective product tray 87 through the slide 86, completing a single test process.

[0079] See attached document Figure 3 The staggered separation feeding mechanism 5 is used to separate batches of workpieces 55 into individual pieces so that they can be picked up by the robot arm at the downstream station.

[0080] The staggered separation feeding mechanism 5 includes a base plate 51. The base plate 51 is horizontally fixed above the worktable 9. A feeding slide 54 is vertically mounted on the front end of the upper surface of the base plate 51. A through guide channel is formed inside the feeding slide 54, and the cross-sectional shape of the channel is clearance-fitted with the horizontal cross-sectional profile of the workpiece 55. Batches of workpieces 55 are stacked from top to bottom and accommodated in the guide channel of the feeding slide 54, and are fed downwards by the gravity of the workpieces 55 themselves.

[0081] A linear drive source is fixedly mounted on the rear end of the upper surface of the base plate 51. In some embodiments, this linear drive source is a hydraulic cylinder 52 to provide smooth thrust. A push rod 53 is fixedly connected to the power output end of the hydraulic cylinder 52. The push rod 53 is arranged along the horizontal longitudinal axis of the base plate 51. The front end face of the push rod 53 is horizontally aligned with the bottom discharge port of the feeding chute 54.

[0082] A fine-tuning screw 56 is installed on one side of the feeding slide 54, and the fine-tuning screw 56 is connected to the adjustable side plate of the feeding slide 54. Rotating the fine-tuning screw 56 can adjust the effective width or limit dimension of the guide channel, so that the guide channel maintains a clearance fit with the outer contour of the workpieces 55 of different specifications, thereby reducing the shaking or jamming of the workpieces 55 during the falling process, and keeping the movement direction of the bottom workpieces 55 aligned with that of the push rod 53, so that the push rod 53 can push out the bottom workpieces 55 one by one.

[0083] The adaptive test module provided by this invention includes the alignment control process of the reference magnetic source when executing the self-calibration program.

[0084] In actual operation, the device establishes a unified global three-dimensional absolute coordinate system based on various electromechanical actuators. Specifically, the origin of the three-dimensional coordinate system is established through the mechanical zero-return action after the equipment is powered on, combined with the limit switches or photoelectric sensors of each axis. Physical spatial alignment is then performed through the following specific sub-steps:

[0085] In step S101, the adaptive test module first controls the measuring end of the No. 2 robotic arm 64 to pass through the magnetically conductive slide plate and extend into the interior of the outer shell 71. The electromagnetic shielding control module controls the sliding cover 75 to close, and then the adaptive test module controls the high magnetic permeability shielding sheet to open.

[0086] Subsequently, the main control unit of the adaptive test module generates a reference addressing command. Inside the housing 71 of the magnetic flux detection mechanism 7, there is a support plate with a workpiece support position and a calibration area. The standard reference magnet 74 is fixedly installed within the calibration area.

[0087] The upper surface of the standard reference magnet 74 is flush with its mounting base, and a high-permeability magnetic partition is provided between the standard reference magnet 74 and the workpiece bearing position. Above the standard reference magnet 74, there is also an openable and closable high-permeability magnetic shielding plate, which is connected to a micro-drive unit fixed on the bearing plate. The micro-drive unit is communicatively connected to the adaptive test module.

[0088] When performing self-calibration, the adaptive test module controls the micro-drive to open the high-permeability magnetic shield; when performing workpiece testing, the micro-drive drives the high-permeability magnetic shield to close, so as to reduce the influence of the superimposed magnetic field generated by the standard reference magnet 74 on the workpiece test area.

[0089] In specific implementation scenarios, the standard reference magnet is made of samarium cobalt to maintain the thermal stability of the magnetic field during long-term operation. The main control unit retrieves the three-dimensional center coordinate parameters of the standard reference magnet pre-set in the underlying system memory and uses them as the target addressing parameters.

[0090] In step S102, the adaptive test module calculates the horizontal axial displacement deviation based on the real-time coordinates of the current No. 2 robot 64 (which are obtained by the main control unit by reading the cumulative pulse count of the position encoder of each axis drive motor) and the target addressing parameters, and converts the displacement deviation into a corresponding motor drive pulse signal, which is then sent to the internal control loop of the conveying mechanism 6.

[0091] In step S103, the motor 62 of the conveying mechanism 6 receives a motor drive pulse signal. The motor 62 drives the connected linear transmission assembly to move, causing the second robotic arm 64 to move linearly along the horizontal guide rail on the gantry 61. A position encoder is coaxially mounted on the motor 62.

[0092] The position encoder collects the angular displacement data of the motor rotor in real time and feeds it back to the adaptive test module. The control system continuously corrects the running error through the position closed-loop algorithm, so that the No. 2 robot 64 moves directly above the standard reference magnet, and the center measurement axis of the No. 2 robot 64 coincides with the center point of the standard reference magnet in the vertical direction.

[0093] In step S104, the adaptive test module generates a vertical downward pressure command and sends it to the vertical lifting module inside the No. 2 robotic arm 64. The vertical lifting module outputs a direct driving force, causing the Hall plate 641 fixed at its end to move at a constant speed in the vertically downward direction.

[0094] When the Hall plate 641 descends to the calibrated height position set at the upper surface of the standard reference magnet, the vertical lifting module stops operating and the mechanism enters the mechanical locking state, thereby completing the closed-loop alignment control of the Hall plate 641 for the reference magnetic source.

[0095] The calibration height position is predetermined by the system based on the linear response range of the Hall plate 641 and the test data of the magnetic field distribution gradient on the surface of the reference magnet. It is usually set to be in a close-fitting state or with a fixed preset gap to meet the physical distance requirements for probe sensitivity calibration.

[0096] After completing the aforementioned closed-loop alignment control for the reference magnetic source, the adaptive test module provided by this invention further executes the raw data sampling process of the reference magnetic field.

[0097] With physical alignment completed and the vertical lifting module mechanically locked, the adaptive test module acquires the raw test data of the reference magnetic field without algorithm compensation through the following sub-steps:

[0098] In step S105, the main control unit of the adaptive test module sends a data acquisition trigger command to the Hall plate 641 at the end of the No. 2 robotic arm 64. The constant current source drive circuit integrated inside the Hall plate 641 is turned on, providing a constant excitation current to the Hall semiconductor element.

[0099] Under the influence of the reference magnetic field generated by the standard reference magnet, charge carriers inside the Hall plate 641 are deflected based on the Hall effect, generating an analog voltage signal at its output terminal that corresponds to the reference magnetic induction intensity. This analog voltage signal characterizes the current environmental conditions and the actual induction status of the probe material itself due to temperature drift or aging.

[0100] In step S106, to reduce the attenuation and interference of the measurement signal during transmission, the analog voltage signal output by the Hall plate 641 is transmitted to the signal conditioning circuit fixedly connected to it via a shielded wire. This signal conditioning circuit is embedded in the hardware integrated board at the end of the No. 2 robotic arm 64, and a differential amplifier circuit and a low-pass filter circuit are connected in series inside it.

[0101] The differential amplifier circuit amplifies the weak analog voltage signal, while the low-pass filter circuit removes interfering high-frequency environmental electromagnetic noise. The conditioned analog signal is then input to an electrically connected analog-to-digital converter for quantization, converting it into a discrete digital magnetic field signal.

[0102] In step S107, the analog-to-digital converter uploads the digital magnetic field signal to the main control unit of the adaptive test module in real time via the industrial communication bus. After receiving the continuous data stream, the main control unit performs multi-point mean filtering within a preset sampling time window to reduce the random white noise error of a single sampling.

[0103] In practice, the size of the preset sampling time window is determined by a combination of the sampling frequency of the hardware analog-to-digital converter and the timeliness requirements of a single test cycle. To avoid distortion of the reference data due to transient anomalies, the main control unit simultaneously calculates the variance of the continuous discrete sampled data within this time window before performing multi-point mean filtering.

