A small rotary ultrasonic testing device

Abstract

The present invention relates to the technical field of ultrasonic detection equipment, and more specifically, to a rotary ultrasonic detection equipment, including a main body shell, in which are arranged: a rotating mechanism unit, located at the front end of the main body shell; a control and processing module, located at the central position inside the main body shell; a display and operation interface, located at the upper part of the main body shell; a power supply and interface module, located at the rear part of the main body shell; wherein the rotating mechanism unit includes a high-precision servo motor, a precision reducer, a rotary encoder and a balancing structure, the high-precision servo motor is connected to the precision reducer, and the precision reducer is coaxially connected to the rotary encoder; it can not only perform 360-degree all-round detection, but also maintain extremely high accuracy during high-speed rotation. It can adapt to workpieces of various complex shapes, greatly improving the flexibility and adaptability of detection. At the same time, through advanced signal processing and data fusion algorithms, the present invention significantly improves the detection efficiency while improving the detection accuracy.

Description

Technical Field

[0001] The present invention relates to the technical field of ultrasonic detection equipment, and more specifically, to a small rotary ultrasonic detection equipment. Background Art

[0002] With the continuous development of industrial manufacturing, the requirements for product quality and safety are increasing day by day. Non-destructive testing technology is particularly important in this context. Ultrasonic testing, as a commonly used non-destructive testing method, has been widely used in various industries due to its advantages such as no radiation, portability and relatively low cost. However, traditional ultrasonic testing technology often seems to be unable to cope with complex-shaped workpieces or scenes that require all-round testing.

[0003] Early ultrasonic testing equipment was mostly fixed, and the operator needed to manually move the probe to complete the test. This method is not only inefficient, but also has poor consistency and repeatability in the test results. With the advancement of technology, some automated testing equipment has begun to appear, but there are still many limitations. For example, the rotary ultrasonic machining system proposed by Zhou et al. in their research (see Zhou, H., Zhang, J., Yu, D., et al. Advances in rotary ultrasonic machining system for hard and brittle materials. Advances in Mechanical Engineering, 2019, 11 (12): 1-13) has made some progress in processing, but it still has shortcomings in all-round detection capabilities and adaptability to complex-shaped workpieces.

[0004] The ultrasonic testing equipment currently on the market mainly has the following problems:

[0005] 1. Limited detection range: Most devices can only detect specified locations on the product and cannot achieve all-round detection. This is particularly disadvantageous when detecting complex-shaped workpieces, as defects in key areas may be missed.

[0006] 2. Lack of flexibility: It is often difficult for inspection equipment to flexibly adjust the inspection position according to different measurement requirements. This results in the need to replace equipment or re-order when facing workpieces of different types or shapes, resulting in a waste of manpower and material resources.

[0007] 3. The contradiction between accuracy and efficiency: Improving detection accuracy usually means reducing detection speed, while improving efficiency may sacrifice accuracy. Existing technologies find it difficult to achieve a good balance between the two.

[0008] 4. Poor adaptability: For workpieces with curved surfaces or irregular shapes, existing equipment often has difficulty maintaining stable detection accuracy. This is a serious problem in high-end manufacturing industries such as aerospace and energy.

[0009] 5. Limited data processing capabilities: Many devices lack advanced signal processing and data fusion capabilities and are unable to fully utilize the advantages of multi-angle detection, resulting in insufficient reliability and comprehensiveness of detection results.

[0010] 6. Poor stability during long-term operation: In industrial environments that require continuous long-term operation, the performance of existing equipment tends to decline significantly over time, affecting the reliability of detection.

[0011] These problems have seriously restricted the application and development of ultrasonic testing technology in high-end manufacturing. Therefore, the development of a new type of ultrasonic testing equipment that can achieve all-round, high-precision and high-efficiency testing has become an urgent need in the industry. Summary of the invention

[0012] The present invention is proposed in this context, aiming to solve the limitations of the prior art and provide a small rotary ultrasonic testing device to achieve all-round detection of workpieces and improve the comprehensiveness and reliability of detection. Improve the flexibility of detection so that the device can adapt to workpieces of various shapes and types. Improve detection efficiency while ensuring high precision. Improve the adaptability of the device to complex curved surfaces and irregularly shaped workpieces. Improve the accuracy and reliability of the detection results through advanced signal processing and data fusion technology. Improve the long-term working stability of the equipment to meet the needs of industrial environments.

[0013] The present invention provides a small rotary ultrasonic detection device, comprising: a main body shell, wherein the main body shell is provided with: a rotary mechanism unit, which is located at the front end of the main body shell; a control and processing module, which is located at the central position inside the main body shell; a display and operation interface, which is located at the upper part of the main body shell; and a power supply and interface module, which is located at the rear part of the main body shell; wherein the rotary mechanism unit comprises a high-precision servo motor, a precision reducer, a rotary encoder and a balancing structure, wherein the high-precision servo motor is connected to the precision reducer, the precision reducer is coaxially connected to the rotary encoder, and the balancing structure is arranged on both sides of the precision reducer; the control and processing module comprises a signal processing and data acquisition unit and an intelligent control unit, wherein the signal processing and data acquisition unit is connected to the intelligent control unit through a data bus; the rotary mechanism unit is connected to a spherical array probe, and is used to realize 360-degree rotation of the spherical array probe; the signal processing and data acquisition unit is connected to the spherical array probe through a high-speed data line, and is used to receive and process ultrasonic signals; the intelligent control unit is respectively connected to the rotary mechanism unit, the signal processing and data acquisition unit, the display and operation interface and the power supply and interface module, and is used to control the operation of the entire device.

[0014] Specifically, the spherical array probe includes: a plurality of piezoelectric crystal elements arranged in a spherical shape, each of the piezoelectric crystal elements is connected to the signal processing and data acquisition unit through an independent signal line; an adaptive focusing module is connected to the intelligent control unit and is used to adjust the focusing parameters in real time according to the current detection angle and depth; a coupling agent automatic replenishment system includes a micro peristaltic pump and a storage tank, the micro peristaltic pump is electrically connected to the intelligent control unit and is used to automatically replenish the coupling agent according to the detection requirements; wherein, when it is detected that the coupling agent reserve is lower than a preset threshold value, the intelligent control unit controls the micro peristaltic pump to start replenishing the coupling agent; if the coupling agent reserve is still lower than the preset threshold value after replenishment, a warning message is issued to the operator.

[0015] Specifically, the signal processing and data acquisition unit includes: a high-speed digital signal processor, used to process the ultrasonic signal from the spherical array probe in real time; an analog-to-digital converter, connected to the spherical array probe and the high-speed digital signal processor, used to convert the analog signal into a digital signal; a data storage module, connected to the high-speed digital signal processor, used to store the collected and processed data; wherein, when the remaining storage space of the data storage module is lower than a predetermined threshold, the intelligent control unit automatically triggers data compression or data transmission operations to ensure continuous data acquisition.

