Hollow shaft inner wall defect detection method

Abstract

The invention provides a hollow shaft inner wall defect detection method, which comprises the following steps: obtaining a detection demand of the inner wall of a hollow shaft, the detection demand at least comprising a fitting state of a reducing inner wall; the posture of the detection device is adjusted through a mechanical arm, and the mechanical arm moves to a target position according to the control instruction; detecting the defect of the inner wall of the hollow shaft through an electromagnetic array eddy current sensor, wherein the electromagnetic array eddy current sensor is attached to the inner wall and generates a detection signal; acquiring fitting state data of the electromagnetic array eddy current sensor and the inner wall through a linear Hall sensor, wherein the fitting state data represents the distance between the detection device and the inner wall; and processing the detection signal through an intelligent eddy current detector and generating a defect detection result, wherein the defect detection result represents the position and the characteristic of the inner wall defect.

Description

Technical Field

[0001] The present invention relates to the field of information technology, and in particular to a method for detecting defects on the inner wall of a hollow shaft. Background Art

[0002] Defect detection on the inner wall of variable-diameter hollow shafts presents a series of complex technical challenges. The primary challenge is achieving high-precision defect identification within a non-uniform bore structure. Due to the varying inner diameter of the hollow shaft, traditional inspection methods struggle to maintain consistent contact between the inspection device and the inner wall, resulting in inaccurate test results. Secondly, the variable diameter structure complicates the positioning and posture adjustment of the inspection device, necessitating real-time calibration of the inspection position and dynamic adjustment of the inspection device's posture. Furthermore, efficient detection of defects distributed along the entire circumference of the hollow shaft's inner wall presents a significant challenge, requiring comprehensive coverage within a limited space. These interrelated issues collectively constitute the core technical challenges of intelligent defect detection for variable-diameter hollow shaft inner walls. Ensuring accurate, comprehensive, and efficient inspection within complex internal structures while overcoming the various interferences introduced by variable diameters are key challenges in this field. This requires not only innovative detection principles but also the flexibility and adaptability of the inspection device and intelligent data processing to meet the urgent demands of modern industry for high-quality and efficient inspection. Summary of the Invention

[0003] The present invention provides a method for detecting defects on the inner wall of a hollow shaft, which mainly includes: Obtaining the inspection requirements of the inner wall of the hollow shaft, the inspection requirements at least include the fitting state of the variable diameter inner wall; adjusting the posture of the inspection device by a robotic arm, and the robotic arm moves to the target position according to the control instruction; detecting defects on the inner wall of the hollow shaft by an electromagnetic array eddy current sensor, and the electromagnetic array eddy current sensor fits the inner wall and generates a detection signal; obtaining the fitting state data of the electromagnetic array eddy current sensor and the inner wall by a linear Hall sensor, and the fitting state data represents the distance between the inspection device and the inner wall; processing the detection signal by an intelligent eddy current detector and generating a defect detection result, and the defect detection result represents the position and characteristics of the inner wall defect. Furthermore, the method of detecting defects on the inner wall of a hollow shaft by using an electromagnetic array eddy current sensor includes: adopting an electromagnetic array eddy current sensor with a double-row S-shaped topology layout, wherein the double-row S-shaped topology layout consists of a first-row four-element detection array and a second-row three-element detection array; forming a spatial phase difference distribution by using the first-row four-element detection array and the second-row three-element detection array; constructing a spiral progressive detection path by embedding a short array in a long array gap; generating a consistent detection field strength by using a sensing unit designed with equal parameters; and optimizing the magnetic circuit and generating the detection signal by using magnetic domain orientation processing of the magnetic core. Furthermore, obtaining data on the fit between the electromagnetic array eddy current sensor and the inner wall using a linear Hall sensor includes: detecting changes in the primary magnetic field, which is excited by the electromagnetic array eddy current sensor, using the linear Hall sensor; obtaining the magnitude of the induced eddy currents based on the changes in the primary magnetic field; determining the fit between the electromagnetic array eddy current sensor and the inner wall based on the relationship between the induced eddy currents and the primary magnetic field; processing the change data on the primary magnetic field through filtering, noise reduction, and signal amplification; generating an analog voltage signal related to the distance through analog-to-digital conversion and transmitting it to a host computer. Furthermore, adjusting the posture of the detection device using a robotic arm includes: obtaining the fit data through a host computer and calculating the error between the actual distance and the target distance; generating a control variable based on the error using a PID algorithm; determining the direction and amplitude of the robotic arm's movement based on the control variable; calculating the target angles of each robotic arm joint using inverse kinematics; driving the robotic arm joints to the target angles using motors; obtaining the actual angles using encoders and comparing them with the target angles for fine-tuning. Furthermore, the processing of the detection signal by the intelligent eddy current detector and the generation of defect detection results include: obtaining the detection signal by the intelligent eddy current detector; extracting the defect characteristics of the inner wall of the hollow shaft according to the detection signal; generating imaging data of the inner wall defect by signal processing; determining the position and size of the defect by the imaging data; and transmitting the imaging data to a host computer to generate a visual detection result.Furthermore, the method of adjusting the posture of the detection device by a robotic arm includes: fixing the hollow shaft with a dedicated clamping mechanism and driving it to rotate about the central axis; delivering the detection device to a designated detection point on the inner wall of the hollow shaft by the robotic arm; continuously acquiring distance data between the detection device and the inner wall by the linear Hall sensor; adjusting the motion trajectory of the robotic arm based on the distance data; and detecting defect signals at the designated detection point by the electromagnetic array eddy current sensor. Furthermore, the method of detecting defects on the inner wall of the hollow shaft by the electromagnetic array eddy current sensor includes: exciting a primary magnetic field by the electromagnetic array eddy current sensor; inducing eddy currents on the inner wall of the hollow shaft based on the primary magnetic field; generating impedance changes by reacting the induced eddy currents to the excitation coil; determining changes in coil voltage based on the impedance changes; and analyzing the changes in coil voltage by the intelligent eddy current detector to generate defect information.