[0104] When the calculated variance exceeds the system's set fluctuation threshold, the main control unit determines that the batch of samples is invalid and automatically triggers a resampling command; when the variance is within the fluctuation threshold range, the main control unit sets the extracted value as the original sampled value of the reference magnetic field on the surface of the standard reference magnet, and this variable is defined as... The main control unit will use this raw sample value The data is stored in a local register to complete the closed-loop capture of the original data, and used as the underlying input variable to provide a data foundation for the subsequent dynamic correction coefficient calculation model.

[0105] The adaptive testing module provided by this invention, after acquiring the original sampled value of the reference magnetic field, further executes the workflow of the dynamic correction coefficient solution model to complete the online closed-loop self-calibration of the test system accuracy.

[0106] The adaptive testing module completes the data calculation and update verification through the following sub-steps:

[0107] In step S108, the main control unit of the adaptive test module accesses the underlying read-only memory and retrieves the theoretical reference magnetic flux density value of the pre-written standard reference magnet, defined as... Its units are based on the international standard system, mT or T.

[0108] The theoretical reference magnetic induction intensity is the actual reference parameter obtained by measuring the center test point of the standard reference magnet 74 with a gaussmeter traceable to national metrological standards during the factory calibration stage, under the set standard room temperature environment, with the sliding cover 75 closed, the high permeability shielding plate open, and the gaussmeter probe at the set calibration height.

[0109] The main control unit will retrieve the theoretical reference magnetic flux density value. Compared with the original sampled values ​​obtained in the previous step It is simultaneously loaded into the motherboard processor's arithmetic register.

[0110] In step S109, the arithmetic logic unit inside the main control unit calls the correction coefficient calculation model. Before performing the division operation, the arithmetic logic unit first determines the original sampled value. Is the absolute value less than the zero-point noise threshold set by the system?

[0111] If the value is less than this threshold, it is determined that the probe is disconnected or has no signal output, and an abnormal interruption is triggered directly; if the value is greater than or equal to this threshold, the main control unit uses the theoretical reference magnetic induction intensity value. As a target reference value, the currently acquired raw sample value is used. Using the current feedback quantity, a division operation is performed to obtain the dynamic correction coefficient of the measurement system under the current state. .

[0112] The specific operational relationship is expressed as follows: In a physical sense, this dynamic correction coefficient primarily characterizes the overall sensitivity and gain deviation of the Hall plate 641 and the signal conditioning circuit under the current calibration state.

[0113] By introducing this ratio calculation model, the system can correct the overall proportional deviation of the measurement link online, providing an algorithmic basis for subsequent measurement compensation.

[0114] In step S110, to prevent distortion of the calculated coefficients due to sudden external strong magnetic interference or physical damage to the Hall plate 641, which could introduce global measurement errors, the adaptive test module performs a safety boundary verification logic before updating the system compensation parameters.

[0115] The main control unit will calculate the dynamic correction coefficient. With respect to the system's internally preset effective tolerance range A comparison is performed. Among them, the lower threshold... With upper limit threshold The value is determined by the system based on a comprehensive evaluation of the maximum full-scale temperature drift rate and the maximum annualized aging rate provided in the Hall 641 datasheet. It is usually set to fluctuate by 10% to 15% above or below the standard ratio of 1.0.

[0116] When the calculated dynamic correction coefficient When the measured value exceeds the preset effective tolerance range, the main control unit determines that there is structural damage or strong environmental interference in the current probe measurement link, triggers the underlying system protection mechanism, stops the test action of the corresponding measurement actuator, and sends a level pulse to drive the alarm light 4 on the outside of the frame 1 to prompt the operator to intervene and check.

[0117] When the dynamic correction coefficient When the data falls within the effective tolerance range, the main control unit determines that the closed-loop calibration data is valid and legal. Then, it writes the dynamic correction coefficient into the system dynamic compensation register and overwrites the old coefficient stored in the previous self-calibration cycle. At the same time, it records the current temperature detected by the temperature sensor as the calibration reference temperature for this self-calibration, thereby completing the entire self-calibration program and preparing data for the subsequent scanning and measurement of the workpiece to be tested.

[0118] The automated loading and unloading module provided by this invention includes a motion control process for push rod driving and stroke limit when executing the workpiece loading program.

[0119] The staggered feeding mechanism is located at the feed end of the frame and includes a hydraulic cylinder and a push rod. The hydraulic cylinder 52 is fixedly mounted on the base plate 51, and its power output end is fixedly connected to one end of the push rod 53. The push rod 53 is arranged along a linear guide structure on the base plate 51, and its other end faces the discharge port at the bottom of the feeding chute 54. The automated loading and unloading module performs power transmission and stroke closed-loop control through the following sub-steps:

[0120] In step S201, the main control unit of the automated loading and unloading module generates a feed command based on the overall operating cycle of the test system. The main control unit converts this command into a digital control signal and sends it to the three-position four-way solenoid directional valve that drives the hydraulic cylinder.

[0121] The coil of the electromagnetic directional valve is energized and generates magnetic force, driving the internal valve core to shift and change the flow path of the hydraulic oil. High-pressure hydraulic oil output from the hydraulic pump station enters the rodless chamber of the hydraulic cylinder via pipeline. A relief valve is connected in parallel on the main hydraulic supply circuit to limit the system's maximum operating pressure.

[0122] In practice, the opening pressure of the relief valve is usually set in the range of 2MPa to 5MPa to provide a smooth driving force and reduce the risk of workpiece deformation under pressure.

[0123] In step S202, the high-pressure hydraulic oil entering the rodless chamber of the hydraulic cylinder pushes the internal piston to slide forward along the inner wall of the cylinder. The linear motion of the piston converts the hydrodynamic power into mechanical thrust, driving the push rod fixedly connected to it to extend forward along the linear guide groove.

[0124] The front end of the push rod contacts the single workpiece at the front of the workpiece array, applying a horizontal force so that the workpiece overcomes the static friction between itself and the base plate and begins to move forward along the feeding slide.

[0125] In step S203, to limit the push rod stroke from exceeding the structural allowable range and to achieve workpiece separation and positioning, a stroke limit component is provided on the side of the push rod's movement trajectory in the staggered separation feeding mechanism. This stroke limit component includes a front limit switch installed at the front end of the guide groove and a zero-return limit switch installed at the rear end. The front limit switch is installed in an adjustment groove extending along the movement direction of the push rod 53, and its trigger position can be adjusted according to the size of the workpiece 55. A sensor lever is securely mounted on the side of the push rod.

[0126] When the push rod extends the workpiece to the set separation position, the sensing lever contacts and triggers the front limit switch. The internal contacts of the front limit switch close, generating a positioning feedback signal and transmitting it to the main control unit of the automated loading and unloading module. The travel distance to this separation position is mechanically fine-tuned according to the physical dimensions of different workpiece models, typically set between 10mm and 50mm.

[0127] In step S204, after receiving the position feedback signal from the front limit switch, the main control unit sends a neutral position lockout command to the three-position four-way solenoid directional valve. The valve core of the solenoid directional valve resets to the neutral position, cutting off the oil inlet to the rodless chamber of the hydraulic cylinder, and the push rod stops extending.

[0128] To prevent hydraulic overload caused by hardware failure of the front limit switch, the main control unit simultaneously starts the internal timing protection program when issuing the feed command;

[0129] If the push rod extends beyond the preset timeout threshold without receiving a positioning feedback signal, the main control unit forcibly cuts off the directional valve output and triggers a system shutdown alarm. At this point, the single workpiece at the front end has been pushed to the preset robotic arm gripping position.

[0130] After the robotic arm completes the workpiece gripping, the main control unit outputs a reverse drive level to control the oil intake of the rod chamber of the hydraulic cylinder, driving the push rod back to the initial position and triggering the zero-return limit switch, thus completing the mechanical cycle of single push rod drive and stroke closed-loop control.