[0016] Specifically, the intelligent control unit includes: a microprocessor for executing control programs and data processing algorithms; a real-time operating system installed on the microprocessor for managing device resources and task scheduling; an artificial intelligence accelerator connected to the microprocessor through a high-speed interface for accelerating the operation of AI algorithms; wherein the intelligent control unit dynamically adjusts the operating frequency of the microprocessor and the operating mode of the artificial intelligence accelerator according to the current task complexity to balance performance and power consumption.

[0017] Specifically, it also includes an adaptive focusing algorithm, which is implemented based on the following steps: obtaining the current rotation angle and the target focusing depth; calculating the three-dimensional coordinates of the desired focusing point based on the rotation angle and the target focusing depth; calculating the distance from the center of the sphere to the desired focusing point based on the desired focusing point coordinates; for each piezoelectric crystal element in the spherical array probe, calculating the sound wave propagation time difference according to its position and the distance; based on the time difference, setting the transmission and reception delay time for each piezoelectric crystal element; applying the delay time to each piezoelectric crystal element to achieve focusing of the sound wave; wherein, if the calculated desired focusing point exceeds the maximum detection range of the device, the focusing point is automatically adjusted to the boundary of the maximum detection range, and a prompt message is sent to the operator.

[0018] Specifically, it also includes a defect recognition AI algorithm, which is implemented based on the following steps: obtaining a preprocessed B-scan image; extracting multi-scale features of the B-scan image through a multi-layer convolutional neural network; using a jump connection structure to fuse feature maps of different levels; based on the fused feature map, generating a feature map of the same size as the input image through a deconvolution operation; applying a Sigmoid activation function to the feature map to obtain a segmentation map representing the probability of the defect; determining the position and size of the defect from the segmentation map by setting a threshold; wherein, when the size or position of the detected defect exceeds a preset range, the system automatically increases the scanning density of the area and re-executes the detection process to improve the reliability of the result.

[0019] Specifically, it also includes a multi-angle data fusion algorithm, which is implemented based on the following steps: initialize the weight of each detection angle to an equal value; calculate the uncertainty of the result of each detection angle, and the uncertainty is based on the signal strength and noise level; update the weight of each detection angle according to the uncertainty, the lower the uncertainty, the higher the weight; use the updated weight to perform weighted averaging on the detection results of each angle to obtain a fusion result; based on the uncertainty of each angle, calculate the comprehensive uncertainty of the fusion result; wherein, if the uncertainty of a certain angle is significantly higher than that of other angles and exceeds a preset threshold, the data of this angle will be excluded from the fusion process and the re-detection process of this angle will be triggered.

[0020] Specifically, the rated power of the high-precision servo motor is 500W, and the maximum speed is 6000rpm; the reduction ratio of the precision reducer is 160:1, and the repeatability accuracy is ±5 arc seconds; the resolution of the rotary encoder is 18 bits, that is, 262,144 resolutions / rev; wherein the intelligent control unit dynamically adjusts the speed of the high-precision servo motor according to the detection requirements, and automatically reduces the speed to ensure stability when the detection accuracy requirements increase; when the error between the actual position detected by the rotary encoder and the target position exceeds a preset threshold, an automatic calibration procedure is triggered.

[0021] Specifically, the spherical array probe contains 64 piezoelectric crystal elements, each of which has a size of 5mm×5mm and an operating frequency of 5MHz; the adaptive focusing module is implemented using a field programmable gate array, and the clock frequency is not less than 100MHz; wherein the intelligent control unit monitors the working status of each piezoelectric crystal element in real time, and when it is detected that the performance degradation of any element exceeds a preset threshold, it automatically adjusts the parameters of other elements to compensate, and marks the status of the element in the test report; if the operating frequency deviates from 5MHz by more than ±0.5MHz due to environmental factors, the automatic frequency calibration procedure is triggered.

[0022] Specifically, it also includes the following work processes: system startup and self-test, including: power-on self-test, checking the connection status and functional integrity of each hardware module; sensor calibration, using the built-in standard test block to calibrate the sensitivity of the spherical array probe; rotation mechanism initialization, returning the rotation mechanism to zero and verifying the encoder reading; if any step fails, the system will report a specific error and enter safe mode; parameter setting, including: loading default parameters, reading the last used parameters from the configuration file; user adjustment, allowing the operator to adjust the detection range and accuracy requirement parameters through the touch screen interface; parameter verification, the system automatically checks the validity of the parameters to prevent setting invalid parameters; if the parameters set by the user exceed the device capability range, the system automatically adjusts to the closest feasible parameters and prompts the user; the detection process, including: rotation positioning, controlling the servo motor to rotate to the specified position according to the set angle interval; coupling agent replenishment, automatically replenishing the appropriate amount of coupling agent after each rotation; ultrasonic transmission and reception, triggering the spherical array probe, and collecting A-scan signals; real-time processing, the high-speed digital signal processor performs pre-processing such as signal filtering and envelope detection; defect recognition, the processed B-scan image is input to the data storage, saving the original data and processing results to the data storage module; wherein, if an abnormality is found during the detection process, the system automatically repeats the detection of the current position, and suspends the detection and alarms if three consecutive abnormalities occur; multi-angle fusion, including: data alignment, aligning the detection data of different angles based on the readings of the rotary encoder; uncertainty estimation, calculating the uncertainty of the detection results of each angle; weight calculation, dynamically adjusting the weight of the data of each angle according to the uncertainty; data fusion, applying the multi-angle data fusion algorithm to generate a comprehensive result; wherein, if the uncertainty of the fused result exceeds the preset threshold, the system automatically increases the number of detection angles and re-executes the detection process; result display and interaction, including: 3D reconstruction, building a three-dimensional model of the detected object based on the fusion result; defect annotation, annotating the position and size of the detected defect on the three-dimensional model; interactive operation, allowing the user to rotate and zoom the three-dimensional model through the touch screen to view detailed information; report generation, automatically generating a PDF format report containing detection parameters, results and three-dimensional models; wherein, if the detected defect exceeds the preset severity, the system automatically sends an alarm message to the designated person.

[0023] The beneficial effects of the present invention are mainly reflected in the following aspects:

[0024] The above goals have been successfully achieved through a series of innovations such as innovative spherical array probe design, high-precision rotation mechanism, adaptive focusing algorithm and multi-angle data fusion technology. This rotary ultrasonic detection equipment can not only perform 360-degree all-round detection, but also maintain extremely high accuracy during high-speed rotation. It can adapt to workpieces of various complex shapes, greatly improving the flexibility and adaptability of detection. At the same time, through advanced signal processing and data fusion algorithms, the present invention significantly improves detection efficiency while improving detection accuracy.

[0025] These characteristics of the present invention make it have broad application prospects in high-end manufacturing industries such as aerospace, energy, and automobiles. It can not only improve product quality and safety, but also reduce production costs and improve production efficiency. In general, the present invention has brought revolutionary progress to the field of industrial non-destructive testing and is expected to become one of the key technologies to promote the development of Industry 4.0 and intelligent manufacturing. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 It is the overall structure logic block diagram of the present invention.