[0004] Furthermore, the electromagnetic array eddy current sensor is a flat flexible electromagnetic array eddy current sensor.

[0005] Furthermore, the linear Hall sensor is a closed-loop linear Hall sensor.

[0006] The technical solution provided by the embodiment of the present invention may have the following beneficial effects: The present invention discloses an intelligent detection method for inner wall defects of variable diameter hollow shafts. In order to solve the problem of hollow shaft detection with non-equal diameter inner hole structure, the present invention adopts an integrated detection device loaded by a robotic arm, combined with a linear Hall sensor and an S-shaped array eddy current sensor, to achieve precise fitting of the detection device with the inner wall. The inner wall defect information is obtained through the eddy current induction principle, and the detection position is calibrated in real time using the Hall sensor to ensure the detection accuracy. The present invention also innovatively uses the PID algorithm and the inverse kinematics calculation of the robotic arm to dynamically adjust the posture of the detection device, solving the detection difficulty brought by the variable diameter structure. Combined with the axial propulsion controlled by the servo motor and the S-shaped sensor topology layout, efficient detection of the full circumferential defect distribution of the inner wall of the hollow shaft is achieved. The present invention significantly improves the accuracy and efficiency of the detection of defects on the inner wall of the variable diameter hollow shaft, and provides strong support for quality control in related fields. BRIEF DESCRIPTION OF THE DRAWINGS

[0007] Figure 1 This is a flow chart of a method for detecting defects on the inner wall of a hollow shaft according to the present invention. In order to make the objectives, technical solutions and advantages of the present invention more clear, the present invention is described in detail below with reference to the accompanying drawings and specific embodiments.

[0008] like Figure 1 In this embodiment, a method for detecting defects on the inner wall of a hollow shaft may specifically include: Step S101 : obtaining a variable diameter hollow shaft sample to be tested, wherein the sample has a non-uniform diameter inner hole structure, and fixing the hollow shaft by a dedicated clamping mechanism.

[0009] Sample data of a variable-diameter hollow shaft to be inspected is obtained, and the geometric parameters of the non-uniform inner bore structure are generated through digital modeling. Based on the geometric parameters, the distribution of variable-diameter feature points of the hollow shaft's inner bore is determined, generating a set of feature point coordinates. Using this set of feature point coordinates, the initial clamping position of the dedicated clamping mechanism is calculated, obtaining the spatial distribution data of the clamping points. Based on this spatial distribution data, motion control instructions for the clamping mechanism are generated, and clamping posture adjustment parameters are determined. Based on the clamping posture adjustment parameters, the dedicated clamping mechanism is driven to perform positioning operations, obtaining the fixed state data of the hollow shaft. Based on this fixed state data, the degree of fit between the clamping mechanism and the inner wall of the hollow shaft is detected, and the fit deviation value is determined. If the fit deviation value exceeds a preset threshold, the controller adjusts the clamping mechanism's posture to obtain optimized fit state data. Based on this optimized fit state data, real-time signals from the Hall effect element and detection coil are collected to obtain preliminary inner wall defect detection results. This preliminary detection result is then integrated with the eddy current detection array channel data to generate accurate distribution information of the hollow shaft's inner wall defects.

