Multi-element X-ray detector automatic calibration system and working method thereof

Abstract

The invention discloses a multi-element X-ray detector automatic calibration system and a working method thereof, and relates to the technical field of nondestructive detection.The multi-element X-ray detector automatic calibration system comprises a box body, a protective cover is arranged on the box body, a Y-axis guide rail is arranged in the protective cover, a Y-axis driving device is arranged on the protective cover, and an X-axis driving device is arranged on a supporting plate and used for driving a sample table to move on an X-axis guide rail on the supporting plate; the detector driving device is arranged on the protective cover and used for driving a detector moving block on a detector guide rail on the top wall of the protective cover to move along the detector guide rail, an X-ray detector is installed at the bottom of the detector moving block through a connecting arm, and a plurality of X-ray tubes are arranged at the bottom of the X-ray detector. A sensor, a vacuum adsorption system and a radiation detector are arranged on the inner rear wall of the protective cover, and a touch display panel is arranged on the outer side of the protective cover. According to the adaptive path planning algorithm, the sample three-dimensional model is analyzed based on deep learning, the optimal calibration path is automatically generated, and the redundant moving distance is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of non-destructive testing, and particularly to an automatic calibration system for a multi-element X-ray detector and its working method. Background Art

[0002] In the calibration process of existing X-ray detectors, the position of the detector is usually adjusted manually, which has problems such as cumbersome operation, low accuracy, and poor repeatability. Especially when detecting complex curved surfaces or large-sized samples, it is difficult to quickly achieve multi-axis collaborative calibration. In addition, existing equipment lacks integrated protection and automatic control modules, is easily interfered by the environment, and has low efficiency. Therefore, there is an urgent need for a device that can automatically complete calibration, support multi-axis adjustment, and have high stability. Summary of the Invention

[0003] The purpose of the present invention is to provide an automatic calibration system for a multi-element X-ray detector and its working method to overcome the above-mentioned defects in the prior art.

[0004] An automatic calibration system for a multi-element X-ray detector includes a box body, a piezoelectric ceramic vibration compensation module, a Y-axis driving device, an X-axis driving device, a detector driving device, a connecting arm, an X-ray detector, a vacuum adsorption system, a radiation detector, and a touch display panel. The piezoelectric ceramic vibration compensation module is embedded in the box body, and a protective cover is provided on it. A Y-axis guide rail is provided in the protective cover. The Y-axis driving device is arranged on the protective cover and is used to drive the pallet to move along the Y-axis guide rail. The X-axis driving device is arranged on the pallet and is used to drive the sample stage to move on the X-axis guide rail on the pallet. The detector driving device is arranged on the protective cover and is used to drive the detector moving block on the detector guide rail on the top wall of the protective cover to move along the detector guide rail. The bottom of the detector moving block is installed with an X-ray detector through a connecting arm. A plurality of X-ray tubes are provided at the bottom of the X-ray detector. A temperature sensor, a TOF sensor, a position sensor, a stress sensor, a vacuum adsorption system, and a radiation detector are also provided in the protective cover. The adsorption end of the vacuum adsorption system is connected to the sample stage. A pressure sensor and a piezoelectric accelerometer are also installed on the sample stage. The touch display panel is arranged on the outside of the protective cover.

[0005] Preferably, the protective cover includes an outer lead alloy shell and an inner polyimide electromagnetic shielding layer. An observation window is provided on the front side of the protective cover.

[0006] Preferably, the thickness of the inner polyimide electromagnetic shielding layer is 0.5 - 1.2 mm.

[0007] Preferably, the connecting arm can help the X-ray detector to perform pitching adjustment within the range of 0° - 180°.

[0008] The working method of the above-mentioned automatic calibration system for a multi-element X-ray detector includes the following steps:

[0009] S1. System initialization and environmental calibration: The hardware conducts self-check. After the device is powered on, the status of each component is detected. Meanwhile, the environmental parameter compensation temperature sensor collects the internal temperature of the box in real time. If the temperature difference > ±0.5 °C, the temperature control module is activated to balance the temperature.