[0027] Figure 2 It is a detailed block diagram of the rotating mechanism unit of the present invention.

[0028] Figure 3 It is a detailed block diagram of the spherical array probe of the present invention.

[0029] Figure 4 It is a detailed block diagram of the control and processing module of the present invention.

[0030] Figure 5 It is a logic block diagram of the overall workflow of the present invention.

[0031] Figure 6 This is a logic block diagram of the adaptive focusing algorithm of the present invention.

[0032] Figure 7 It is the defect identification AI algorithm logic box of the present invention.

[0033] Figure 8 This is a logic block diagram of the multi-angle data fusion algorithm of the present invention.

[0034] Fig. 9 This is the detailed process of system startup and self-checking of the present invention.

[0035] Fig.10 The following is a detailed flow chart of the detection process of the present invention.

[0036] Fig.11 It is a dynamic detection efficiency waveform diagram of the present invention and comparative example 1-2.

[0037] Fig.12It is a comparison chart of the 3D surface detection capabilities of the present invention and comparative examples 1-2. DETAILED DESCRIPTION

[0038] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the specific implementation methods, structures, features and effects thereof are described in detail below in conjunction with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0039] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0040] The present invention will be further described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0041] Please refer to Figure 1-12 The present invention relates to a rotary ultrasonic testing device, which is particularly suitable for all-round non-destructive testing of complex-shaped workpieces. In industrial production, integrity testing of key components such as pipelines and pressure vessels is of great importance. Traditional fixed ultrasonic testing equipment often has difficulty in dealing with complex curved surfaces or irregularly shaped workpieces, and the rotary design of the present invention just solves this problem.

[0042] Specifically, the rotary ultrasonic detection device of the present invention is mainly composed of the following parts: a main body shell 1, a rotary mechanism unit 2, a control and processing module 3, a display and operation interface 4, and a power supply and interface module 5. These components work together to achieve efficient and accurate all-round detection.

[0043] First, let's take a look at the core of the equipment - the rotating mechanism unit 2. This unit is located at the front end of the main housing 1 and consists of a high-precision servo motor 21, a precision reducer 22, a rotary encoder 23 and a balancing structure 24. The high-precision servo motor 21 (model: AKM52G-ANCNR-00) provides strong and precise power, with a rated power of 500W and a maximum speed of up to 6000rpm. This high-performance motor ensures smooth movement of the probe during the inspection process. The maximum speed of the high-precision servo motor 21 is set to 6000rpm, but in the actual inspection process, we usually limit it to less than 4000rpm. This is because although a higher speed can improve the inspection efficiency, it may also introduce more vibration and instability. 4000rpm is the best balance between efficiency and stability, which can meet most inspection needs.

[0044] Connected to the motor is a precision reducer 22 (model: SHG-20-160-2UH), with a reduction ratio of 160:1 and an amazing repeatability of ±5 arc seconds. This high-precision reducer is essential to ensure the accuracy of the test. Especially when testing narrow spaces such as the inner wall of a pipe, a small positioning error can lead to significant deviations in the test results.

[0045] The rotary encoder 23 (model: RON786C) provides 18-bit high resolution, that is, 262,144 resolution points per revolution. This accuracy enables the system to accurately track the position of the probe, providing reliable spatial information for subsequent data processing and image reconstruction.

[0046] It is worth mentioning that this equipment also adopts an innovative balanced structure design. Adjustable tungsten alloy counterweights (density 18g / cm 3 ). This design cleverly solves the vibration problem during high-speed rotation and ensures the stability of detection. According to our experiments, when the rotation speed exceeds 3000rpm, if there is no balance structure, the vibration amplitude of the probe may exceed 50μm, which will seriously affect the detection accuracy. With the balance structure, even at a high speed of 6000rpm, the vibration of the probe can be controlled within 10μm.

[0047] Next, let's focus on the "brain" of the device - the control and processing module 3. This module is located in the central position inside the main body and includes a signal processing and data acquisition unit 31 and an intelligent control unit 32. The signal processing and data acquisition unit 31 uses a high-speed digital signal processor 311 (DSP), a high-performance analog-to-digital converter 312 (ADC) and a large-capacity data storage module 313.

[0048] Specifically, the DSP uses the TMS320C6678 multi-core processor from Texas Instruments, with a main frequency of up to 1.25GHz and 8 processing cores. This powerful processing capability enables the system to process massive ultrasonic signals from the spherical array probe in real time. The ADC uses AD's AD9276, which has a 16-bit resolution and a sampling rate of 80MSPS. This high-performance ADC ensures high-fidelity acquisition of the original signal.

[0049] For data storage, we chose a 512GB NVMe SSD (model: Samsung 970PRO). This high-speed storage device not only provides ample storage space, but its ultra-high read and write speeds (read speed up to 3500MB / s, write speed up to 2700MB / s) also ensures fast access to data, effectively avoiding bottlenecks in the data processing process. When the remaining storage space of the data storage module 313 is less than 15%, the system automatically triggers data compression or transfer operations. This threshold is determined based on the data volume and storage requirements of typical detection tasks. 15% of the remaining space is usually sufficient to cope with sudden large amounts of data generation, while also leaving enough time for the system to perform necessary data management operations.

[0050] The intelligent control unit 32 is the command center of the entire system, which includes a high-performance microprocessor 321, a real-time operating system 322 and an AI accelerator 323. The microprocessor 321 uses the Intel Core i7-1185G7, 4 cores and 8 threads, and a maximum frequency of 4.8GHz. This powerful processor provides sufficient computing resources for complex control algorithms and data analysis. The operating frequency of the microprocessor is dynamically adjusted from 1.5GHz to 4.8GHz. When the task load is light (such as the system is idle or only simple data processing is performed), the processor frequency will drop to 1.5GHz to save energy; when facing complex real-time data processing tasks, the frequency can be increased to 4.8GHz to ensure performance. This dynamic adjustment strategy can effectively reduce the system's power consumption and heat generation while ensuring processing performance.

[0051] The real-time operating system 322 uses QNX Neutrino RTOS 7.1. This professional real-time operating system has extremely low latency and high reliability, which is very suitable for application scenarios such as this device that have extremely high real-time requirements. Especially when processing large amounts of data generated by high-speed rotating probes, QNX's efficient scheduling algorithm can ensure the timely execution of key tasks.

[0052] AI Accelerator 323 uses Google Edge TPU (model: Coral USB Accelerator), which can provide 4TOPS of computing power. This dedicated AI chip greatly accelerates the image recognition and defect analysis process of the device. According to our tests, after using Edge TPU, the execution speed of complex defect recognition algorithms increased by about 5 times, making real-time analysis possible.