[0010] Step S102: Use a robotic arm to load an integrated detection device, wherein the integrated detection device includes a linear Hall sensor and an S-type array eddy current sensor, and extend the detection device into the inner hole of the hollow shaft.

[0011] The robotic arm receives control commands, adjusts its inspection posture, and loads the integrated inspection device. The integrated inspection device extends into the inner bore of the hollow shaft, activating the linear Hall sensor and S-shaped array eddy current sensor. The linear Hall sensor detects changes in the magnetic induction of the induction coil and obtains distance data from the inner wall of the hollow shaft. Based on the magnetic induction change data, an algorithm is used to determine the degree of fit between the inspection device and the inner wall of the hollow shaft. If the fit is below a set threshold, the position of the inspection device is adjusted based on the structural parameters of the robotic arm. The S-shaped array eddy current sensor scans the inner surface of the hollow shaft and collects defect signals at specified points. The defect signals are received by the intelligent eddy current detector to obtain preliminary defect distribution data for the inner bore of the hollow shaft. After the hollow shaft completes a full rotation inspection, the robotic arm gradually advances the inspection device along the axis. Flaw signals are obtained at each point during the advancement process, and through data fusion, the complete defect distribution of the inner bore of the hollow shaft is determined. The linear Hall sensor can be a closed-loop type, and the electromagnetic array eddy current sensor can be a flat, flexible electromagnetic array eddy current sensor.

[0012] Specifically, after receiving control commands, the robotic arm uses a PID control algorithm to adjust its detection posture, setting a position accuracy of ±0.1mm, and then installs an integrated detection device. The detection device extends into the inner bore of the hollow shaft, activating a linear Hall effect sensor and an S-shaped array eddy current sensor. The linear Hall effect sensor detects changes in the magnetic induction of the induction coil at a sampling frequency of 10kHz, acquiring distance data from the inner wall of the hollow shaft with a measurement range of 0.5mm to 5mm. The magnetic induction change data is processed using a Kalman filter algorithm to calculate the fit between the detection device and the inner wall. If the fit is less than 90%, the position of the detection device is adjusted using an inverse kinematics algorithm based on the robotic arm's six degrees of freedom parameters to ensure that the fit meets the required standards. The S-shaped array eddy current sensor scans the inner bore surface at a 5mm interval, acquiring defect signals. The signals are then subjected to an FFT transform to extract characteristic frequency components. After receiving the signals, the intelligent eddy current detector uses a support vector machine algorithm to classify the defect type and generate a preliminary defect distribution map. After the hollow shaft completes a 360° rotation test, the robotic arm advances along the axis in 1mm steps, recording the defect signals at each point in real time. Finally, multiple sets of data are fused through the Gaussian mixture model to generate a complete three-dimensional defect distribution map of the inner hole of the hollow shaft.

[0013] Step S103 , exciting the primary magnetic field by the array eddy current sensor to induce eddy currents on the inner wall surface of the hollow shaft, wherein the induced eddy currents react to the impedance of the excitation coil to obtain an impedance change signal.

[0014] An alternating current is applied to the excitation coil of the array eddy current sensor to generate a primary magnetic field. Eddy currents are induced by the action of the primary magnetic field on the inner wall surface of the hollow shaft. The secondary magnetic field generated by the induced eddy current is obtained and reacted to the excitation coil. The excitation coil receives the influence of the secondary magnetic field and changes the impedance of the coil. An intelligent eddy current detector is used to collect impedance changes and obtain a voltage signal. If the voltage signal exceeds the preset threshold, the change is filtered by the signal processing module to determine the preliminary location of the defect. Based on the filtered signal data and the fitting status information collected by the linear Hall sensor, the distance between the detection device and the inner wall is determined. The judgment result is received by the controller, and the posture of the detection device is adjusted. The adjusted posture data is obtained, the detection parameters of the array eddy current sensor are updated, and accurate defect information is obtained.

[0015] Specifically, an alternating current with a frequency of 100kHz is applied to the excitation coil of the array eddy current sensor, generating a primary magnetic field with an intensity of 0.5T. The primary magnetic field acts on the inner surface of the hollow shaft, inducing eddy currents with a frequency of 100kHz and an amplitude of 10mA. The induced eddy currents generate a secondary magnetic field with an intensity of 0.1T, which reacts back to the excitation coil. The excitation coil receives the influence of the secondary magnetic field, changing the coil impedance within a range of 5Ω to 10Ω. An intelligent eddy current detector is used to detect this impedance change, generating a voltage signal with an amplitude of 0.1V to 0.5V. If the voltage signal exceeds a preset threshold of 0.3V, the signal processing module performs a low-pass filter with a cutoff frequency of 1kHz, determining the initial defect location as 50mm from the detection starting point. Based on the filtered signal data and the alignment status information collected by the linear Hall effect sensor, the distance between the detection device and the inner wall is determined to be 0.5mm. The controller receives this determination and adjusts the detection device's posture by a 5° angle. The adjusted posture data was obtained, the detection parameters of the array eddy current sensor were updated, the excitation frequency was adjusted to 120 kHz, and accurate defect information was obtained. The defect depth was 0.2 mm and the length was 10 mm.