[0010] S2. Sample loading and model import: Place the workpiece to be measured on the sample stage, start the vacuum adsorption system for adsorption, and confirm the fitting degree through the pressure sensor. If the workpiece is an irregular curved surface, switch to the magnetic quick-release fixture to adapt to the groove or convex structure, and upload the graphic model of the workpiece through the touch display panel. The system automatically analyzes the geometric features, and the algorithm generates calibration points.

[0011] S3. Multi-axis collaborative movement and path planning: The calibration path optimization adaptive algorithm sorts the calibration points according to the "Z-shaped" or "spiral" path based on the geometric features of the workpiece to minimize the idle stroke. If the model contains holes or weak areas, the invalid detection points are automatically skipped to avoid the risk of collision. The motion control executes the rough positioning stage and the fine positioning stage.

[0012] S4. X-ray calibration and data acquisition: The multi-wavelength adaptive detection automatically switches the X-ray wavelength according to the material type, and dynamically adjusts the light source power to avoid overexposure or insufficient signal. The X-ray detector collects the diffraction signal with real-time data feedback, and the stress distribution heat map is displayed in real time through the touch display panel, and the out-of-tolerance area is marked.

[0013] S5. Error compensation and closed-loop correction: Multisource data is fused through Kalman filtering for compensatory movement, and the adaptive PID parameters are adjusted.

[0014] S6. Calibration result output and cloud synchronization: Report generation, stress distribution cloud map, maximum / minimum value statistics, key point data table, environmental parameter record.

[0015] Preferably, in step S1, check the status of the motors of the X-axis drive device, Y-axis drive device and detector drive device, verify the stability of the X-ray detector and the X-ray tube light source, and test the airtightness of the vacuum adsorption system. If an abnormality is detected, an alarm is given through the touch display panel and the maintenance items are prompted.

[0016] Preferably, in step S3, during the rough positioning stage of the motion control, the pallet is driven by the Y-axis drive device to quickly move along the Y-axis guide rail to the target area at a moving speed of 500 mm / s, and the X-axis guide rail is advanced horizontally synchronously to cover the full length of the workpiece. During the fine positioning stage, the pitch angle of the X-ray detector is adjusted through the connecting arm to align the probe axis with the normal line of the workpiece surface, and the X-axis drive device can finely adjust the position of the sample stage.

[0017] Preferably, in step S5, the Kalman filter fuses multi-source data and inputs: the motor encoder positions of the Y-axis driving device and the X-axis driving device, temperature sensors, vibration spectra, and X-ray intensities, and outputs: predicted mechanical transmission errors and probe angle offsets. For dynamic correction: if the deviation of the normal direction of the detector is > 0.2°, immediately trigger the fine-tuning of the connecting arm angle; if the accumulated positioning error > ±0.01 mm, pause the calibration and start the X / Y-axis reverse compensation movement.

[0018] Preferably, in the adaptive PID parameter adjustment in step S5, according to the real-time error e = target position - actual position, the formula for dynamically updating the control parameters is as follows:

[0019] K p = K po ·(1 + a|e|);

[0020] K d = K do / (1 + β|e|);

[0021] Where: K po : basic proportionality, K do : differential coefficient, a and β: environmental adaptation factors, obtained through training with historical data. By adjusting the PID output, the steady-state error is converged within ±0.005 mm.

[0022] The beneficial effects achieved by the present invention are as follows:

[0023] 1. The adaptive path planning algorithm of the present application: Analyzes the three-dimensional model of the sample based on deep learning (CNN + RNN), automatically generates the optimal calibration path, reduces the redundant moving distance, and shortens the calibration time. Supports the dynamic obstacle avoidance function, automatically detours when encountering sample protrusions or holes, and avoids the risk of collision. Kalman filter multi-source fusion of sensor data, real-time compensation for mechanical thermal expansion and vibration offset. The algorithm has a short iteration period, ensuring the real-time correction ability under high-speed movement. Cloud collaborative optimization: The calibration data is uploaded to the cloud, and the global parameters are optimized through big data analysis to continuously improve the system performance.