[0131] The automated loading and unloading module provided by this invention, after pushing the single workpiece at the foremost end of the workpiece array away from the workpiece array by the push rod, further executes the action control process of separating the workpieces in the array one by one, so as to reduce the risk of multiple materials being mis-loaded due to workpiece surface adhesion or magnetic attraction.

[0132] To reduce the risk of multiple items being ejected simultaneously due to magnetic attraction between adjacent permanent magnets, a material blocking assembly is provided at the bottom of the feeding slide 54. The material blocking assembly includes lateral blocking cylinders respectively disposed on both sides of the feeding slide 54 and a non-magnetic positioning block connected to the output end of the lateral blocking cylinders. A micro-motion trigger switch is provided at the gripping position.

[0133] Before the hydraulic cylinder 52 drives the push rod 53 to move, the lateral blocking cylinder first drives the positioning pressure block to clamp the workpiece group located above the bottom workpiece, so that the bottom workpiece is kept in an ejectable state; after the push rod 53 pushes the bottom workpiece to the predetermined gripping position, the micro-motion trigger switch outputs the position signal.

[0134] After receiving the arrival signal, the main control unit maintains the clamping state of the lateral blocking cylinder. Once the first robotic arm 63 picks up the bottom workpiece and the push rod 53 retracts, the lateral blocking cylinder is released, allowing the remaining workpiece to fall by one workpiece thickness under gravity. If the arrival signal is not received within the set time, the push rod 53 and the lateral blocking cylinder are reset, and the alarm light 4 is triggered.

[0135] See attached document Figure 4 The conveying mechanism 6 includes a gantry frame 61, a motor 62, a first robotic arm 63, a second robotic arm 64, a third robotic arm 65, and a cylinder 66. The gantry frame 61 is fixedly mounted on the worktable 9. The motor 62 is connected to the X-axis linear transmission assembly mounted on the gantry frame 61, and is used to drive the first robotic arm 63, the second robotic arm 64, and the third robotic arm 65 to move horizontally along the workpiece conveying direction.

[0136] The lifting end of robotic arm 63 is detachably connected to suction cup 632 via buckle 631, and the lifting end of robotic arm 65 is detachably connected to suction cup 652 via buckle 651; robotic arm 63 and robotic arm 65 are respectively driven to lift by corresponding cylinders 66.

[0137] The second robotic arm 64 includes a Y-axis linear transmission assembly connected to the moving end of the X-axis linear transmission assembly, a Z-axis drive mechanism and an attitude adjustment assembly disposed at the moving end of the Y-axis linear transmission assembly, and a Hall plate 641 and a laser displacement sensor 642 are mounted at the end of the attitude adjustment assembly, enabling the Hall plate 641 to move along the X-axis, Y-axis and Z-axis.

[0138] The automated loading and unloading module provided by this invention, after the conveying mechanism 6 moves to the predetermined material picking position, further executes the control process of material picking and loading into the test chamber to smoothly transport the peeled and positioned individual workpieces to the testing station in the test chamber.

[0139] S301, when the conveying mechanism 6 moves to the predetermined material picking position and the automated loading and unloading module receives the workpiece separation signal, it controls the cylinder 66 corresponding to the No. 1 robot 63 to extend, so that the suction cup 632 descends and adheres to the upper surface of the workpiece 55.

[0140] S302, the automated loading and unloading module connects to the negative pressure generator connected to suction cup 632, and determines whether suction cup 632 has adsorbed the workpiece 55 based on the feedback from the negative pressure sensor; if the negative pressure does not reach the set threshold, the transfer stops and the alarm light 4 is triggered.

[0141] S303, after confirming that the adsorption is effective, cylinder 66 drives robot arm 63 to rise, and motor 62 drives robot arm 63 to move horizontally to above the workpiece bearing position of magnetic flux detection mechanism 7.

[0142] S304, robot arm 63 descends to the unloading height, cuts off the negative pressure and releases workpiece 55, causing workpiece 55 to fall into the workpiece bearing position on the bearing plate. Then robot arm 63 rises and exits the opening and closing range of magnetic flux detection mechanism 7.

[0143] In actual industrial settings, the operation of surrounding equipment motors, power grid fluctuations, and external magnetic fields such as the Earth's magnetic field continuously cause crosstalk to the weak surface magnetic field test. In order to attenuate the DC magnetic field strength in the test area to within the set safety limit, such as below 50 A / m, this invention sets up a dynamic isolation hardware structure at the test station and executes the corresponding construction process through an electromagnetic shielding control module.

[0144] See attached document Figure 5 The magnetic flux detection mechanism 7 includes a housing 71, a drive cylinder 72, a sealing strip 73, a standard reference magnet 74, and a sliding cover 75. The housing 71 is fixedly installed on the worktable 9, and a support plate is provided inside the housing 71. The support plate is provided with a workpiece support position and a calibration area.

[0145] The top of the housing 71 has an opening for the workpiece 55 and the measuring end of the No. 2 robotic arm 64 to enter and exit. The sliding cover 75 is slidably connected to the guide structure on the top of the housing 71. The drive cylinder 72 is fixed to the side of the housing 71, and its output end is connected to the sliding cover 75 to drive the sliding cover 75 to open or close the opening in the horizontal direction. The sealing strip 73 is provided around the opening and at the overlap position between the sliding cover 75 and the housing 71.

[0146] To achieve magnetic field isolation, the overall frame of the outer shell 71 is made of permalloy plates with high initial permeability, which are welded together to bypass and guide external low-frequency magnetic lines of force by utilizing the magnetic shielding effect of the high permeability material.

[0147] The thickness and number of layers of the permalloy plate are selected based on the internal dimensions of the outer shell 71 and the background magnetic field strength of the environment in which the equipment is located. After assembly, the background magnetic field strength of the cavity is checked to ensure that the background magnetic field strength at the workpiece bearing position does not exceed the set limit.

[0148] The inner wall of the outer shell 71 is fixed with an electromagnetic wave absorbing material layer by an adhesive, which is used to absorb and dissipate the high-frequency electromagnetic reflected waves generated by the operation of the electromechanical components inside the test chamber.

[0149] The sealing strip 73 is made of conductive rubber and flexible high magnetic permeability material. It is used to fill the assembly gap between the sliding cover 75 and the housing 71, and reduce the electromagnetic field and low-frequency external magnetic field from entering the housing 71 through the gap.

[0150] A two-dimensional overlapping magnetically conductive slide plate is stacked on top of the sliding cover 75 for the measuring end of the No. 2 robotic arm 64 to pass through. The magnetically conductive slide plate is supported by a slide frame connected to the X-axis and Y-axis moving ends of the No. 2 robotic arm 64 and can slide relative to the sliding cover 75 along the X-axis and Y-axis. The lifting end of the No. 2 robotic arm 64 extends into the housing 71 through the central hole on the magnetically conductive slide plate.

[0151] The magnetically conductive slide plate can move along the X and Y axes with the No. 2 robotic arm 64, and always remains overlapping with the sliding cover 75 during the movement. When the sliding cover 75 is opened or closed, the magnetically conductive slide plate remains supported by the carriage and slides relative to the sliding cover 75 to avoid interference between the sliding cover 75 and the No. 2 robotic arm 64.

[0152] The specific method for the electromagnetic shielding control module to execute dynamic electromagnetic shielding cavity construction control includes the following sub-steps:

[0153] S401, the automated loading and unloading module confirms that robot arm 63 has exited the movement range of slide cover 75, and controls the measuring end of robot arm 64 to extend into the interior of housing 71 through the magnetic sliding plate. After confirming that there is no movement interference between each robot arm and slide cover 75, the electromagnetic shielding control module releases the movement interlock of slide cover 75.

[0154] S402, the electromagnetic shielding control module controls the drive cylinder 72 to extend or retract, driving the sliding cover 75 to move horizontally to the closed position along the guide structure at the top of the housing 71.