[0053] Another key component of the present invention is the spherical array probe 6. This innovative probe design includes 64 piezoelectric crystal elements 61 arranged in a spherical shape. Each piezoelectric crystal element has a size of 5mm×5mm and an operating frequency of 5MHz. This design ensures good detection performance at any angle. The 64 piezoelectric crystal elements of the spherical array probe 6 have an operating frequency of 5MHz, but the system supports fine-tuning within the range of 4-6MHz. This fine-tuning capability enables the device to adapt to different detection objects and environments. For example, for denser materials, we may slightly reduce the frequency to improve penetration; for detection tasks that require higher resolution, the frequency can be slightly increased.

[0054] The probe is also equipped with an adaptive focusing module 62, which uses an FPGA (model: Xilinx Virtex-7XC7VX485T) to implement a real-time focusing algorithm. The FPGA has a clock frequency of up to 100MHz and can quickly adjust the phase of each piezoelectric element to achieve precise beam focusing. This adaptive focusing technology greatly improves the resolution and sensitivity of the detection, especially when detecting complex curved surfaces.

[0055] In order to ensure the acoustic coupling effect during the detection process, this device is also equipped with an innovative automatic coupling agent replenishment system63. The system includes a micro peristaltic pump (flow range: 0.1-100ml / min) and a storage tank with a capacity of 5ml. The coupling agent is evenly distributed on the surface of the probe through micropores (diameter 0.2mm). The system monitors the use of the coupling agent in real time, and automatically starts the replenishment procedure when the reserve is lower than the preset threshold (usually set to 20%). The 20% threshold can not only ensure an adequate supply of coupling agent during the detection process, but also avoid too frequent replenishment operations, thereby improving the overall detection efficiency. The amount of coupling agent replenished each time is controlled between 0.1-0.2ml. This range can minimize waste while ensuring a good acoustic coupling effect.

[0056] Now, let's take a closer look at the working process of this device.

[0057] Step S1: System startup and self-test. First, the device performs a power-on self-test to check the connection status and functional integrity of each hardware module. This process usually takes 30-60 seconds, depending on the complexity of the system. Next, the spherical array probe 6 is calibrated for sensitivity using a built-in standard test block. This calibration process usually takes 2-3 minutes and ensures that the probe has consistent detection sensitivity at different angles and depths. Finally, the system initializes the rotation mechanism, zeros it, and verifies the encoder reading. If any abnormality is found during the self-test, the system immediately enters safe mode and reports the error message in detail on the display to guide the operator to troubleshoot.

[0058] Step S2: Parameter setting. At this stage, the system first loads the last used parameters from the configuration file. These parameters include detection range, rotation speed, data acquisition frequency, etc. The operator can then adjust these parameters through the 10.1-inch high-resolution touch screen (1920×1200 pixels). For example, for pipeline inner wall detection, we usually recommend setting the rotation speed between 1000-2000rpm and the data acquisition frequency to 50kHz. These settings can achieve a good balance between detection effect and efficiency in most cases.

[0059] It is worth noting that the system automatically checks the validity of the parameters. For example, if the detection depth set by the operator exceeds the maximum detection range of the equipment (usually 300mm), the system will automatically adjust it to the maximum feasible value and prompt the operator through a pop-up window. This intelligent adjustment mechanism effectively prevents detection errors caused by improper parameter settings.

[0060] Step S3: Detection process. This is the core part of the entire workflow. First, according to the set angle interval, the high-precision servo motor 21 controls the spherical array probe 6 to rotate to the specified position. Our experience shows that for most applications, collecting data every 15° is a good balance point, which can ensure the comprehensiveness of the detection without generating too much redundant data.

[0061] After each rotation, the coupling agent automatic replenishment system 63 will accurately replenish the appropriate amount of coupling agent. Through precise control, we ensure that the amount of coupling agent replenished each time is between 0.1-0.2ml, which can ensure good acoustic coupling effect without causing excessive waste.

[0062] Next, the system triggers the spherical array probe 6 to emit ultrasonic waves and receive echo signals. In this process, the adaptive focusing algorithm plays a key role. The algorithm is based on the principle of time reversal mirror and phased array technology, and its core steps are as follows:

[0063] 1. Get the current rotation angle θ and target focus depth d;

[0064] 2. Calculate the desired focus point coordinates: x f =d·sin(θ),y f =0,z f =d·cos(θ);

[0065] 3. Calculate the distance from the center of the sphere to the desired focus point:

[0066] 4. Calculate the delay time for each array element i:

[0067] Where θ is the current rotation angle (unit: radians), d is the target focus depth (unit: meters) x f ,y f ,z f is the three-dimensional coordinate of the desired focal point (unit: meter), R is the distance from the center of the sphere to the desired focal point (unit: meter), x i ,y i ,z i is the three-dimensional coordinate of the ith array element (unit: meter), τ i is the delay time of the ith array element (unit: seconds), and c is the propagation speed of ultrasound in the medium (unit: meter / second).

[0068] When the calculated desired focus point exceeds the maximum detection range of the device (usually 300mm), the system automatically adjusts the focus point to 290mm. 290mm is chosen instead of 300mm to ensure the stability of the focus effect, because the focus quality may be reduced at the extremes of the detection range.

[0069] This adaptive focusing technology enables our device to maintain excellent spatial resolution at different depths and angles. According to our tests, at a depth of 200mm, this technology can achieve a lateral resolution of less than 1mm.

[0070] Subsequently, the high-speed DSP processes the received A-scan signal in real time, including signal filtering, envelope detection, etc. We use a bandpass filter (passband is 3.5MHz-6.5MHz) to remove noise, and use Hilbert transform for envelope detection. These processing steps significantly improve the signal-to-noise ratio of the signal, laying the foundation for subsequent defect identification.

[0071] Next, the processed B-scan image is fed into the AI ​​model for defect recognition analysis. Our AI algorithm is based on a convolutional neural network with a U-Net architecture, and its main steps include:

[0072] 1) Convert the raw A-scan data into a 256×256 pixel B-scan image.

[0073] 2) Multi-scale features are extracted through 5 convolutional blocks, each of which contains two 3×3 convolutional layers and one 2×2 maximum pooling layer.

[0074] 3) Use 5 upsampling blocks for feature reconstruction, each block contains a 2×2 upsampling layer and two 3×3 convolutional layers.

[0075] 4) Use skip connections to concatenate the feature map of the corresponding layer of the encoder with the feature map of the decoder.

[0076] 5) Finally, a 1×1 convolutional layer and a Sigmoid activation function are used to generate a segmentation map representing the defect probability.

[0077] This deep learning model performed well in our tests, with a detection rate of over 95% and a false alarm rate of less than 5% for common defect types (such as cracks, pores, inclusions, etc.). During defect identification, if the detected defect size exceeds 10mm or the depth exceeds 5mm, the system automatically increases the scanning density of the area. Specifically, the angular resolution of the area is increased from the standard 15° to 7.5° to obtain more detailed data. These thresholds are determined based on years of inspection experience and a large amount of experimental data, and can effectively balance inspection accuracy and efficiency.