[0024] 2. The high-precision adaptation of the connecting arm of the present application: Dynamic adaptation of the guide rail, supports continuous pitch adjustment from 0° to 180°, and adapts to complex curved surfaces such as aviation blades (curvature radius 150 - 300 mm) and turbine disks. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 It is a schematic structural diagram of the whole of the present invention.

[0026] In the figure, 1 is the box body; 2 is the protective cover; 3 is the Y-axis guide rail; 4 is the pallet; 5 is the sample stage; 6 is the X-axis guide rail; 7 is the detector guide rail; 8 is the detector moving block; 9 is the X-ray detector; 91 is the X-ray tube; 10 is the connecting arm; 11 is the vacuum adsorption system; 12 is the radiation detector; 13 is the touch display panel. Specific implementation mode

[0027] The following is a further detailed description of the specific implementation mode of the present invention by describing the embodiments with reference to the accompanying drawings, so as to help those skilled in the art have a more complete, accurate and in-depth understanding of the concept and technical solution of the present invention.

[0028] As Figure 1 shown, the present invention constructs a two-dimensional motion platform through the orthogonally arranged X-axis guide rail 6 and Y-axis guide rail 3, and combines the micro-step drive of the sample stage 5 to realize the full-plane coverage positioning of the X-ray detector 9 and the sample.

[0029] It is further set that: the touch display panel 13 has a built-in path planning algorithm: automatically generating and optimizing the motion trajectory based on the detection points of the three-dimensional model;

[0030] Three-dimensional model analysis and feature extraction:

[0031] 1. Input data: Import the CAD model (STEP or IGES format) of the workpiece, including geometric topology, surface curvature and key detection areas (such as welds, hole positions).

[0032] 2. Curvature sensitivity analysis: Use differential geometry algorithms (such as Gaussian curvature calculation) to identify high-curvature areas (such as the leading edge of a turbine blade), and preferentially plan dense detection points.

[0033] 3. Key point marking: Based on edge detection (Canny algorithm) and region growing method (Region Growing), mark the stress concentration areas (such as fillet transitions).

[0034] 4. Detection point generation:

[0035] Adaptive mesh division: Adopt a non-uniform grid algorithm (Non-Uniform Grid), and increase the grid density to 1mm×1mm in high-curvature areas and reduce it to 5mm×5mm in flat areas.

[0036] Path optimization goal: Minimize the total moving distance and time, and at the same time ensure that the normal direction of the X-ray detector 9 is perpendicular to the surface (angle error ≤ 0.01°).

[0037] Path planning algorithm:

[0038] 1. Path optimization based on deep reinforcement learning (DRL):

[0039] State space: including the coordinates of the detection points, the pitch angle of the X-ray detector 9, and the joint angles of the robotic arm.

[0040] Action space: the moving speeds of the X / Y axes and the angle adjustment amount of the X-ray detector 9.

[0041] 2. Reward function:

[0042] Positive reward: reducing the path length and decreasing the angular deviation;

[0043] Negative penalty: collision risk and exceeding the mechanical limit.

[0044] 3. Training process: Generate 100,000 sets of random models through a virtual simulation environment (such as Gazebo) to train the neural network (PPO algorithm) to optimize the strategy.

[0045] Dynamic obstacle avoidance function:

[0046] Real-time point cloud processing: Collect the point cloud data of the sample surface through the TOF sensor, compare it with the CAD model, and detect protrusions or depressions.

[0047] RRT (Rapidly-exploring Random Tree) algorithm: When an obstacle is detected, generate a detour path in real time to ensure collision-free movement. And the real-time error compensation algorithm: multi-sensor data fusion and dynamic correction.

[0048] Real-time error compensation algorithm:

[0049] (1) Multi-sensor data fusion

[0050] Temperature sensor: used to monitor and adjust the temperature of the box to make the temperature difference stable (temperature difference < ±0.5°C) to maintain the stability of the detection environment;

[0051] Position sensor: provides real-time feedback on the mechanical position, and monitors the absolute positions of the pallet 4 and the detector moving block 8 in real time, with an accuracy of ±0.001 mm;

[0052] Stress sensor: The X-ray detector 9 calculates the residual stress values on the surface and inside of the sample by analyzing the diffraction angle shift (accuracy ±3 MPa);

[0053] Piezoelectric accelerometer, used to measure the vibration acceleration during the operation of the sample stage and the sample;

[0054] Motor encoders of the X-axis drive device and the Y-axis drive device: 17-bit absolute value encoders with an angular resolution of 0.001°.