[0155] S403, during the movement of the sliding cover 75, the electromagnetic shielding control module judges the movement state of the sliding cover 75 based on the position feedback and operating current of the drive cylinder 72; if the sliding cover 75 has not reached the theoretical closed position and the operating current exceeds the set threshold, it is determined that there is mechanical jamming, and the drive cylinder 72 is controlled to stop and reverse reset.

[0156] S404, when the sliding cover 75 reaches the closed position and triggers the closed position signal, the sliding cover 75 and the sealing strip 73 fit together, the two-dimensional overlapping magnetic sliding plate and the sliding cover 75 remain overlapping, and the electromagnetic shielding control module controls the drive cylinder 72 to maintain the current position, thereby forming a relatively closed low background magnetic field test space.

[0157] After the dynamic electromagnetic shielding cavity is constructed and it is confirmed that there is no external physical interference in the operating environment, in order to reduce the spatial attitude tilt error of the workpiece 55 surface caused by machining tolerances, uneven material thickness, or internal fitting clearance of the bearing plate, the adaptive test module executes the program to obtain the three-dimensional non-collinear coordinates of the surface.

[0158] In terms of hardware configuration, the end effector of the No. 2 robotic arm 64 inside the handling mechanism 6 is not only fixedly installed with a Hall plate 641 for magnetic field detection, but also has a laser displacement sensor 642 fixedly installed side by side on its side.

[0159] The measuring optical axis of the laser displacement sensor 642 is parallel to the measuring center axis of the Hall plate 641 by a known offset distance, and there is a pre-calibrated hardware mounting height difference between the two in the vertical direction. The elevation difference The parameters are pre-determined and embedded in the system's underlying parameters during the factory calibration phase using standard gauge blocks. The adaptive test module is connected to the motor encoder and laser displacement sensor 642 inside the conveying mechanism 6 via an industrial bus electrical signal.

[0160] The specific method for the adaptive testing module to obtain the three-dimensional non-collinear coordinates of the workpiece surface 55 includes the following sub-steps:

[0161] S405, test reference point plane trajectory planning. The adaptive test module first controls the A-axis drive mechanism and the B-axis drive mechanism to return to the mechanical zero position, so that the measuring optical axis of the laser displacement sensor 642 remains parallel to the Z-axis of the equipment;

[0162] Subsequently, the adaptive testing module retrieves the two-dimensional contour dimension parameters of the workpiece 55 pre-stored in the storage unit and plans three test reference points on the theoretical horizontal projection surface of the workpiece 55.

[0163] To ensure that these three points are non-collinear in mathematical and physical space to satisfy the subsequent conditions for establishing the spatial plane, the system generates coordinate planning according to the vertex distribution rules of isosceles or equilateral triangles. The adaptive testing module defines the planar coordinates of these three test reference points as follows: , and And use it as the target position command for driving the conveying mechanism 6 to translate.

[0164] S406, first spatial reference point distance measurement and coordinate calculation. The adaptive test module sends control commands to the servo drive circuit inside the handling mechanism 6, driving the second robotic arm 64 to move horizontally along the X and Y axes to the first test reference point. The position above.

[0165] After positioning is completed, the emitting end of the laser displacement sensor 642 emits a probe laser beam towards the surface of the workpiece 55, and the receiving end receives the light signal reflected from the surface. The internal processing circuit of the sensor calculates the measured vertical distance from the probe to the surface of the workpiece 55 based on the principle of triangulation, defined as... .

[0166] In the principle of distance measurement, a laser emitter emits a beam of light that shines on the surface of the workpiece. The diffusely reflected light is focused by a lens and imaged onto an internal photoelectric receiving element. The displacement change of the light spot position is calculated and converted into a measured distance value.

[0167] To reduce the single-time distance measurement jump error caused by local surface roughness, machining scratches, or sudden changes in reflectivity of workpiece 55, the adaptive testing module extracts multiple consecutive distance measurement data within a preset sampling time window, which is set to 10ms to 50ms. After removing extreme values ​​that exceed three times the standard deviation, the arithmetic mean is calculated and used as the final effective distance value.

[0168] Subsequently, the adaptive test module reads the absolute height of the Z-axis encoder of the current No. 2 robotic arm 64. And retrieve the pre-stored hardware installation height difference The coordinate calculation model is called to calculate the height coordinates of that point on the surface of workpiece 55. The specific solution formula is as follows: .

[0169] The adaptive testing module combines the acquired x and y coordinates with the calculated height coordinates to generate the first spatial point cloud coordinates. And store it in the cache.

[0170] S407, Remaining reference point tracking and point cloud coordinate generation. The adaptive test module cyclically calls the underlying logic of the above-mentioned servo positioning, laser ranging, extreme value elimination, mean filtering, and height calculation, sequentially driving the No. 2 robot arm 64 to move to the second and third test reference points. and .

[0171] The laser displacement sensor 642 collects and the system calculates the height coordinates of the corresponding positions. and The adaptive testing module synchronously combines the calculation results and records them as the coordinates of the second spatial point cloud. and the third spatial point cloud coordinates .

[0172] S408, Spatial Non-collinear Data Verification and Abnormal Resampling. To prevent data acquisition failure at a certain point due to mechanical transmission step loss or workpiece 55 surface edge collapse, which could lead to three points becoming collinear and causing a control dead zone in subsequent normal vector calculation due to division by zero overflow, the adaptive test module executes mathematical fault-tolerant verification logic after completing three samplings.

[0173] The main control unit of the adaptive test module extracts the first spatial vector pointing from P1 to P2 and the vector from P1 to P2. point to The second spatial vector is used to perform a three-dimensional vector cross product operation to solve for the magnitude of the cross product. This magnitude geometrically represents the size of the area of ​​the triangle formed by the two spatial vectors mentioned above; the smaller the area, the closer it is to a collinear state of the three points.

[0174] The main control unit compares the magnitude of the cross product with the system's preset collinearity tolerance threshold. This collinearity tolerance threshold is determined based on a comprehensive evaluation of the minimum resolution of the laser displacement sensor 642 and the minimum baseline distance between the two points.

[0175] When the magnitude of the cross product is less than the collinearity tolerance threshold, the system determines that the three extracted points are too close together or have an abnormal distribution in physical space, and therefore determines that the current sampling group is invalid. At this time, the adaptive testing module automatically regenerates new reference points around the originally planned planar coordinates with a preset fixed offset, and controls robot arm 64 to perform a secondary supplementary test.

[0176] To avoid dead loop control abnormalities, the adaptive test module has a maximum retry counter. If the number of consecutive resampling exceeds the set limit, the system determines that the surface morphology of the current workpiece is abnormal or that no workpiece is placed, stops the test and triggers an alarm.

[0177] When the magnitude of the cross product is greater than or equal to the collinearity tolerance threshold, the system determines that the verification passes and... , , It is established as an effective non-collinear coordinate input source, providing data support for subsequent solution of spatial normal vectors and execution of multi-degree-of-freedom constant distance compensation control.

[0178] The adaptive testing module obtains the coordinates of three effective non-collinear spatial point clouds on the surface of workpiece 55 through fault tolerance verification. , as well as Subsequently, to characterize the current actual tilting posture of workpiece 55, the main control unit of the adaptive test module calls the built-in spatial normal vector calculation model. This calculation model, based on an analytical geometry algorithm in memory, converts discrete point cloud data into attitude deflection angles that can guide the movement of the mechanical transmission shaft.

[0179] The specific method for the adaptive testing module to execute the spatial normal vector calculation model includes the following sub-steps:

[0180] S409, Coplanar Space Vector Construction. The main control unit of the adaptive test module extracts the coordinates of three valid non-collinear point cloud spaces within the cache area. , and The main control unit is based on points. Taking the point as the starting point of spatial geometry, respectively and points As the spatial geometric endpoint, two coplanar spatial vectors are constructed by performing difference operations on the corresponding coordinate axes.