[0078] Finally, the system saves the original data and processing results to a high-speed SSD. Considering the importance of the data, we adopt a real-time backup strategy to ensure the security of the data.

[0079] It is worth noting that during the entire detection process, the system will continue to monitor the signal quality. If an abnormality is found (for example, the signal strength suddenly decreases by more than 50%), the system will automatically repeat the detection of the current position. If there is still an abnormality after three consecutive detections, the system will pause the detection and issue an alarm to prompt the operator to check the equipment status or the surface condition of the workpiece.

[0080] Step S4: Multi-angle data fusion. This is another innovation of the present invention. First, the system aligns the detection data of different angles based on the accurate reading of the rotary encoder 23. Then, the uncertainty of the detection result of each angle is calculated. Here we use the uncertainty estimation method based on signal strength and noise level:

[0081] σ i = k·(N i / S i )

[0082] Among them, σ i is the uncertainty of the ith angle, N i is the noise level at the ith angle, S i is the signal strength at the ith angle, and k is an empirical coefficient (usually between 0.1 and 0.3).

[0083] Next, the system dynamically adjusts the weight of each angle data according to the uncertainty:

[0084]

[0085] Finally, the weighted average method is used for data fusion:

[0086] F=∑(w i ·Di ) / ∑w i

[0087] Among them, w i is the weight of the i-th angle, D i is the detection result of the i-th angle, F is the final result after fusion, and ∑ represents the sum of all angles.

[0088] This multi-angle fusion technology significantly improves the reliability of detection. According to our experiments, compared with single-angle detection, this method can improve the defect location accuracy by about 30% and the detection rate by about 15%.

[0089] If the uncertainty of the fused result exceeds a preset threshold (usually set to 20%), the system will automatically increase the number of detection angles and re-execute the detection process to ensure the reliability of the result.

[0090] In the data fusion process, if the uncertainty σ of a certain angle i If the signal-to-noise ratio is greater than 0.5 (i.e., the signal-to-noise ratio is less than 2:1), the data at that angle will be excluded from the fusion process. This threshold is set to ensure the reliability of the fusion result. Our experiments show that when the signal-to-noise ratio is less than 2:1, the credibility of the data is significantly reduced, and including it in the fusion process may reduce the accuracy of the overall result.

[0091] Step S5: Result display and interaction: This is the last step of the entire detection process and also the key stage for the operator to interact with the system.

[0092] First, the system builds a 3D model of the detected object based on the fused detection results. Here we use the improved Marching Cubes Algorithm to achieve efficient 3D reconstruction. The core idea of ​​the algorithm is to regard the detection data as a 3D voxel field, and then determine the position and shape of the isosurface by analyzing the eight vertex values ​​of each voxel. We optimized the traditional algorithm and introduced adaptive meshing technology, which increased the reconstruction speed by about 40% while ensuring the accuracy of the model.

[0093] The steps of the improved March cube algorithm are as follows:

[0094] 1. Initialization: Define the voxel field V(x,y,z), where x,y,z are spatial coordinates, and set the threshold T of the target isosurface.

[0095] 2. Adaptive grid division: Dynamically adjust the grid size according to the gradient information of the detection data. Grid size adjustment formula:

[0096]

[0097] Use a smaller mesh in areas with larger gradients and a larger mesh in areas with smaller gradients. Where Δs is the adjusted mesh size and Δs base is the base grid size, α is the adjustment factor (0<α<1), is the gradient magnitude of the voxel field.

[0098] Traverse the voxels: a. Determine the vertex state: For each voxel, perform the following steps: For the 8 vertices v of the voxel i (i=0,1,…,7), calculate: s i =1ifV(v i )≥T,else s i =0;

[0099] b. Look up the table to determine the triangle configuration: According to the vertex status, look up the predefined 256 configuration tables to determine the intersection of the isosurface and the voxel.

[0100] c. Calculate the intersection position: For each voxel edge that intersects the isosurface, use linear interpolation to calculate the intersection position P, P = v1 + (TV (v1)) · (v2-v1) / (V (v2)-V (v1)), v1, v2 are the two endpoints of the voxel edge, V (v1), V (v2) are the voxel field values ​​corresponding to the endpoints;

[0101] d. Generate triangles: Generate triangles representing isosurfaces based on the intersection positions.

[0102] 4. Mesh optimization: Optimize the generated triangular mesh, including: remove duplicate vertices; smoothing: use Laplace smoothing algorithm, P ′ i =P i +λ·∑(P j -P i ) / N, where P ′ i is the vertex position after smoothing, P i is the original vertex position, P j is the position of adjacent vertices, λ is the smoothing coefficient (0<λ<1), N is the number of adjacent vertices, and the optimized triangular mesh is output to represent the reconstructed 3D model.

[0103] Next, the system marks the location and size of the detected defects on the 3D model. In order to intuitively indicate the severity of the defects, we use a color coding scheme: green for minor defects (less than 1mm in depth), yellow for medium defects (1-3mm in depth), and red for severe defects (greater than 3mm in depth). This intuitive visualization method greatly improves the readability of the inspection results.

[0104] The operator can then interact with the 3D model through the touch screen, including rotating, zooming, and viewing detailed information. We have developed a smooth gesture control interface, where the operator can pinch to zoom to adjust the model size and swipe with one finger to rotate the model. When a defect mark is clicked, the system will pop up a detailed information window showing the specific parameters of the defect, such as location coordinates, size, depth, etc.

[0105] Finally, the system automatically generates a PDF report containing the inspection parameters, results and 3D model.

[0106] The report includes the following key information:

[0107] 1. Inspection overview: date, time, operator information, inspected workpiece information, etc.

[0108] 2. Detection parameters: type of probe used, detection frequency, rotation speed, data acquisition frequency, etc.

[0109] 3. Summary of test results: number of defects detected, distribution, maximum defect size, etc.

[0110] 4. 3D model screenshot: Display the 3D model with defect annotations from multiple angles.

[0111] 5. Detailed defect list: including the location, size, depth, type and other information of each defect.

[0112] 6. Conclusions and recommendations: Preliminary assessment and follow-up recommendations based on the test results.

[0113] It is worth mentioning that if the detected defects exceed the preset severity, the system will automatically send an alarm message to the designated personnel. For example, when a crack with a depth of more than 5mm is detected, the system will immediately notify the quality inspection supervisor through the preset communication channel (such as email, SMS, etc.) so that necessary measures can be taken in time.

[0114] Preferably, in one embodiment of the present invention, we also integrate an intelligent temperature management system. Considering that high-speed rotation and continuous data processing may cause the system temperature to rise, we designed a multi-level temperature control mechanism:

[0115] 1. When the temperature of any key component (such as DSP, motor, etc.) exceeds 60℃, the system will automatically increase the speed of the cooling fan.

[0116] 2. If the temperature continues to rise to 70°C, the system will appropriately reduce the processor frequency and motor speed to reduce heat generation.