[0055] Data synchronization: Ensure the time sequence alignment of multi-source data through the hardware timestamp (PTP protocol).

[0056] (2) Dynamic correction by Kalman Filter

[0057] State equation: X k = AX k-1 + Bu k + W k ;

[0058] Where: X k : State vector (position, angle, temperature drift);

[0059] u k : Control input (motors of the X-axis drive device, motors of the Y-axis drive device, motor drive commands of the detector drive device (including rotational speed, direction commands));

[0060] W k : Process noise (mechanical transmission error);

[0061] X k-1 : State vector at the previous moment.

[0062] Observation equation:

[0063] Z k = HX k + V k ;

[0064] Where:

[0065] H: Observation matrix, used to map the state vector X K to the observation space;

[0066] Z k : Observation values of multiple sensors (including position, stress, temperature);

[0067] V k : Observation noise (errors of position, stress, temperature sensors).

[0068] Real-time correction process:

[0069] 1. Prediction: Predict the state at the next moment according to the instructions of the touch display panel 13;

[0070] 2. Update: Integrate the data of position, stress, and temperature sensors, calculate the Kalman gain, and correct the prediction error;

[0071] 3. Output: Generate compensation instructions (such as the angle offset of the X-ray detector 9).

[0072] (3) Adaptive PID parameter adjustment

[0073] According to the real-time error e = target position - actual position, the formula for dynamically updating the control parameters is:

[0074] K p = K po ·(1 + a|e|);

[0075] K d = K do / (1 + β|e|);

[0076] Where: K po : Base ratio, K do : Differential coefficient, a and β: Environmental adaptation factors, obtained through training with historical data, and by adjusting the PID output, the steady-state error is converged within ±0.005 mm.

[0077] The protective cover 2 uses a 5 mm lead alloy outer shell (radiation attenuation rate ≥ 99%) + 1 mm polyimide electromagnetic shielding layer, and a visible window with a lead equivalent of 0.5 mmPb is provided at the front end. The piezoelectric ceramic vibration compensation module is embedded in the box body 1, with a response frequency of 0 - 200 Hz to offset external vibration interference; the sensor monitors environmental parameters in real time and triggers active adjustment.

[0078] The X-ray detector 9 integrates a tungsten target (K α = 0.021 nm) and a molybdenum target (K α = 0.071 nm) dual light sources, the electromagnetic drive shutter switching time < 0.1 s, and the wavelength stability error < ±0.015 nm. The surface of the sample stage 5 is provided with a honeycomb-shaped vacuum adsorption hole array (hole diameter 1 mm, pitch 5 mm), and the adsorption pressure is adjustable in stages (-50 kPa to -90 kPa), suitable for carbon fiber and titanium alloy materials.

[0079] Automatically generate a calibration path: 1. Model format support: The touch display panel 13 has a built-in CAD parsing engine, supports mainstream 3D model formats (such as STEP, IGES, STL), and automatically extracts the following key geometric features: Surface curvature: Calculate the local curvature radius (R) and identify high-curvature regions (such as blade edges, welded joints). Normal direction: Generate the normal vector of each detection point based on the grid vertex data (for aligning the X-ray detector 9).

[0080] 2. Dynamically select the optimal path mode according to the model features: Zigzag scan: Suitable for flat or low-curvature surfaces, covering the detection area at a fixed pitch (default 2 mm). Spiral path: For circular or rotary workpieces (such as turbine disks), spiral outwards from the center to reduce the empty travel. Layered detection: For multi-layer structures (such as composite material laminates), generate paths by depth layering (interval of 0.5 mm per layer). Dynamic obstacle avoidance and optimization: Use the AABB (axis-aligned bounding box) algorithm to detect potential interference between the probe and the workpiece, and automatically skip dangerous points or adjust the path height.