[0181] Specifically, the first spatial vector is defined as Its coordinates are represented as The second spatial vector is defined as follows: Its coordinates are represented as Geometrically, the two spatial vectors mentioned above intersect at a point. Furthermore, they are not collinear, thus determining the actual inclined plane where the surface of workpiece 55 is located in physical space.

[0182] S410, solving for the cross product of the original normal vector. The main control unit calls the underlying mathematical operation library to perform calculations on the first spatial vector. With the second space vector Perform a three-dimensional vector cross product operation to obtain the original normal vector perpendicular to the actual surface of the workpiece 55. .

[0183] Original normal vector It includes projection components along three orthogonal coordinate axes, denoted as follows: , and The internal calculation logic is as follows:

[0184] ;

[0185] ;

[0186] .

[0187] The three components mentioned above together define the normal deflection tendency of the plane in three-dimensional space.

[0188] S411, Normal Vector Direction Verification and Normalization. Since the vector direction obtained from the cross product operation is affected by the order of the input vectors, to match the actual working perspective of the No. 2 robotic arm 64 inside the handling mechanism 6 from top to bottom, the main control unit needs to normalize the original normal vector... Z-axis component Perform symbol verification. When determined... When the value is less than zero, it indicates that the calculated normal vector points into the interior of workpiece 55, and the main control unit then... , and Simultaneously perform the inversion operation to adjust it into a valid normal vector pointing to the test space above.

[0189] After completing the direction verification, the main control unit calculates the original normal vector. Spatial modulus In terms of operational form, the modulus... Depend on , and The square root of the sum of the squares of the three factors is taken. Thanks to the hardware error-proofing and software threshold interception for the collinearity of the three points in the pre-measurement step, the modulus here is... The value is greater than the minimum resolution threshold for floating-point numbers set by the system, thus avoiding the risk of program errors caused by division by zero in subsequent processing.

[0190] Subsequently, the main control unit will... , and Dividing each by the modulus L yields a unit space normal vector with a constant modulus of 1. This eliminates the numerical interference of the original coordinate physical distance on the attitude angle calculation.

[0191] S412, Spatial attitude deflection angle calculation. The adaptive testing module uses the normalized unit space normal vector. The task is to determine the tilt orientation of workpiece 55's surface relative to the theoretical horizontal reference plane of the equipment. The theoretical horizontal reference plane of the equipment is parallel to the worktable surface 9, and its unit normal vector is fixed as a reference vector (0,0,1) in the positive Z-axis direction. The main control unit calculates the unit spatial normal vector using the inverse trigonometric function calculation module. The Euler angles corresponding to each axial component.

[0192] To address the truncation error caused by floating-point operations in the underlying processor, the main control unit first performs a trigonometric function calculation on the input components. and Execute clamp protection logic to restrict its value to the mathematical domain of [-1, 1].

[0193] In the actual calculation, the main control unit defines the tilt angle of the workpiece 55 surface rotating around the X-axis as the roll angle, and its value corresponds to the component. The arcsine value; the tilt angle of the workpiece surface 55 rotating about the Y-axis is defined as the pitch angle, and its value corresponds to the component. The arcsine operation takes the inverse value.

[0194] The calculated roll and pitch angle data are packaged and written into the underlying registers of the motion controller in real time. These two angle parameters characterize the current physical assembly tilt of workpiece 55 and serve as the reference input parameters for driving robot arm 64 to adjust the spatial pose of Hall plate 641 and perform multi-degree-of-freedom constant distance compensation.

[0195] After obtaining the roll and pitch angles of the workpiece surface 55 through the spatial normal vector calculation model, the adaptive test module performs multi-degree-of-freedom attitude adjustment and constant distance compensation control, driving the magnetic flux detection mechanism to adjust the spatial attitude, so that the test probe and the workpiece surface are adjusted to a set distance parallel to each other, thereby completing the standardized scanning of the surface magnetic distribution.

[0196] In terms of hardware mechanical transmission, the end effector of the second manipulator 64 of the handling mechanism 6 is equipped with an attitude adjustment component with two-axis rotational degrees of freedom. The attitude adjustment component includes an orthogonally arranged A-axis drive mechanism and a B-axis drive mechanism. The flange of the A-axis drive mechanism is fixedly connected to the Z-axis lifting end of the second manipulator 64 to realize rotational movement around the X-axis;

[0197] The B-axis drive mechanism's mounting base is installed at the output end of the A-axis drive mechanism to achieve rotational motion around the Y-axis. A Hall effect sensor 641 is fixedly connected to the end flange of the B-axis drive mechanism. Each servo drive mechanism has a built-in absolute encoder and is electrically connected to the main control unit of the adaptive test module via an industrial bus.

[0198] The specific method for the adaptive test module to perform multi-degree-of-freedom attitude adjustment and constant distance compensation control includes the following sub-steps:

[0199] S413, end-effector spatial attitude follow-up deflection. The main control unit of the adaptive test module reads the roll and pitch angle data calculated by the upper-level logic. Before executing the drive command, the main control unit compares these two angle parameters with the system's preset mechanical limit threshold. This mechanical limit threshold is set based on the physical interference boundary of the attitude adjustment component and the torsional limit of the internal wiring harness.

[0200] When the calculated angle exceeds the threshold, the system determines that the workpiece 55's placement posture is out of tolerance, stops the test, and executes a reset alarm action; when the calculated angle is within the threshold range, the main control unit combines the mechanical reduction ratio of the A-axis drive mechanism and the B-axis drive mechanism to convert the roll angle and pitch angle into the corresponding underlying motion control pulse sequence.

[0201] The A-axis drive mechanism and the B-axis drive mechanism rotate synchronously according to pulse commands, causing the Hall plate 641 to deflect in attitude, so that its detection plane and the actual inclined surface of the workpiece 55 are in a spatially parallel state.

[0202] S414, Spatial rotation center offset compensation calculation. Due to the pre-calibrated structural offset distance between the physical rotation center of the attitude adjustment component and the actual sensing center of the Hall plate 641 in the Z-axis direction, when the A-axis and B-axis rotate, the sensing center of the Hall plate 641 will generate a corresponding vertical height deviation in the Z-axis direction.

[0203] The main control unit invokes the built-in kinematic matrix compensation algorithm to calculate the accompanying height displacement based on the structural offset distance parameters stored in the system and the current roll and pitch angles. These structural offset distance parameters are pre-measured and calibrated by a coordinate measuring machine during the equipment assembly phase and are stored in the controller's non-volatile memory.

[0204] Geometrically, this operation is equivalent to the difference in projection of the arc chord length relative to the center of rotation onto the Z-axis coordinate system, which the main control unit records as the attitude height compensation value.

[0205] S415, constant lift-off distance positioning and anti-collision pressing. Based on the workpiece 55 surface height coordinates obtained in the previous steps, the calculated attitude height compensation value, and the target lift-off distance set by the system, the main control unit calculates the final target coordinates of the No. 2 robot 64 in the Z-axis direction.

[0206] The target lifting distance is set according to the process requirements of the magnetic field testing specification, with a set range of 0.5mm to 1.0mm. The main control unit sends an absolute positioning command to the Z-axis drive mechanism of the second robotic arm 64, driving the Hall plate 641 to move downwards to the final target coordinates.

[0207] During the descent, the Z-axis servo driver samples the output torque current of the motor in real time. If the current value suddenly increases and exceeds the set collision protection threshold before reaching the target coordinate, the system determines that physical interference has occurred, cuts off the feed pulse, and reverses and retracts, forming a hardware anti-collision closed loop for the test probe.