[0117] 3. If the temperature still rises to 80°C, the system will pause the detection process and issue an overheating warning.

[0118] This intelligent temperature control mechanism ensures the stability and reliability of the device during long-term continuous operation. According to our long-term tests, at room temperature of 25°C, this device can work continuously for more than 8 hours without overheating.

[0119] During the system startup and self-check phase, if any step fails and cannot be automatically recovered within 60 seconds, the system will enter safe mode. The 60-second time threshold is determined based on a lot of practical experience. It provides enough time for the system to recover itself without excessively delaying the operator's work.

[0120] During the detection process, if the signal strength of three consecutive tests decreases by more than 50%, the system will suspend the detection and alarm. The setting of this threshold and number of times is intended to balance sensitivity and false alarm rate. A single or occasional drop in signal strength may be caused by temporary factors, while three consecutive large drops are likely to indicate equipment failure or significant changes in the detection environment.

[0121] Finally, in the result display and interaction stage, we adopted the following color coding scheme for the detected defects: green indicates minor defects with a depth of less than 1mm, yellow indicates medium defects with a depth of 1-3mm, and red indicates severe defects with a depth of more than 3mm. This classification scheme is based on multiple industry standards and practical application experience, and can intuitively reflect the severity of defects and help operators quickly identify areas that need to be focused on.

[0122] In addition, considering the various complex situations that may be encountered in actual applications, we have also developed a set of adaptive detection strategies. For example, when severe rust or coating is detected on the surface of the workpiece, the system will automatically adjust the ultrasonic frequency and energy to improve the penetration ability. Specifically, for ordinary metal surfaces, we use a standard frequency of 5MHz; for severely corroded surfaces, the system will automatically switch to a low-frequency mode of 2.25MHz and appropriately increase the transmission energy (usually by 20-30%). This adaptive strategy significantly improves the adaptability of the equipment under various complex working conditions.

[0123] In general, the rotary ultrasonic testing equipment of the present invention realizes efficient, accurate and all-round testing of workpieces with complex shapes through the innovative design of spherical array probe 6, high-precision rotating mechanism, advanced signal processing and artificial intelligence algorithms, and humanized interactive interface. It not only greatly improves the testing efficiency and accuracy, but also significantly reduces the difficulty of operation, bringing revolutionary progress to the field of industrial non-destructive testing.

[0124] In practical applications, this equipment has demonstrated excellent performance in many fields. For example, during the annual inspection of the pressure vessel of a nuclear power plant, our equipment successfully detected tiny cracks (only 0.8mm in length) that were difficult to detect with traditional methods, providing key information for timely prevention of potential failures. In another case, an airline used our equipment to conduct routine inspections of passenger aircraft wings, which not only shortened the inspection time by 40%, but also increased the detection rate by about 20%, greatly improving flight safety.

[0125] Looking ahead, we plan to further optimize the algorithm and improve the intelligence level of the equipment. For example, we will introduce machine learning technology to achieve automatic defect classification and damage assessment. At the same time, we are also exploring the possibility of combining this technology with other non-destructive testing methods (such as eddy current testing, X-ray testing, etc.) to develop a more comprehensive and powerful integrated testing system.

[0126] In summary, the present invention provides an efficient, accurate and easy-to-use solution for non-destructive testing of complex workpieces, which is expected to play an important role in many fields such as industrial quality control and safety inspection, and make important contributions to improving product quality and production safety.

[0127] In order to fully verify the performance of the rotary ultrasonic testing equipment of the present invention, we designed a series of rigorous tests covering the key indicators of the equipment. These tests not only verified the basic functions of the equipment, but more importantly, they highlighted the innovation and superiority of the present invention.

[0128] First, we conducted an in-depth evaluation of the equipment's detection accuracy. We produced a series of standard test blocks containing artificial defects of different sizes, depths, and shapes. These test blocks simulated various defect conditions that might be encountered in actual industrial environments. Our test team repeatedly tested these test blocks using this equipment, with each test block tested at least 50 times to ensure statistical significance of the results.

[0129] Secondly, we focused on testing the rotational stability and positioning accuracy of the device. This is an important innovation of the present invention and is directly related to the reliability of the test results. We designed a high-precision laser displacement sensor system to monitor the tiny displacement of the probe in real time during the rotation process.

[0130] In addition, we also evaluated the adaptive focusing capability of the device. We used curved test blocks with different curvatures to simulate workpieces of various complex shapes. This test was designed to verify the superiority of the invention when processing irregular surfaces.

[0131] Finally, we conducted a long-term continuous operation test to verify the stability and reliability of the equipment. This test lasted for 72 hours and simulated the high-intensity usage scenario in an actual industrial environment.

[0132] Here is a summary of our test results:

[0133]

[0134] These test results clearly demonstrate the superior performance of our invention. First, in terms of defect detection accuracy, our equipment can reliably detect tiny defects of 0.2mm, which is more than double the industry average. This high accuracy stems from our innovative spherical array probe design and advanced signal processing algorithms. It means that our equipment can detect potentially dangerous defects at an early stage, greatly improving the safety of industrial production.

[0135] The test results of rotational stability are also impressive. Even at a high speed of 6000rpm, the displacement of the probe is controlled within 10μm. This is far better than the industry average, fully demonstrating the effect of our high-precision servo motor and innovative balanced structure design. High rotational stability ensures that extremely high accuracy can be maintained even in high-speed detection, which is crucial to improving detection efficiency.

[0136] The test results of the adaptive focusing effect further highlight the unique advantages of the present invention. Our equipment can still maintain a detection accuracy of 0.3mm on surfaces with a curvature radius greater than 50mm, which greatly exceeds the existing technology. This means that our equipment can accurately detect workpieces of various complex shapes, such as aircraft engine blades, pressure vessels, etc., greatly expanding the scope of application.

[0137] The results of long-term stability tests are also noteworthy. After 72 hours of continuous operation, the performance of our equipment has decayed by less than 5%, while the industry average is 15%. This excellent stability is due to our carefully designed cooling system and intelligent power management strategy. It means that our equipment can operate reliably for a long time in a high-intensity industrial environment, reducing downtime for maintenance and improving production efficiency.

[0138] Significant improvement in detection speed (200cm 2 / min vs the industry average of 50cm 2 / min) fully demonstrates the advantages of our all-round detection. This high efficiency comes not only from high-speed rotation, but more importantly from our innovative multi-angle data fusion algorithm, which enables us to obtain comprehensive detection data in one scan.

[0139] In general, these test results strongly demonstrate the significant advantages of the present invention in terms of detection accuracy, efficiency and adaptability. It not only greatly improves the reliability and efficiency of detection, but also expands the application scope of ultrasonic detection technology. This performance leap will bring revolutionary changes to industrial quality control and is expected to play an important role in high-end manufacturing industries such as aerospace, energy, and automobiles.

[0140] In order to explore the advantages of the present invention, we conducted a series of comprehensive experiments to compare the performance of the present invention with the prior art under various conditions. These experiments not only verified the effectiveness of our innovative design, but also pointed out the direction for future improvements.