[0016] In step S104, a linear Hall sensor is used to detect the change in the primary magnetic field, and an analog voltage signal related to the distance from the inner wall of the hollow shaft is output. The signal is processed by filtering, noise reduction, and analog-to-digital conversion to obtain digital distance data.

[0017] A linear Hall sensor detects changes in the primary magnetic field and outputs an analog voltage signal related to the distance from the inner wall of the hollow shaft. The analog voltage signal is filtered and processed for noise reduction, resulting in a smoothed analog signal with noise removed. The analog signal is then smoothed through signal amplification and normalization, resulting in a standardized analog signal with consistent amplitude. The standardized analog signal is processed through analog-to-digital conversion to generate digitized distance data. The host computer software receives this digitized distance data and obtains the actual distance dactual between the Hall element and the inner wall of the hollow shaft. Based on preset parameters, the target distance dtarget is determined when the detection device and the hollow shaft are in close contact. The difference between the actual distance dactual and the target distance dtarget is determined. The error, e(t), is calculated as dtarget - dactual to obtain deviation data on the contact between the detection device and the inner wall of the hollow shaft. The deviation data, e(t), is processed using a PID algorithm to generate the control variable u(t). The control variable u(t) is calculated using inverse kinematics to determine the target angles for each joint of the robotic arm.

[0018] Step S105 , calculating the error between the actual distance between the detection device and the inner wall of the hollow shaft and the target distance based on the digitized distance data, generating a control variable through a PID algorithm, and determining the adjustment direction and amplitude of the robotic arm.

[0019] The Hall effect sensor detects changes in the primary magnetic field and outputs an analog voltage signal related to distance. The analog voltage signal is filtered and de-noised to obtain a de-noised voltage signal. The de-noised voltage signal is amplified and normalized to obtain a standardized analog signal. The standardized analog signal is processed through analog-to-digital conversion to obtain digitized distance data. Based on the digitized distance data, the actual distance dactual between the detection device and the inner wall of the hollow shaft is calculated to determine the actual distance value. The target distance dtarget, when the detection device is tightly fitted to the hollow shaft, is obtained to determine the target distance value. The error between the actual distance and the target distance is calculated using the formula e(t) = dtarget - dactual to obtain the error value e(t). Based on the error value e(t), the PID algorithm generates the control variable u(t) to determine the direction and amplitude of the robot arm adjustment. Using inverse kinematics, the target angles of each joint are calculated based on the control variable u(t) to obtain the target parameters for the robot arm adjustment.

[0020] Specifically, the Hall sensor detects changes in the primary magnetic field and outputs an analog voltage signal related to distance. For example, when the distance between the detection device and the inner wall of the hollow shaft is 2 mm, the sensor outputs a 0.5 volt voltage signal. The analog voltage signal is filtered and de-noised using a Butterworth low-pass filter with a cutoff frequency set to 100 Hz to remove high-frequency noise and obtain a de-noised voltage signal. The de-noised voltage signal is amplified and normalized using an operational amplifier to amplify the signal to a range of 0 to 5 volts, thereby obtaining a standardized analog signal. The standardized analog signal is processed through analog-to-digital conversion using a 12-bit ADC chip with a sampling rate of 1 kHz to convert the analog signal into digitized distance data. Based on the digitized distance data, the actual distance dactual between the detection device and the inner wall of the hollow shaft is calculated. Linear interpolation is used to map the voltage value to a distance value, for example, 0.5 volt corresponds to 2 mm, to determine the actual distance value. The target distance dtarget, which is the distance between the Hall element and the inner wall of the hollow shaft when the detection device is in close contact with the hollow shaft, is determined. The target distance is set to 1 mm, and the target distance value is then determined. The error between the actual distance and the target distance is calculated using the formula e(t) = dtarget - dactual. For example, if the target distance is 1 mm and the actual distance is 2 mm, the error e(t) is -1 mm, and the error value e(t) is obtained. Based on the error value e(t), the control variable u(t) is generated using the PID algorithm. The proportional coefficient Kp is set to 0.5, the integral coefficient Ki is set to 0.1, and the differential coefficient Kd is set to 0.2. The calculated control variable u(t) is -0.6, and the direction and amplitude of the robot arm adjustment are determined. Based on the control variable u(t), the target angles of each joint are calculated using inverse kinematics, and the Jacobian matrix is ​​used to solve the joint angles. For example, when the control variable u(t) is -0.6, the angle of joint 1 is 30 degrees and the angle of joint 2 is 45 degrees, thus obtaining the target parameters for the robot arm adjustment.

