Tank defect self-adaptive imaging method based on ultrasonic flaw detection technology

Abstract

The invention discloses a tank defect self-adaptive imaging method based on an ultrasonic flaw detection technology. The method comprises the following steps: S1, setting a driving path of an ultrasonic flaw detection trolley according to the detection requirement of a detected part; s2, the ultrasonic flaw detection trolley runs along a preset running path, an ultrasonic probe starts to work at the same time, ultrasonic waves are sent to the detected part, and a signal reflected back from the interior of the detected part is received and recorded and is transmitted back to the trolley data processing center; s3, the data processing center processes the ultrasonic signals by adopting a Golden Litsea rotundifolia algorithm based on a composite index CI to optimize a variational mode decomposition algorithm, then establishes a target function # imgabs2 # based on an algorithm approximation index value # imgabs0 # and a correlation coefficient index value # imgabs1 #, and performs denoising reconstruction on the signals by utilizing a self-adaptive reconstruction algorithm. According to the method, through signal processing and feature extraction, the gray level image of the internal defect condition of the anti-tank body is generated. Through analysis of the grayscale image, automatic identification and positioning of internal defects of the tank body can be realized.

Description

Technical Field

[0001] The invention relates to the technical field of ultrasonic flaw detection imaging, and in particular to a tank body defect adaptive imaging method based on ultrasonic flaw detection technology. Background Art

[0002] Ultrasonic flaw detection technology is a non-destructive testing method that detects defects inside materials by sending high-frequency sound waves and receiving reflected waves. When ultrasonic waves encounter defects inside the material, reflections occur, and the reflected waves are received by the detector and converted into electrical signals. By analyzing these signals, the location, size and shape of the defects can be inferred. In the prior art, defect detection of container shells mainly focuses on the detection of surface fatigue cracks. For example, the patent with application number 202111231964.3 discloses a nondestructive detection method and system for fatigue cracks on the surface of a pressure vessel shell. The detection steps are: cleaning the surface of the shell, then applying a penetrant on the surface, and then placing the shell in a high-pressure chamber for constant pressure periodic vibration. The purpose is to allow the penetrant to penetrate into the shell, so as to facilitate the subsequent use of a developer to develop the shell, and further facilitate the high-definition camera to scan the developed image to obtain a defect image, and further use ultrasound to perform ultrasonic detection on the shell after cleaning, so as to obtain characteristic data of the shell defects; the technical solution disclosed in the patent application does not specify whether the high-definition camera can shoot the inner wall of the container shell in the later stage. At the same time, in the technical solution, the developed image is first scanned by a high-definition camera to determine whether the container is damaged and then determine whether to perform the next step of ultrasonic detection. This step is likely to cause omissions of non-surface defects of the shell, that is, internal defects.

[0003] On the other hand, in the prior art, the ultrasonic signal collected by the ultrasonic probe usually needs to be converted into a grayscale image so that the defective part and the normal part can be contrasted in grayscale, which is helpful for discovering and identifying the defect.

[0004] In existing image processing algorithms, histogram equalization adjusts the grayscale histogram of the image to make it uniformly distributed, thereby enhancing the contrast. However, it may lead to noise enhancement, affecting image quality, and has limited effect on images with good contrast. In filtering technology, median filtering is computationally complex and has poor effect on large-scale noise. Gaussian filtering blurs image details and affects edge detection. In this application, power transformation, bilinear interpolation and mean filtering techniques are used to process the grayscale image of defects. Summary of the invention

[0005] In view of this, the purpose of the present invention is to develop an adaptive imaging method for tank defects based on ultrasonic flaw detection technology, to visualize ultrasonic scanning defects, to convert the scanning results into grayscale images, to adapt the scanning forward speed according to the defect depth of the scanning surface, to improve the detection efficiency while ensuring the accuracy of the detection results.

[0006] The tank defect adaptive imaging method based on ultrasonic flaw detection technology of the present invention comprises the following steps:

[0007] S1. Set the driving path of the ultrasonic flaw detection trolley according to the inspection requirements of the inspected parts.

[0008] S2. The ultrasonic flaw detection vehicle travels along the preset driving path, and the ultrasonic probe starts working, emitting ultrasonic waves to the inspected object. The signal reflected back from the inspected object is received and recorded, and then transmitted back to the vehicle data processing center.