[0208] S416, Mesh scanning and dynamic data acquisition. After positioning, the Hall plate 641 establishes an equidistant parallel relationship with the surface of the workpiece 55 in three-dimensional space at a specified distance. Based on the two-dimensional contour dimension parameters of the workpiece 55 and the aforementioned three reference point data, the main control unit generates a scanning trajectory within the XY projection plane of the workpiece 55 according to the set test step distance;

[0209] For planar workpieces, a planar mesh trajectory is generated. For curved workpieces, the main control unit dynamically updates the Z-axis position, local normal direction, and target lift-off distance of each scanning node based on the surface height data continuously acquired by the laser displacement sensor 642 during the scanning process. The test step distance is set to 1.0mm to 5.0mm based on the effective sensing area of ​​the Hall plate 641 and the test resolution requirements. The main control unit controls the second robotic arm 64 to perform interpolation motion along the X and Y axes, driving the Hall plate 641 to translate along the set trajectory.

[0210] During the translation interpolation process, the main control unit monitors the current coordinates of the X-axis and Y-axis in real time. If the coordinate data exceeds the safety margin boundary set by the theoretical two-dimensional contour expansion of the workpiece, the system triggers the over-boundary stop protection to prevent the probe from falling or hitting the fixture.

[0211] During movement within the safety boundary, the Hall plate 641 senses magnetic field lines that pass perpendicularly through its surface. The internal semiconductor Hall effect circuit uses the Lorentz force deflection principle to convert the magnetic field strength into a corresponding voltage signal.

[0212] The voltage signal, after being amplified and converted from analog to digital by the signal conditioning circuit, is read by the adaptive test module and mapped to the surface magnetic flux density value corresponding to each coordinate point. The main control unit combines and packages the acquired coordinate data and magnetic flux density values, stores them in the memory, and completes the process of acquiring the surface magnetic field distribution characteristics of the single workpiece.

[0213] After completing the spatial attitude compensation and initial positioning of the workpiece 55 surface, in order to reduce the interference of local warping or micro-waves on the magnetic field detection accuracy caused by the sintering process on the surface of the workpiece 55, the adaptive test module executes the adaptive constant distance scanning trajectory control program to control the gap between the magnetic flux detection mechanism and the workpiece surface to be kept within the set dynamic tolerance range.

[0214] In terms of mechanical transmission and hardware layout, the X-axis linear transmission assembly is set on the gantry 61, and the No. 2 robot 64 is installed on the moving end of the X-axis linear transmission assembly via the Y-axis linear transmission assembly. The Z-axis drive mechanism of the No. 2 robot 64 is installed on the moving end of the Y-axis linear transmission assembly.

[0215] The Z-axis drive mechanism of robot arm 64 is fixedly mounted on the Y-axis slide saddle. The end flange of robot arm 64 is fixedly connected to the attitude adjustment assembly, and the laser displacement sensor 642 and Hall plate 641 are fixedly mounted side by side at the execution end of the attitude adjustment assembly. The drive motors of each linear guide are equipped with high-resolution position encoders and establish a closed-loop electrical connection with the main control unit of the adaptive test module.

[0216] The specific method for the adaptive constant-distance scanning trajectory control executed by the adaptive test module includes the following sub-steps:

[0217] S501, Scan Area Mapping and Planar Trajectory Look-Ahead Generation. The main control unit retrieves the two-dimensional contour dimensions of workpiece 55 from memory and establishes a physical boundary model in the virtual control system. Considering the physical characteristics of magnetic field line distortion and divergence in the edge region of the permanent magnet, the main control unit shrinks the safety offset distance inward based on the physical boundary to generate an effective test boundary.

[0218] The safety offset distance is set between 1.0 mm and 3.0 mm based on the magnetic field divergence characteristics. Within the effective test boundary, the main control unit plans discrete test coordinate nodes according to the preset test step distance, generating a continuous reciprocating XY projection scanning path. The main control unit sends the planned path to the underlying motion controller, controlling the X-axis linear guide and Y-axis linear guide of the conveying mechanism 6 to perform planar linkage interpolation.

[0219] S502, real-time ranging and acquisition of surface micro-morphology. During the XY plane interpolation motion, if the Z-axis height is simply kept constant, local deformation errors on the surface of workpiece 55 will cause the actual lifting distance to deviate from the target set value. The main control unit controls the laser displacement sensor 642 to start continuous measurement mode.

[0220] The laser displacement sensor 642 emits a probe beam to the surface of the workpiece 55 directly below it at a sampling frequency of 1000Hz set by the system, and obtains the relative physical distance between the probe and the surface in real time.

[0221] Since the laser displacement sensor 642 and the Hall plate 641 have a known installation offset vector in their spatial structure, the main control unit stores the acquired distance data in a first-in-first-out buffer queue, and converts the installation offset vector to the device's global coordinate system based on the current rotation angles of the A-axis and B-axis. Then, it calculates the position delay compensation amount by combining the real-time feed speeds of the X-axis and Y-axis, and converts the laser ranging data into the actual lift-off distance at the corresponding position of the sensing center of the Hall plate 641, thereby achieving a physical mapping and matching between the ranging data and the spatial coordinates of the Hall plate 641.

[0222] The main control unit monitors the depth of the buffer queue in real time. If the data accumulation exceeds the set maximum queue length limit, it determines that the interpolation speed and sampling frequency are mismatched and triggers a feed deceleration command to prevent data overflow and loss.

[0223] S503 features Z-axis dynamic tracking and constant-distance closed-loop compensation. The main control unit extracts relative physical distance data matching the current coordinates from the buffer queue and performs a subtraction operation with the target lift distance set by the system to solve for the dynamic height deviation.

[0224] To prevent sudden distance changes caused by deep pits or severe scratches on the surface of workpiece 55, which could trigger a violent Z-axis servo response and cause the workpiece to collide, the main control unit performs a limiting clamping judgment before inputting data to the regulator. When the rate of change of the dynamic height deviation exceeds the set safety threshold, the system rejects the abnormal jump value and maintains the output command of the previous cycle;

[0225] When the dynamic height deviation is within the safety threshold, the main control unit inputs the dynamic height deviation to the built-in PID closed-loop controller. The controller then sequentially performs proportional gain amplification, integral error accumulation elimination, and derivative overshoot suppression operations to calculate the dynamic compensation pulse command for the Z-axis.

[0226] The main control unit, based on the height data of the current ranging point and at least two adjacent ranging points located in two different plane directions, ensures that the three ranging points are not collinear in the XY projection plane, calculates the local normal vector at the current scanning node, and synchronously controls the A-axis and B-axis to adjust the direction of the Hall plate 641 according to the local normal vector, while controlling the Z-axis to correct the lifting distance between the Hall plate 641 and the surface of the workpiece 55.

[0227] Through the coordinated control of the X, Y, Z, A and B axes, the measuring axis of the Hall plate 641 is set along the local normal direction of the workpiece 55 surface, and the Hall plate 641 is kept at a set distance from the workpiece 55 surface.

[0228] S504, high-speed position triggering and dynamic data latching. Addressing the issue of measurement noise introduced by mechanical vibration during servo motor operation, the main control unit performs S-shaped velocity curve planning and acceleration / deceleration filtering on the output pulses of the motion controller to reduce inertial impact at mechanical commutation nodes.

[0229] The motion controller's underlying hardware trigger module uses position feedback from the X, Y, and Z axes, as well as angle feedback from the A and B axes, for real-time comparison. When the actual position and angle of each axis enter the set tolerance range of the corresponding test node, and the actual lift-off distance is within the set tolerance range, the underlying hardware trigger module sends a hardware synchronization trigger signal to the data acquisition card.

[0230] Upon receiving the trigger signal, the data acquisition card latches and reads the simulated magnetic field strength signal output by the Hall plate 641. The main control unit packages and binds the spatial three-dimensional coordinate data with the converted magnetic flux density value and stores it in non-volatile memory, completing the acquisition of magnetic field characteristics under high-precision dynamic morphology tracking.

[0231] After obtaining the original magnetic flux density value in the dynamic data latching stage, in order to reduce the measurement error introduced by ambient temperature drift and dynamic following hysteresis of the servo system, the main control unit of the adaptive test module calls the built-in real-time compensation algorithm for measurement data to reconstruct the original data numerically.