[0021] Step S106, calculating the target angle of each joint through inverse kinematics of the robotic arm, and driving the robotic arm to adjust the posture of the detection device. The adjustment is based on the link coordinate system transformation established by the DH parameters to obtain the detection position that fits the inner wall.

[0022] The connecting rod coordinate system of the hollow shaft inner wall detection device is established using the DH parameters to determine the relative position and orientation of each joint. The initial position and posture of the detection device's end relative to the hollow shaft inner wall are calculated based on the current joint angles of the robotic arm and the transformation of the connecting rod coordinate system. The analog voltage signal output by the Hall effect sensor is filtered and noise-reduced to produce an electrical signal related to the actual distance between the detection device and the hollow shaft inner wall. The electrical signal is processed through signal amplification, standardization, and analog-to-digital conversion to obtain the actual distance dactual between the detection device and the hollow shaft inner wall. Based on the preset target distance dtarget when the detection device is in close contact with the hollow shaft inner wall, the error e(t) is calculated as dtarget - dactual. The PID algorithm is used to process the error e(t) and generate the control variable u(t), which determines the direction and amplitude of the robotic arm's adjustment. The control variable u(t) is solved using inverse kinematics to calculate the target angles for each joint of the robotic arm. The robotic arm is driven according to the target angles of each joint, adjusting the detection device's posture to achieve the detection position that is in close contact with the hollow shaft inner wall. The new signal output by the adjusted Hall sensor is obtained, and through analog-to-digital conversion processing, it is determined whether the fit between the detection device and the inner wall of the hollow shaft reaches the target distance dtarget.

[0023] Specifically, the connecting rod coordinate system of the hollow shaft inner wall detection device was established using the DH parameters. The initial angles of each joint of the robotic arm were set to θ1 = 0°, θ2 = 90°, and θ3 = 45°, thereby determining the relative position and orientation of each joint. Based on the current joint angles of the robotic arm and the transformation of the connecting rod coordinate system, a homogeneous transformation matrix was used to calculate the initial position of the detection device end relative to the hollow shaft inner wall: (x = 100 mm, y = 50 mm, z = 200 mm), and the posture: (α = 0°, β = 0°, γ = 0°). The analog voltage signal output by the Hall effect sensor was obtained, with an amplitude of 2.5V. This signal was filtered and denoised using a Butterworth low-pass filter with a cutoff frequency set to 10Hz. This electrical signal, representing the actual distance between the detection device and the hollow shaft inner wall, was obtained. This signal was amplified to 5V by a signal amplifier circuit and standardized. A 12-bit analog-to-digital converter was used to convert the analog signal into a digital signal, resulting in the actual distance dactual = 2.1mm between the detection device and the hollow shaft inner wall. Based on the preset target distance dtarget = 2.0 mm when the detection device is in close contact with the inner wall of the hollow shaft, the error e(t) is calculated as 2.0 mm - 2.1 mm = -0.1 mm. A PID algorithm is used to process the error e(t), setting the proportional coefficient Kp = 1.0, the integral coefficient Ki = 0.1, and the differential coefficient Kd = 0.01. This generates a control variable u(t) = 0.12, and the robot arm is adjusted in the positive direction with an amplitude of 0.12 mm. The control variable u(t) is solved using inverse kinematics, and the Jacobian matrix is ​​used to calculate the target angles for each joint of the robot arm: θ1 = 0.1°, θ2 = 89.8°, and θ3 = 45.2°. The robot arm is driven according to the target angles of each joint, and the detection device's posture is adjusted. The detection position corresponding to the inner wall of the hollow shaft is (x = 100.12 mm, y = 49.88 mm, z = 200.12 mm). The new signal output by the adjusted Hall sensor is obtained, with a signal amplitude of 2.48V. The signal is processed by a 12-bit analog-to-digital converter to determine whether the fit between the detection device and the inner wall of the hollow shaft reaches the target distance dtarget = 2.0mm.

[0024] In step S107 , an intelligent eddy current detector is used to analyze and process the impedance change signal to extract surface defect information of the inner wall of the hollow shaft. The information is transmitted to a host computer to generate defect imaging results.