[0141] We selected two typical comparative examples: traditional fixed ultrasonic testing equipment and common rotating ultrasonic testing equipment. These comparative examples represent the mainstream technologies and the latest attempts in the current market. Our embodiments fully reflect the core innovations of the present invention, including spherical array probes, high-precision rotating mechanisms, and advanced signal processing algorithms.

[0142] In the experiment, we focused on several key indicators: detection accuracy, rotation stability, adaptive focusing capability, detection speed, and long-term working stability. These indicators directly reflect the performance improvement brought by our innovations. We used standard test blocks, curved workpieces, and long-term operation tests simulating industrial environments to comprehensively evaluate the performance of each device.

[0143] The following is a summary of our experimental results:

[0144]

[0145] From these data, we can clearly see that the present invention performs well in all aspects, especially in processing complex-shaped workpieces and high-speed detection.

[0146] Please refer to Fig.11 This dynamic waveform diagram can clearly demonstrate the advantages of the present invention in maintaining high efficiency and high stability during long-term continuous operation. Observers can intuitively see that the waveform line of the present invention always remains at a high position with little fluctuation, while the waveform line of the comparative example is relatively low and may have obvious fluctuations. Please refer to Fig.12 Through this 3D diagram, we can intuitively see that the "detection accuracy surface" of the present invention maintains the smallest distance and the best consistency with the workpiece surface, while the comparative example will show an obvious decrease in accuracy in the complex surface area, which is reflected as a deviation from the workpiece surface.

[0147] In terms of the minimum detectable defect, the 0.2mm of the present invention is far superior to other solutions. This is due to our innovative spherical array probe design and advanced signal processing algorithms. In practical applications, this means that we can detect potentially dangerous defects at an early stage, greatly improving the safety and reliability of products. For example, in the aerospace field, this high-precision detection capability can help detect tiny cracks or material fatigue early and prevent failures that may lead to catastrophic consequences.

[0148] The test results of rotational stability are even more impressive. At a high speed of 6000rpm, our equipment can still control the probe displacement within 10μm, which is far superior to other rotary devices. High rotational stability ensures accuracy in high-speed detection and greatly improves detection efficiency. On the production line of the automotive manufacturing industry, this high-speed and high-precision detection capability can significantly improve the efficiency of product quality control without sacrificing detection accuracy.

[0149] The breakthrough in curved surface inspection capabilities highlights the unique advantages of this invention. The ability to accurately inspect surfaces with a radius of curvature of only 50mm far exceeds existing technologies. This means that our equipment can easily handle workpieces of various complex shapes, such as turbine blades, pressure vessels, etc. In the energy industry, this capability allows us to more comprehensively inspect key components in nuclear power plants and improve overall safety.

[0150] Significant improvement in detection speed (200cm 2 / min) fully demonstrates the advantages of our all-round detection. This high efficiency comes not only from high-speed rotation, but more importantly from our innovative multi-angle data fusion algorithm. In large-scale manufacturing industries, this efficient detection can significantly shorten production cycles and increase production capacity.

[0151] The results of long-term stability testing are equally impressive. After 72 hours of continuous operation, the performance of our equipment has declined by less than 5%, which is far better than other solutions. This excellent stability is due to our carefully designed heat dissipation system and intelligent power management strategy. In industrial environments that require 24-hour uninterrupted operation, such as the petrochemical industry, this stable performance ensures the continued reliability of detection and reduces downtime for maintenance.

[0152] In general, these experimental results strongly demonstrate the significant advantages of the present invention in terms of detection accuracy, efficiency and adaptability. It not only greatly improves the reliability and efficiency of detection, but also expands the application scope of ultrasonic detection technology. This performance leap will bring revolutionary changes to industrial quality control and is expected to play an important role in high-end manufacturing industries such as aerospace, energy, and automobiles.

[0153] The best implementation is undoubtedly the one that uses our full set of innovative technologies. It perfectly combines spherical array probes, high-precision rotation mechanisms, adaptive focusing algorithms, and multi-angle data fusion technology. This combination not only achieves the best in all indicators, but also demonstrates excellent comprehensive performance. Especially when processing complex-shaped workpieces, the advantages of this solution are more obvious.

[0154] In the future, we plan to further optimize the algorithm, especially in real-time data processing and automatic defect classification. We believe that with the development of artificial intelligence technology, our equipment will be able to identify and classify defects more intelligently, further improving the efficiency and accuracy of inspection. At the same time, we are also exploring the possibility of combining this technology with other non-destructive testing methods to develop a more comprehensive and powerful integrated inspection system.

[0155] In conclusion, these experimental results not only verify the superiority of the present invention, but also point out the future development direction for us. We believe that this technology will play an important role in improving the quality and safety of industrial production and bring practical benefits to various industries.

[0156] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modification, replacement, and improvement made within the principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A rotary ultrasonic testing device, characterized in that: include: A main body shell, wherein the main body shell is provided with: a rotating mechanism unit, located at the front end of the main body shell; The control and processing module is located in the central position of the main shell; the display and operation interface is located in the upper part of the main shell; The power supply and interface module is located at the rear of the main body shell; wherein the rotating mechanism unit includes a high-precision servo motor, a precision reducer, a rotary encoder and a balancing structure, the high-precision servo motor is connected to the precision reducer, the precision reducer is coaxially connected to the rotary encoder, and the balancing structure is arranged on both sides of the precision reducer; the control and processing module includes a signal processing and data acquisition unit and an intelligent control unit, and the signal processing and data acquisition unit is connected to the intelligent control unit through a data bus; the rotating mechanism unit is connected to the spherical array probe to achieve 360-degree rotation of the spherical array probe; the signal processing and data acquisition unit is connected to the spherical array probe through a high-speed data line to receive and process ultrasonic signals; the intelligent control unit is respectively connected to the rotating mechanism unit, the signal processing and data acquisition unit, the display and operation interface and the power supply and interface module to control the operation of the entire device; The spherical array probe includes: a plurality of piezoelectric crystal elements arranged in a spherical shape, each of the piezoelectric crystal elements is connected to the signal processing and data acquisition unit through an independent signal line; an adaptive focusing module is connected to the intelligent control unit and is used to adjust the focusing parameters in real time according to the current detection angle and depth; a coupling agent automatic replenishment system includes a micro peristaltic pump and a storage tank, the micro peristaltic pump is electrically connected to the intelligent control unit and is used to automatically replenish the coupling agent according to the detection requirements; wherein, when it is detected that the coupling agent reserve is lower than a preset threshold value, the intelligent control unit controls the micro peristaltic pump to start replenishing the coupling agent; if the coupling agent reserve is still lower than the preset threshold value after replenishment, a warning message is issued to the operator.