[0232] In terms of hardware configuration and connectivity, a temperature sensor (e.g., a platinum resistance temperature sensor) is mounted on the end of the attitude adjustment component of the conveying mechanism 6, closely attached to the Hall plate 641. The sensing end of the temperature sensor is fixedly connected to the substrate surface of the Hall plate 641 using thermally conductive adhesive. The output pin of the temperature sensor is connected to the temperature conditioning circuit inside the main control unit via a shielded cable to obtain the actual operating temperature of the sensing element.

[0233] The specific method for the adaptive testing module to execute the real-time compensation algorithm for measurement data includes the following sub-steps:

[0234] S505, Temperature Drift Compensation. The main control unit first multiplies the original magnetic flux density value latched by the data acquisition card with the dynamic correction coefficient obtained in step S100 to obtain the online calibration magnetic flux density value; then it reads the current operating temperature output by the temperature conditioning circuit, and performs a difference operation between the current operating temperature and the calibration reference temperature recorded in the most recent self-calibration time to obtain the temperature deviation.

[0235] To prevent the compensation value from jumping due to a disconnection of the temperature sensor or strong external interference, the main control unit judges the effective range of the current operating temperature. When the current operating temperature exceeds -10℃ to 60℃, temperature compensation is stopped and an abnormal prompt is output.

[0236] When within this range, the online calibration flux density value is compensated based on the temperature deviation. Physically, the carrier mobility of the semiconductor Hall element changes with temperature fluctuations, causing temperature drift in the output voltage.

[0237] Based on this linear response principle, the main control unit uses the formula... Calculate temperature-compensated magnetic flux density In this formula, The original magnetic flux density value latched by the acquisition card; It is the temperature comprehensive compensation coefficient pre-programmed into the non-volatile memory inside the system, and its physical unit is ppm / ℃.

[0238] S506, Dynamic Lift-off Residual Error Compensation. Due to the physical phase lag in the PID closed-loop regulation of the Z-axis drive mechanism, a residual microscopic height deviation will occur between the Hall plate 641 and the workpiece 55 surface during high-speed interpolation. The main control unit extracts the difference between the actual physical distance fed back by the laser displacement sensor 642 and the target lift-off distance within the current sampling period, and records it as the residual height deviation. .

[0239] The main control unit performs a threshold check when the residual height deviation... When the absolute value is greater than the set tolerance threshold of 0.1mm, the current coordinate point is determined to be in a servo out-of-synchronization state, and the point is marked as invalid and discarded data;

[0240] When the residual height deviation is within the tolerance threshold, the main control unit calls the pre-established lift-off distance-magnetic induction intensity compensation model to perform height compensation on the current magnetic flux density value.

[0241] The compensation model is established by controlling the Hall plate 641 to measure standard samples at multiple known lift-off distances during the equipment calibration stage, and based on the magnetic induction intensity data corresponding to each lift-off distance. It can be established using a lookup table, a piecewise fitting function, or an attenuation model established for workpieces of corresponding specifications.

[0242] S507, Sensor Nonlinear Mapping Reconstruction. To correct the input-output nonlinear characteristics of the Hall effect sensor 641 semiconductor material itself, the main control unit calls a preset compensation model to perform dimensionality reduction calculations. The specific calculation steps are as follows:

[0243] First, calculate the height-compensated magnetic flux density. The power is raised to the first, second, and third powers, and then each power is multiplied by its corresponding nonlinear weighting coefficient.

[0244] Finally, the product results are summed with the zeroth-order bias constant to reconstruct the final standard magnetic flux density. .

[0245] The aforementioned weighting coefficients and bias constants are calculated and generated by the multi-point magnetic field calibration platform during factory calibration and then imported into the system. The main control unit then calculates the standard magnetic flux density. The original data in the cache is replaced and stored in the three-dimensional coordinate mapping matrix for subsequent magnetic field uniformity evaluation.

[0246] See attached document Figure 6 The feeding mechanism 8 includes a support plate 81, a main body 82, a drive module 83, a positioning pin 84, an ejector pin 85, a slide rail 86, and a defective material tray 87. The support plate 81 is fixedly installed on the workbench 9, the main body 82 is fixedly installed on one side of the support plate 81, and the drive module 83 is installed on the main body 82.

[0247] Ejector pin 85 is connected to the moving end of drive module 83, which drives ejector pin 85 to move up and down. Positioning pin 84 is located on the upper part of support plate 81 and is used to position replaceable clamps. Slide 86 is located on one side of main body 82, and the outlet of slide 86 faces defective material tray 87.

[0248] S601, after the test is completed, the No. 2 robotic arm 64 rises and exits the movement range of the sliding cover 75, while simultaneously moving the two-dimensional overlapping magnetic sliding plate to a reset position that does not obstruct the opening action of the sliding cover 75; subsequently, the electromagnetic shielding control module controls the drive cylinder 72 to move the sliding cover 75 to the open position.

[0249] S602, the conveying mechanism 6 drives the No. 3 robot arm 65 to move above the workpiece 55, and the cylinder 66 drives the No. 3 robot arm 65 to descend, so that the suction cup 652 adheres to the workpiece 55; after the negative pressure is established and reaches the set threshold, the No. 3 robot arm 65 picks up the workpiece 55 and rises.

[0250] S603, when the test result of workpiece 55 is qualified, robot arm 65 moves workpiece 55 to the material clamp positioned by positioning pin 84 and releases it, drive module 83 drives ejector pin 85 to move, so that workpiece 55 is loaded into the material clamp in sequence; when the test result of workpiece 55 is unqualified, robot arm 65 moves to the slide 86 and releases workpiece 55, so that workpiece 55 falls into defective material tray 87 along slide 86.

[0251] S604, after the unloading is completed, robot arm 3 65 returns to the initial position; after the clamp is full, the operator releases the positioning and replaces the empty clamp, and then begins the testing process for the next workpiece.

Claims

1. A device for testing the magnetic field distribution on the surface of a permanent magnet, characterized in that, Includes a frame (1) and a worktable (9). A touch control screen (10) is installed on one side of the outer wall of the worktable (9). The upper surface of the worktable (9) is arranged with a staggered separation feeding mechanism (5), a conveying mechanism (6), a magnetic flux detection mechanism (7) and a unloading mechanism (8) in sequence along the conveying direction of the workpiece (55). The conveying mechanism (6) is equipped with a measuring execution end; The magnetic flux detection mechanism (7) includes a housing (71) and a support plate. The support plate inside the housing (71) is provided with a workpiece support position and a calibration area. A standard reference magnet (74) is fixedly installed in the calibration area. A high-permeability shielding sheet that can be opened and closed is also provided above the standard reference magnet (74). The testing device also includes an automated loading and unloading module that is communicatively connected to the misaligned separation loading mechanism (5), the conveying mechanism (6) and the unloading mechanism (8), an electromagnetic shielding control module that is communicatively connected to the magnetic flux detection mechanism (7), and an adaptive testing module that is communicatively connected to the touch control screen (10) and the measurement execution end respectively. The adaptive testing module is used to process spatial coordinate data, calculate motion compensation, and generate the final measurement results.

2. The device for testing the magnetic field distribution on the surface of a permanent magnet according to claim 1, characterized in that, The transport mechanism (6) includes: A gantry frame (61) is fixedly mounted on the workbench (9); The motor (62) is connected to the X-axis linear transmission assembly mounted on the gantry (61) and is used to drive the No. 1 robot (63), the No. 2 robot (64) and the No. 3 robot (65) to move in the horizontal direction; The No. 1 robotic arm (63) and the No. 3 robotic arm (65) are respectively driven to lift by corresponding cylinders (66). The lifting end of the No. 1 robotic arm (63) is detachably connected to the first suction cup (632) through the first buckle (631), and the lifting end of the No. 3 robotic arm (65) is detachably connected to the second suction cup (652) through the second buckle (651). The second manipulator (64) includes a Y-axis linear transmission component connected to the moving end of the X-axis linear transmission component, a Z-axis drive mechanism and an attitude adjustment component disposed at the moving end of the Y-axis linear transmission component, and the attitude adjustment component includes an orthogonally arranged A-axis drive mechanism and a B-axis drive mechanism. A laser displacement sensor (642) and a Hall plate (641) are fixedly installed side by side at the end of the attitude adjustment assembly. The measuring optical axis of the laser displacement sensor (642) and the measuring center axis of the Hall plate (641) are parallel by a known offset distance, and there is a pre-calibrated hardware installation height difference between the two in vertical height. A temperature sensor is mounted close to the Hall plate (641), and its sensing end is fixedly connected to the substrate surface of the Hall plate (641) by thermally conductive adhesive.