[0025] The intelligent eddy current detector collects impedance change signals transmitted by the probe through array channels and transmits them to the processing unit via the second and third terminals. The processing unit amplifies and filters the impedance change signals to remove background noise, generating clear defect-related signal data. Based on this defect-related signal data, a digital signal processing algorithm is used to extract characteristic parameters of the inner hole surface defect and determine preliminary defect location and morphology. By analyzing these characteristic parameters and combining them with linear Hall sensor data, the fit between the probe and the inner wall of the hollow shaft is determined, and a fit adjustment coefficient is obtained. Based on this fit adjustment coefficient, the controller adjusts the detection device's posture and generates signal acquisition instructions for real-time fit optimization. Based on this optimized signal acquisition instruction, the intelligent eddy current detector drives the probe along the inner hole axis, generating a full circumferential defect signal dataset. From this defect signal dataset, an array eddy current signal deconvolution algorithm is used to extract the depth and width of the inner shaft inner wall defect and determine the spatial distribution of the defect. Based on this spatial distribution, the intelligent eddy current detector transmits the defect information to a host computer, generating a raw data packet containing the defect location and size. The host computer receives the original data packet, processes the data using an image reconstruction algorithm, and generates imaging results of surface defects on the inner wall of the hollow shaft.

[0026] Specifically, the intelligent eddy current detector collects the impedance change signal transmitted by the probe through the array channel, and the signal is transmitted to the processing unit through the second terminal and the third terminal. The processing unit uses an amplifier with a gain of 60dB to amplify the signal, and removes background noise with a frequency higher than 10kHz through a Butterworth low-pass filter to obtain clear defect-related signal data. Based on the defect-related signal data, the fast Fourier transform algorithm is used to extract the characteristic parameters of the inner hole surface defect to determine the preliminary position and morphological information of the defect, for example, the defect length is about 2mm and the depth is 0.5mm. Through characteristic parameter analysis, combined with linear Hall sensor data, the fit state of the probe and the inner wall of the hollow shaft is judged, and the fit adjustment coefficient is obtained. If the Hall sensor detects that the distance between the probe and the inner wall is greater than 0.1mm, the adjustment coefficient is generated as 0.8. After obtaining the fit adjustment coefficient, the controller adjusts the posture of the detection device according to the coefficient and generates a real-time fit optimized signal acquisition instruction, for example, adjusting the probe movement speed to 5mm / s. Based on the optimized signal acquisition instructions, the intelligent eddy current detector drives the probe to move axially along the inner hole, obtaining a defect signal dataset with full circumferential coverage. The dataset contains 1,000 sampling points. Using the defect signal dataset, an array eddy current signal deconvolution algorithm is used to extract the depth and width of surface defects on the inner wall of the hollow shaft and determine the spatial distribution characteristics of the defects, such as a defect width of 1.5 mm and a depth of 0.6 mm. Based on the spatial distribution characteristics, the intelligent eddy current detector transmits the defect information to the host computer, generating a raw data packet containing the defect location and size. The data packet size is 1MB. The host computer receives the raw data packet and processes the data using an image reconstruction algorithm to generate an imaging result of the surface defect on the inner wall of the hollow shaft with an imaging resolution of 0.1 mm.

[0027] In step S108, the hollow shaft and the detection device are controlled by a servo motor to move relative to each other, and gradually advance along the axial direction to obtain defect data of different areas of the inner wall. The data is calibrated in real time by the linear Hall sensor to determine the fitting status.

[0028] The host computer controls the variable-diameter hollow shaft module's movement mechanism via a servo motor, gradually advancing it axially to a set initial position and acquiring relative position data between the long circular tube under test and the inner wall of the hollow shaft. The long circular tube under test is equipped with a linear Hall-type eddy current array sensor. Based on the initial position data, it collects contact status signals from the hollow shaft's inner wall and generates real-time calibration parameters. The array eddy current detector transmits the contact status signals via the second and third terminals. A signal processing algorithm is used to optimize the calibration parameters and determine the stability of the inner wall contact coupling. The servo motor adjusts the movement speed of the variable hollow shaft based on the optimized calibration parameters, advancing the shaft axially to the next test area and acquiring inner wall surface signals from that area. The linear Hall-type sensor uses the real-time calibration parameters to adjust the test device's posture and acquire adjusted contact status data to determine the tightness between the test fixture and the inner wall. The array eddy current detector, driven by a chain, extends its probe along a guide sleeve into the inner bore of the hollow shaft. Based on the adjusted contact status data, it acquires defect signals and generates raw data from different areas of the inner wall. A PC connected to the array eddy current detector performs time-domain filtering on the raw data. A multi-physics coupling algorithm is used to analyze the defect signal characteristics and determine the distribution of inner wall defects. The robotic arm control device adjusts the propulsion angle of the long circular tube based on this defect distribution information. A servo motor drives the probe forward steadily to acquire supplementary defect data for the next area. The array eddy current detector integrates this supplementary defect data through a signal processing module and generates complete inner wall defect data based on the calibration results of the linear Hall effect sensor, producing the final inspection report.