[0081] 3. Curvature-driven: The calibration point density is doubled (spacing 1 mm) in the high-curvature region (R < 50 mm), and relaxed to 3 mm in the low-curvature region (R > 200 mm). Stress weight: According to historical data or simulation results, increase the detection point density in areas prone to residual stress (such as welds, holes).

[0082] 4. Motion parameter matching: Speed grading: High-speed movement (500 mm / s) is used in flat areas, and the speed is reduced to 100 mm / s on complex curved surfaces to ensure accuracy. Angle prediction: According to the change rate of the normal direction of adjacent detection points, the pitch angle of the X-ray detector 9 is adjusted in advance (prediction error < 0.05°). Energy parameter optimization: X-ray intensity: Automatically calculate the optimal voltage / current based on the material thickness (model analysis) (e.g., for titanium alloy: 50 kV / 10 mA, for carbon fiber: 30 kV / 5 mA). Multi-wavelength switching: For multi-layer materials, plan the detection order of different wavelengths (e.g., first scan the surface layer with Kα rays, and then penetrate the deep layer with L-series rays).

[0083] 5. Online calibration verification: After every 10 detection points are completed, the system compares the measured stress value with the model prediction value. If the deviation > 10%, trigger path re-planning. Fusion of encoder position and temperature drift data through a Kalman filter to correct the coordinates of the remaining path points in real time (accuracy compensation ±0.003 mm).

[0084] Multi-axis collaborative motion control: Coarse positioning stage: The Y-axis drive device drives the pallet 4 to move quickly along the Y-axis guide rail 3 to the target area (speed 500 mm / s, error ±5 μm). The X-axis guide rail 6 is advanced horizontally synchronously to cover the full length of the workpiece (straightness error < 0.01 mm / m). Fine positioning stage: Adjust the pitch angle (0° - 180°) of the X-ray detector 9 through the connecting arm 10 to ensure that the probe axis is aligned with the surface normal (angle error < 0.1°). The X-axis drive device finely adjusts the position of the sample stage 5, and the final positioning accuracy reaches ±0.005 mm. Motion parameter adaptation: Speed grading control: High speed (500 mm / s) in flat areas, speed reduction in complex curved surfaces (100 mm / s). Angle prediction compensation: According to the change rate of the normal direction of adjacent points, adjust the probe angle in advance (prediction error < 0.05°). The inner layer of the protective cover 2 is a polyimide shielding layer (1 mm) to isolate high-frequency noise. The vacuum adsorption system 11 generates a reverse waveform in real time to cancel vibrations of 20 - 1000 Hz.

[0085] Working principle and process:

[0086] Turn on the power of the starting device, load the system control program through the touch interface of the touch display panel 13, and initialize the motors of the X-axis drive device, the Y-axis drive device, the motor of the detector drive device, and each sensor module. Perform reference calibration using a standard calibration sample: Fix the standard calibration sample on the surface of the sample stage 5, start the vacuum adsorption system 11 (set pressure -80 kPa), and ensure no displacement. Perform zero calibration: The X-ray detector 9 returns to zero along the detector guide rail 7 (pitch angle 0°), and the sample stage 5 is moved to the mechanical origin by the motors of the X-axis drive device and the Y-axis drive device. Place the workpiece to be measured (such as an aviation aluminum alloy turbine blade) on the sample stage 5, and select the adsorption pressure (-60 kPa to -90 kPa) according to the surface curvature to ensure that the contact surface fits without warping. Import the 3D CAD model of the workpiece into the touch display panel 13, set the boundary conditions of the detection area (such as the blade tenon groove and the blade body area), and the system automatically generates a calibration path based on the adaptive calibration algorithm, and plans the X / Y-axis movement trajectory and the pitch angle sequence of the X-ray detector 9. Select the X-ray wavelength: Select the molybdenum target Kα ray (λ = 0.071 nm) through the multi-wavelength switching module. Since it is necessary to match the X-ray wavelength with the crystal plane spacing, it is necessary to adapt to the diffraction requirements of the aluminum alloy lattice, and set the molybdenum target tube x-ray voltage to 35 kV and the molybdenum target X-ray tube 91 current to 20 mA. Start the automatic calibration mode: The Y-axis drive device drives the support plate 4 to move along the Y-axis guide rail 3 to the starting coordinate (Y = 0 mm), and the X-axis drive device drives the sample stage 5 to move horizontally to the first detection point (X = 10 mm, Y = 5 mm). Drive the X-ray detector 9 along the detector moving block 8 through the connecting arm 10 to adjust to the normal direction (θ = 45°), and synchronously trigger the laser rangefinder to verify the distance between the detector and the sample surface (set value 50 mm ± 0.1 mm). Emit X-rays and collect diffraction signals: The detector receives the crystal plane diffraction peak, calculates the peak offset through the energy spectrum analysis module, and combines the stress-strain model (such as sin 2 ψ method) to invert the residual stress value, and the data is transmitted to the touch display panel 13 in real time. Dynamic error compensation: The piezoelectric ceramic vibration compensation module in the box 1 cancels the external vibration interference based on the piezoelectric ceramic driver (response frequency 1 kHz), and the temperature sensor feeds back the environmental data to the Kalman filter to correct the mechanical drift caused by thermal expansion. After completing the full-area calibration, the system automatically generates a stress distribution cloud map and a statistical report (including mean value, standard deviation, and maximum deviation), and uploads them to the MES system through the Wi-Fi module. Repeatability verification: Perform calibration on the same detection point three times, calculate the repeatability error (RSD ≤ 0.8%), and if it exceeds the limit, trigger an alarm and prompt for re-calibration. Before shutting down the system, the X-ray detector 9 returns to the safe position, the protective cover 2 closes, and the radiation dose detector confirms that the environmental dose rate ≤ 0.5 μSv / h.

[0087] The embodiments of the present invention described above do not constitute a limitation on the protection scope of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the claims of the present invention.

Claims

1. An automatic calibration system for a multi - element X - ray detector, comprising a box body (1), a piezoelectric ceramic vibration compensation module, a Y - axis driving device, an X - axis driving device, a detector driving device, a connecting arm (10), an X - ray detector (9), a vacuum adsorption system (11), a radiation detector (12) and a touch display panel (13), characterized in that: The box body (1) is embedded with a piezoelectric ceramic vibration compensation module and is provided with a protective cover (2) thereon. A Y-axis guide rail (3) is provided in the protective cover (2). The Y-axis driving device is arranged on the protective cover (2) and is used to drive the pallet (4) to move along the Y-axis guide rail (3). The X-axis driving device is arranged on the pallet (4) and is used to drive the sample stage (5) to move on the X-axis guide rail (6) on the pallet (4). The detector driving device is arranged on the protective cover (2) and is used to drive the detector moving block (8) on the detector guide rail (7) on the top wall of the protective cover (2) to move along the detector guide rail (7). The bottom of the detector moving block (8) is installed with an X-ray detector (9) through a connecting arm (10). A plurality of X-ray tubes (91) are provided at the bottom of the X-ray detector (9). A temperature sensor, a TOF sensor, a position sensor, a stress sensor, a vacuum adsorption system (11) and a radiation detector (12) are also provided in the protective cover (2). The adsorption end of the vacuum adsorption system (11) is connected to the sample stage (5). A pressure sensor and a piezoelectric accelerometer are also installed on the sample stage (5). A touch display panel (13) is provided on the outer side of the protective cover (2).

2. The automatic calibration system of a multi - element X - ray detector according to claim 1, characterized in that: The protective cover (2) includes an outer lead alloy shell and an inner polyimide electromagnetic shielding layer. An observation window is provided on the front side of the protective cover (2).

3. The automatic calibration system of a multi - element X - ray detector according to claim 2, characterized in that: The thickness of the inner polyimide electromagnetic shielding layer is 0.5 - 1.2 mm.

4. The automatic calibration system of a multi - element X - ray detector according to claim 1, characterized in that: The connecting arm (10) can help the X-ray detector (9) perform pitching adjustment within the range of 0° - 180°.