2. The rotary ultrasonic testing device according to claim 1, characterized in that: The signal processing and data acquisition unit includes: a high-speed digital signal processor, which is used to process the ultrasonic signal from the spherical array probe in real time; an analog-to-digital converter, which is connected to the spherical array probe and the high-speed digital signal processor, and is used to convert the analog signal into a digital signal; a data storage module, which is connected to the high-speed digital signal processor, and is used to store the collected and processed data; wherein, when the remaining storage space of the data storage module is lower than a predetermined threshold, the intelligent control unit automatically triggers data compression or data transmission operations to ensure continuous data acquisition.

3. The rotary ultrasonic testing device according to claim 1, characterized in that: The intelligent control unit includes: a microprocessor for executing control programs and data processing algorithms; a real-time operating system installed on the microprocessor for managing device resources and task scheduling; an artificial intelligence accelerator connected to the microprocessor via a high-speed interface for accelerating the operation of the AI ​​algorithm; wherein the intelligent control unit dynamically adjusts the operating frequency of the microprocessor and the operating mode of the artificial intelligence accelerator according to the current task complexity to balance performance and power consumption.

4. The rotary ultrasonic testing device according to claim 1, characterized in that: It also includes an adaptive focusing algorithm, which is implemented based on the following steps: obtaining a current rotation angle and a target focusing depth; calculating the three-dimensional coordinates of a desired focusing point according to the rotation angle and the target focusing depth; Based on the desired focus point coordinates, calculating the distance from the sphere center to the desired focus point; For each piezoelectric crystal element in the spherical array probe, the sound wave propagation time difference is calculated according to its position and the distance; based on the time difference, the transmission and reception delay time is set for each piezoelectric crystal element; the delay time is applied to each piezoelectric crystal element to achieve the focusing of the sound wave; wherein, if the calculated expected focusing point exceeds the maximum detection range of the device, the focusing point is automatically adjusted to the boundary of the maximum detection range, and a prompt message is sent to the operator.

5. The rotary ultrasonic testing device according to claim 1, characterized in that: Also included is a defect recognition AI algorithm, which is implemented based on the following steps: obtaining a pre-processed B-scan image; Extracting multi-scale features of the B-scan image through a multi-layer convolutional neural network; Using the skip connection structure, feature maps of different levels are fused; based on the fused feature maps, a deconvolution operation is performed to generate a feature map of the same size as the input image; A Sigmoid activation function is applied to the feature map to obtain a segmentation map representing the defect probability; the position and size of the defect are determined from the segmentation map by setting a threshold; wherein, when the size or position of the detected defect exceeds a preset range, the system automatically increases the scanning density of the area where the defect is located and re-executes the detection process to improve the reliability of the result.

6. The rotary ultrasonic testing device according to claim 1, characterized in that: It also includes a multi-angle data fusion algorithm, which is implemented based on the following steps: initializing the weight of each detection angle to an equal value; for the result of each detection angle, calculating its uncertainty, the uncertainty is based on the signal strength and the noise level; according to the uncertainty, updating the weight of each detection angle, the lower the uncertainty, the higher the weight; The updated weights are used to perform weighted averaging on the detection results of each angle to obtain the fusion result. Based on the uncertainty of each angle, the comprehensive uncertainty of the fusion result is calculated. If the uncertainty of a certain angle is significantly higher than that of other angles and exceeds the preset threshold, the data of this angle will be excluded from the fusion process and the re-detection process of this angle will be triggered.

7. The rotary ultrasonic testing device according to claim 1, characterized in that: The rated power of the high-precision servo motor is 500W, and the maximum speed is 6000rpm; the reduction ratio of the precision reducer is 160:1, and the repeatability accuracy is ±5 arc seconds; the resolution of the rotary encoder is 18 bits, i.e. 262,144 resolutions / rev; wherein the intelligent control unit dynamically adjusts the speed of the high-precision servo motor according to the detection requirements, and automatically reduces the speed to ensure stability when the detection accuracy requirements increase; when the error between the actual position detected by the rotary encoder and the target position exceeds a preset threshold, an automatic calibration procedure is triggered.

8. The rotary ultrasonic testing device according to claim 1, characterized in that: The spherical array probe includes 64 piezoelectric crystal elements, each of which has a size of 5mm×5mm and an operating frequency of 5MHz; the adaptive focusing module is implemented using a field programmable gate array, and the clock frequency is not less than 100MHz; wherein the intelligent control unit monitors the working status of each piezoelectric crystal element in real time, and when it is detected that the performance degradation of any element exceeds a preset threshold, the parameters of other elements are automatically adjusted to compensate, and the status of the element is marked in the test report; if the operating frequency deviates from 5MHz by more than ±0.5MHz due to environmental factors, the automatic frequency calibration procedure is triggered.

9. The rotary ultrasonic testing device according to claim 1, characterized in that: It also includes the following workflows: system startup and self-test, including: power-on self-test, checking the connection status and functional integrity of each hardware module; sensor calibration, using the built-in standard test block to calibrate the sensitivity of the spherical array probe; rotation mechanism initialization, returning the rotation mechanism to zero and verifying the encoder reading; if any step fails, the system will report a specific error and enter safe mode; parameter setting, including: loading default parameters, reading the last used parameters from the configuration file; user adjustment, allowing the operator to adjust the detection range and accuracy requirement parameters through the touch screen interface; parameter verification, the system automatically checks the validity of the parameters to prevent invalid parameters from being set; if the parameters set by the user exceed the equipment capability range, the system automatically adjusts to the closest feasible parameters and prompts the user; the detection process, including: rotation positioning, controlling the servo motor to rotate to the specified position according to the set angle interval; coupling agent replenishment, automatically replenishing the appropriate amount of coupling agent after each rotation; ultrasonic transmission and reception, triggering the spherical array probe, and collecting A-scan signals; real-time processing, the high-speed digital signal processor performs signal filtering and envelope detection preprocessing; defect recognition, inputting the processed B-scan image into the A I model for analysis; data storage, saving the original data and processing results to the data storage module; wherein, if an abnormality is found during the detection process, the system automatically repeats the detection of the current position, and suspends the detection and alarms if three consecutive abnormalities occur; multi-angle fusion, including: data alignment, aligning the detection data of different angles based on the readings of the rotary encoder; uncertainty estimation, calculating the uncertainty of the detection results of each angle; weight calculation, dynamically adjusting the weight of the data of each angle according to the uncertainty; data fusion, applying the multi-angle data fusion algorithm to generate a comprehensive result; wherein, if the uncertainty of the fused result exceeds the preset threshold, the system automatically increases the number of detection angles and re-executes the detection process; result display and interaction, including: 3D reconstruction, building a three-dimensional model of the detected object based on the fusion result; defect annotation, annotating the position and size of the detected defect on the three-dimensional model; interactive operation, allowing the user to rotate and zoom the three-dimensional model through the touch screen to view detailed information; report generation, automatically generating a PDF format report containing detection parameters, results and three-dimensional models; wherein, if the detected defects exceed the preset severity, the system automatically sends an alarm message to the designated person.

Images