[0029] Step S109 , comparing the defect imaging result with a preset threshold value to determine the position and extent of the defect on the inner wall of the hollow shaft, and storing the determination result in a host computer database.

[0030] An array eddy current probe collects defect imaging data from the inner wall of a hollow shaft, generating an analog voltage signal indicating changes in the primary magnetic field. This analog voltage signal is filtered and subjected to noise reduction to obtain clear defect imaging data. Signal amplification and normalization techniques are used to process the defect imaging data, generating a standardized digital signal. This standardized digital signal is converted to a digital quantity through analog-to-digital conversion, resulting in a digitized defect imaging result. The digitized result is compared with a preset threshold to determine the initial location and extent of the defect. If the defect extent exceeds the preset threshold, host computer software analyzes the defect imaging data to determine the specific location coordinates of the defect. Based on the defect location coordinates and the distance data output by the Hall effect sensor, information is correlated between the defect depth and the inner wall fit. A PID algorithm is used to calculate the error between the defect imaging result and the target fit, generating adjusted detection device posture data. Based on this adjusted posture data, the defect location and extent information is transmitted to a host computer database, which stores the judgment result.

[0031] Step S1010: extracting detection data from the host computer database, and obtaining the circumferential defect distribution information of the inner wall of the hollow shaft through the S-shaped topological layout characteristics of the array eddy current sensor.

[0032] The test data is extracted from the host computer database and sorted according to the signal acquisition timestamp of the array eddy current sensor. Based on the sorted test data, the sampling points of the array eddy current sensor are spatially mapped using the S-shaped topological layout characteristics. The spatial mapping results are used to obtain the full circumferential coordinate system distribution information of the inner wall of the hollow shaft. Based on the full circumferential coordinate system distribution information, the correspondence between the voltage change signal collected by the array eddy current sensor and the defect location is determined. If the voltage change signal exceeds the preset threshold, the contact state between the probe and the inner wall is determined using the fitting data collected by the linear Hall sensor. Based on the contact state data, the attitude parameters of the detection device adjusted by the controller are obtained. Based on the test data after the attitude parameters are corrected, the defect signal distribution of the inner wall of the hollow shaft is obtained. Based on the defect signal distribution, a signal processing algorithm is used to determine the depth and width information of the full circumferential defect. Based on the depth and width information, a full circumferential defect distribution map of the inner wall of the hollow shaft is obtained.

[0033] Specifically, when extracting inspection data from the host computer database, a timestamp sorting algorithm is used to align the raw signals collected by the array eddy current sensors. For example, the signals are segmented into 1ms intervals to ensure temporal continuity. Based on the S-shaped topological layout, the coordinates of each eddy current probe's sampling points are mapped into a three-dimensional spatial model. A polar coordinate conversion algorithm is used, with the center of the hollow shaft as the origin. The angular and radial distances of each sampling point are calculated based on a 5mm probe spacing. Based on the spatial mapping results, a full circumferential coordinate system for the hollow shaft's inner wall is established. For example, a 360° circumference is divided into 72 equal regions, each corresponding to a 5° detection range. When evaluating voltage change signals, a defect detection threshold is set to ±10% of the standard impedance value. If the signal fluctuation exceeds this range, the linear Hall sensor's alignment detection module is triggered. The Hall sensor outputs the gap data between the probe and the inner wall with a resolution of 0.1mm. Based on this gap data, the controller uses a PID algorithm to calculate posture adjustment parameters, such as outputting stepper motor drive pulses to adjust the extension and retraction of the probe support arm to within a ±2mm range. The corrected inspection data is filtered through a Kalman filter to eliminate mechanical vibration noise and extract the defect signature. A wavelet transform algorithm is used to analyze the signal's frequency domain characteristics. When determining defect depth, the probability distribution of defect depth within the range of 0.1-2 mm is calculated based on the relationship between the eddy current attenuation coefficient and frequency. The defect coordinates, depth, and width data are input into a host computer, ultimately generating a full-circumferential defect distribution map.

[0034] The above only lists some preferred embodiments of the present invention, but the present invention is not limited thereto, and many improvements and modifications can be made. As long as the improvements and modifications are made on the basis of the basic principles of the present invention, they should be considered to fall within the scope of protection of the present invention.