5. A working method of an automatic calibration system for a multi - element X - ray detector as described in any one of claims 1 - 4, characterized in that: It includes the following steps: S1. System initialization and environment calibration: The hardware performs self-check. After the device is powered on, the states of each component are detected. At the same time, the environmental parameter compensation temperature sensor real-time collects the internal temperature of the box body (1). If the temperature difference > ±0.5 °C, the temperature control module is started to balance the temperature; S2. Sample loading and model import: Place the workpiece to be measured on the sample stage (5), start the vacuum adsorption system (11) for adsorption, confirm the fitting degree through the pressure sensor. If the workpiece is an irregular curved surface, switch to the magnetic quick-release fixture to adapt to the groove or protrusion structure, upload the graphic model of the workpiece through the touch display panel (13), the system automatically analyzes the geometric features, and the algorithm generates calibration points; S3. Multi-axis collaborative movement and path planning: The calibration path optimization adaptive algorithm sorts the calibration points according to the "Z-shaped" or "spiral-shaped" path according to the geometric features of the workpiece, minimizes the idle stroke. If the model contains holes or weak areas, automatically skip the invalid detection points to avoid collision risks. The motion control executes the rough positioning stage and the fine positioning stage; S4. X-ray calibration and data collection: The multi-wavelength adaptive detection automatically switches the X-ray wavelength according to the material type, dynamically adjusts the light source power to avoid overexposure or insufficient signal. The real-time data feedback X-ray detector (9) collects the diffraction signal, and the stress distribution heat map is displayed in real time through the touch display panel (13), and the out-of-tolerance area is marked; S5. Error compensation and closed-loop correction: Fuse multi-source data through Kalman filtering, perform compensated movement, and adaptively adjust the PID parameters; S6. Calibration result output and cloud synchronization: report generation, stress distribution cloud map, maximum / minimum value statistics, key point data table, environmental parameter recording.

6. The working method of an automatic calibration system for a multi-element X-ray detector according to claim 5, characterized in that: In step S1, check the status of the motors of the X-axis drive device, Y-axis drive device, and probe drive device, verify the light source stability of the X-ray detector (9) and the X-ray tube (91), and test the airtightness of the vacuum adsorption system (11). If any abnormality is detected, alarm through the touch display panel (13) and prompt the repair items.

7. The working method of an automatic calibration system for a multi - element X - ray detector according to claim 5, characterized in that: In step S3, during the rough positioning stage of motion control, drive the pallet (4) to quickly move along the Y-axis guide rail (3) to the target area through the Y-axis drive device at a moving speed of 500 mm / s, and the X-axis guide rail (6) is advanced horizontally synchronously to cover the full length of the workpiece. During the fine positioning stage, adjust the pitch angle of the X-ray detector (9) through the connecting arm (10) to align the probe axis with the normal of the workpiece surface, and the X-axis drive device can finely adjust the position of the sample stage (5).

8. The working method of an automatic calibration system for a multi-element X-ray detector according to claim 5, characterized in that: In step S5, the input in the Kalman filter for fusing multi-source data includes: the motor encoder positions of the Y-axis drive device and the X-axis drive device, temperature sensors, vibration spectra, and X-ray intensities. The output includes: predicted mechanical transmission errors and probe angle offsets. For dynamic correction: if the deviation of the detector normal direction > 0.2°, immediately trigger the adjustment of the angle of the connecting arm (10). If the accumulated positioning error > ±0.01 mm, pause the calibration and start the X / Y-axis reverse compensation movement.

9. The working method of an automatic calibration system for a multi - element X - ray detector according to claim 5, characterized in that: In step S5, during the adaptive PID parameter adjustment, according to the real-time error e = target position - actual position, the formula for dynamically updating the control parameters is as follows: K p = K po ·(1 + ae); K d = K do / (1 + β|e|); Among them: K po : Base ratio, K do : Differential coefficient, a and β: Environmental adaptation factors, obtained through training with historical data. By adjusting the PID output, the steady-state error is converged within ±0.005 mm.

Images