3. The device for testing the magnetic field distribution on the surface of a permanent magnet according to claim 2, characterized in that, The top of the outer casing (71) is provided with an opening for the workpiece (55) and the measuring end of the second robotic arm (64) to enter and exit. The magnetic flux detection mechanism (7) further includes: The sliding cover (75) is slidably connected to the guide structure on the top of the outer casing (71); A two-dimensional overlapping magnetic guide plate is stacked on top of the slide cover (75) and supported by a slide frame connected to the X-axis and Y-axis moving ends of the No. 2 robot (64). The lifting end of the No. 2 robot (64) extends into the interior of the outer shell (71) through the central hole on the two-dimensional overlapping magnetic guide plate. A drive cylinder (72) is fixed to the side of the housing (71) and is used to drive the sliding cover (75) to open or close the opening; A sealing strip (73) is provided at the periphery of the opening and at the overlapping position of the sliding cover (75) and the outer shell (71).

4. The device for testing the magnetic field distribution on the surface of a permanent magnet according to claim 1, characterized in that, The staggered separation feeding mechanism (5) includes: The base plate (51) is horizontally fixed above the workbench surface (9); The feeding chute (54) is vertically installed on the front end of the upper surface of the base plate (51); The fine-tuning screw (56) is connected to the adjustable side plate of the feeding slide (54) via a transmission. A hydraulic cylinder (52) is fixedly installed on the base plate (51); The push rod (53) is fixedly connected to the power output end of the hydraulic cylinder (52), and its front end face is horizontally aligned with the bottom discharge port of the feeding slide (54); The material blocking assembly is located at the bottom of the feeding slide (54) and includes a lateral blocking cylinder and a non-magnetic positioning block connected to its output end.

5. The testing apparatus according to claim 1, characterized in that, The feeding mechanism (8) includes: Support plate (81); The main body (82) is fixed to the side of the support plate (81); Drive module (83) mounted on main body (82); Positioning pins (84) for positioning the material clamp; The ejector pin (85) is connected to the moving end of the drive module (83), and the slide (86) leads to the defective material tray (87).

6. A method for testing the magnetic field distribution on the surface of a permanent magnet, characterized in that, The device for testing the magnetic field distribution on the surface of a permanent magnet, as described in any one of claims 1-5, comprises the following steps: The adaptive test module drives the No. 2 robotic arm (64) to move above the standard reference magnet (74) to sample the reference data and calculate the dynamic correction coefficient of the Hall plate (641); The automated loading and unloading module controls the misaligned separation loading mechanism (5) to separate the workpiece (55) and push it into the gripping position; The automated loading and unloading module uses a suction cup (632) to pick up the workpiece (55) and transfer it to the workpiece bearing position on the bearing plate; The electromagnetic shielding control module controls the drive cylinder (72) to drive the sliding cover (75) to move to close and fit the sealing strip (73), forming an electromagnetic shielding space; The adaptive testing module acquires the three-dimensional coordinate data of the workpiece (55) surface, calculates the normal vector and distance deviation of the workpiece (55) surface, and controls the No. 2 robot (64) to perform spatial posture and height compensation. The adaptive test module drives the Hall plate (641) to perform magnetic field scanning at a set lift-off distance along a preset three-dimensional trajectory, reads the original sampling data, substitutes it into the dynamic correction coefficient, performs multiplication, and generates surface magnetic distribution data. The automated loading and unloading module controls the No. 3 robot (65) to transfer the tested workpiece (55) to the unloading mechanism (8) for sorting and unloading.

7. The method for testing the magnetic field distribution on the surface of a permanent magnet according to claim 6, characterized in that, The specific steps for calculating the dynamic correction coefficient include: Obtain the original sampled value of the reference magnetic field on the surface of the standard reference magnet; retrieve the pre-written theoretical reference magnetic flux density value; When the absolute value of the original sampled value is determined to be greater than or equal to the preset zero-point noise threshold, the theoretical reference magnetic induction intensity value is used as the target reference quantity, and the original sampled value is used as the current feedback quantity. A division operation is performed to obtain the dynamic correction coefficient. When the calculated dynamic correction coefficient falls within the preset effective tolerance range, it is written into the system dynamic compensation register, and the current temperature is recorded as the calibration reference temperature.

8. The method for testing the magnetic field distribution on the surface of a permanent magnet according to claim 6, characterized in that, The steps for calculating the deviation between the normal vector and the distance include: Multiple consecutive distance measurement data are extracted using a laser displacement sensor (642). After removing extreme values ​​that exceed three times the standard deviation, the arithmetic mean is calculated and used as the final effective distance value. Read the absolute height of the Z-axis encoder of the current No. 2 robot (64), retrieve the pre-stored hardware installation height difference, call the coordinate calculation model to calculate the height coordinate of the point on the surface of the workpiece (55), and generate three effective non-collinear spatial point cloud coordinates; Two coplanar space vectors are constructed in this way, and the original normal vector is obtained by performing a three-dimensional vector cross product operation. The sign of the Z-axis component of the original normal vector is checked, and each component of the original normal vector is divided by its modulus to obtain the unit space normal vector. Clamping protection logic is applied to each component, and roll and pitch angles are calculated using inverse trigonometric functions.

9. The method for testing the magnetic field distribution on the surface of a permanent magnet according to claim 6, characterized in that, The steps for performing spatial attitude and height compensation include: The kinematic matrix compensation algorithm is invoked to calculate the attitude height compensation value based on the fixed structural offset distance parameters and the current roll and pitch angles. The final target coordinates on the Z-axis are calculated based on the height coordinates, attitude height compensation values, and target lift-off distance, and the Hall plate (641) is driven to move downwards. Extract the current relative physical distance and the target lift-off distance, perform a subtraction operation to solve for the dynamic height deviation; when it is within the safe threshold, input it to the PID closed-loop controller to calculate the dynamic compensation pulse command for the Z-axis; The local normal vector at the current scanning node is obtained to control the A-axis and B-axis to adjust the direction of the Hall plate (641), while the Z-axis is controlled to correct the lifting distance between the Hall plate (641) and the surface of the workpiece (55).

10. The method for testing the magnetic field distribution on the surface of a permanent magnet according to claim 6, characterized in that, The steps for generating the surface magnetic distribution data include: The original magnetic flux density value latched is multiplied by the dynamic correction coefficient to obtain the online calibration magnetic flux density value; The current operating temperature is obtained through the built-in temperature measurement and conditioning circuit. When it is within the effective range, the difference between the temperature and the calibrated reference temperature is used to obtain the temperature deviation. The online calibration magnetic flux density value is compensated by combining the pre-programmed temperature comprehensive compensation coefficient, and the temperature compensation magnetic flux density is calculated. Extract the residual height deviation, and when it is within the tolerance threshold, call the lift-off distance-magnetic induction intensity compensation model to perform height compensation on the current magnetic flux density value; Calculate the first, second, and third powers of the height-compensated magnetic flux density, and multiply them by the corresponding nonlinear weighting coefficients respectively. The product results are summed with the zero-order bias constant to reconstruct the final standard magnetic flux density and store it in the three-dimensional coordinate mapping matrix.

Images