Claims

1. A method for detecting defects on the inner wall of a hollow shaft, characterized in that: include: Obtaining the inspection requirements of the inner wall of the hollow shaft, the inspection requirements at least include the fitting state of the variable diameter inner wall; adjusting the posture of the inspection device by a robotic arm, and the robotic arm moves to the target position according to the control instruction; detecting defects on the inner wall of the hollow shaft by an electromagnetic array eddy current sensor, and the electromagnetic array eddy current sensor fits the inner wall and generates a detection signal; obtaining the fitting state data of the electromagnetic array eddy current sensor and the inner wall by a linear Hall sensor, and the fitting state data represents the distance between the inspection device and the inner wall; processing the detection signal by an intelligent eddy current detector and generating a defect detection result, and the defect detection result represents the position and characteristics of the inner wall defect.

2. The method for detecting inner wall defects of a hollow shaft according to claim 1, wherein: The method of detecting defects on the inner wall of a hollow shaft using an electromagnetic array eddy current sensor includes: adopting an electromagnetic array eddy current sensor with a double-row S-shaped topology layout, wherein the double-row S-shaped topology layout is composed of a first-row four-element detection array and a second-row three-element detection array; forming a spatial phase difference distribution by using the first-row four-element detection array and the second-row three-element detection array; constructing a spiral progressive detection path by embedding a short array in the gap of a long array; generating a consistent detection field strength through sensing units designed with equal parameters; and optimizing the magnetic circuit and generating the detection signal through magnetic domain orientation processing of the magnetic core.

3. The method for detecting inner wall defects of a hollow shaft according to claim 1, wherein: The method of obtaining the fitting status data of the electromagnetic array eddy current sensor and the inner wall through the linear Hall sensor includes: detecting the change of the primary magnetic field through the linear Hall sensor, and the primary magnetic field is excited by the electromagnetic array eddy current sensor; obtaining the size of the induced eddy current according to the change of the primary magnetic field; determining the fitting status of the electromagnetic array eddy current sensor and the inner wall through the relationship between the induced eddy current and the primary magnetic field; processing the change data of the primary magnetic field through filtering, noise reduction and signal amplification; generating an analog voltage signal related to the distance through analog-to-digital conversion and transmitting it to the host computer.

4. The method for detecting inner wall defects of a hollow shaft according to claim 1, wherein: The method of adjusting the posture of the detection device by a robotic arm includes: obtaining the fitting state data through a host computer and calculating the error between the actual distance and the target distance; generating a control amount through a PID algorithm according to the error; determining the direction and amplitude of movement of the robotic arm according to the control amount; calculating the target angle of each joint of the robotic arm through inverse kinematics; driving the robotic arm joint to the target angle through a motor; obtaining the actual angle through an encoder and comparing it with the target angle for fine-tuning.

5. The method for detecting inner wall defects of a hollow shaft according to claim 1, wherein: The method of processing the detection signal and generating a defect detection result by an intelligent eddy current detector includes: obtaining the detection signal by the intelligent eddy current detector; extracting the defect characteristics of the inner wall of the hollow shaft according to the detection signal; generating imaging data of the inner wall defect by signal processing; determining the position and size of the defect by the imaging data; and transmitting the imaging data to a host computer to generate a visual detection result.

6. The method for detecting defects on the inner wall of a hollow shaft according to claim 1, wherein: The method of adjusting the posture of the detection device by a robotic arm includes: fixing the hollow shaft by a dedicated clamping mechanism and driving it to rotate around the central axis; sending the detection device to a designated detection point on the inner wall of the hollow shaft by the robotic arm; continuously obtaining the distance data between the detection device and the inner wall by the linear Hall sensor; adjusting the motion trajectory of the robotic arm according to the distance data; and detecting the defect signal of the designated detection point by the electromagnetic array eddy current sensor.

7. The method for detecting inner wall defects of a hollow shaft according to claim 1, wherein: The method of detecting defects on the inner wall of a hollow shaft using an electromagnetic array eddy current sensor includes: exciting a primary magnetic field using the electromagnetic array eddy current sensor; inducing eddy currents on the inner wall of the hollow shaft according to the primary magnetic field; generating an impedance change by reacting the induced eddy currents to an excitation coil; determining a change in coil voltage based on the impedance change; and analyzing the change in coil voltage using the intelligent eddy current detector to generate defect information.

8. The method for detecting inner wall defects of a hollow shaft according to claim 2, wherein: The electromagnetic array eddy current sensor is a flat and flexible electromagnetic array eddy current sensor.

9. The method for detecting inner wall defects of a hollow shaft according to claim 8, wherein: The linear Hall sensor is a closed-loop linear Hall sensor.

Images