[0009] S3. The data processing center uses the Golden Jackal algorithm based on the composite index CI to optimize the variational mode decomposition algorithm to process the ultrasonic signal, and then based on the algorithm approximation index value And the correlation coefficient index value Establishing the objective function , the signal is denoised and reconstructed using an adaptive reconstruction algorithm, and the defect grayscale image is obtained under full traversal path planning, and the defect grayscale image is transmitted back to the display end using the internal communication device of the car.

[0010] In a possible implementation, in step s2, the ultrasonic flaw detection vehicle automatically identifies the defect density of the inspected parts in the driving area during driving. If the defect density is higher than a preset value, the driving speed of the vehicle is slowed down to increase the scanning accuracy in the defect-dense areas; if the defect density is lower than the preset value, the vehicle increases the driving speed while ensuring the minimum scanning accuracy to increase the scanning efficiency.

[0011] In a possible implementation, in step s3, the specific steps of processing the ultrasonic signal by using the algorithm of optimizing the variational mode decomposition based on the composite index CI using the golden jackal algorithm include:

[0012] a. Input ultrasonic signal, initialize GJO parameters, set \left [ {K,\alpha} \right ] The search range of the game randomly generates golden jackal pairs and prey locations.

[0013] b. VMD was used to decompose the input signal under different positions of golden jackal pairs, and the value of the composite index CI of the fitness function was calculated. The optimal solution was the position of the male jackal, and the suboptimal solution was the position of the female jackal.

[0014] Calculate Levy motion random numbers and prey escape energy ,like , then Formula I, Formula II, and Formula III are used to update the prey position, and the golden jackal performs a global search; otherwise, Formula III, Formula IV, and Formula V are used to update the prey position, and the golden jackal performs a local search and attack; the formula I is:

[0015] ; The formula II is: ; The formula III is: .

[0016] The formula IV is: ; The formula V is: .

[0017] c. Update the position of the golden jackal pair according to the location of the prey.

[0018] d. Iterate and return to step b until the optimal \left [ {K,\alpha} \right ] is output after the maximum number of iterations and the minimum composite index.

[0019] e. The best \left [ {K,\alpha} \right ] obtained in step d Results The signal is decomposed by VMD to obtain K IMF components.

[0020] In a possible implementation, in step s3, the specific steps of obtaining the defect grayscale image include:

[0021] (1) Ultrasonic signal After VMD decomposition, the reconstructed signal is expressed as:

[0022] .

[0023] In the formula, is the intrinsic mode function; is the number of natural mode components; For the remainder.

[0024] (2) The ultrasonic signal is expressed as: , define any algorithm The root mean square error of the denoising deviation is:

[0025] .

[0026] In the formula, For the The ultrasonic signal value at the time, ; is the total number of sampled signals; For the algorithm right Noise reduction results, ; Used to evaluate the error between the reconstructed signal and the original signal.

[0027] definition For the algorithm Correlation coefficient:

[0028] .

[0029] In the formula, For the The ultrasonic signal value at the time, ; For the algorithm right Noise reduction results, ; It is used to evaluate the similarity between the reconstructed signal and the original signal; it is the inner product operation.

[0030] (3) According to the approximation index value And the correlation coefficient index value , the following good noise reduction criteria are established:

[0031] Constraints:

[0032] .

[0033] Denoising and reconstruction algorithm The objective function is:

[0034] .

[0035] (4) Iteratively run the algorithm Ra to obtain the denoised and reconstructed ultrasonic signal and extract the measured thickness value.

[0036] (5) The ultrasonic flaw detection vehicle scans the entire path to obtain the ultrasonic grayscale image of the defect.

[0037] Beneficial effects of the present invention: The present invention visualizes the scanned defects, converts the scan results into grayscale images, and adaptively controls the forward scanning speed of the ultrasonic flaw detection trolley according to the defect density of the scanned surface. Under the premise of ensuring the minimum scanning accuracy, if the surface defects are sparse, the scanning speed of the trolley is increased to increase the scanning efficiency; if the surface defects are dense, the scanning speed of the trolley is reduced to facilitate further accurate drawing of the grayscale image and increase the scanning accuracy. Compared with other solutions, the present invention is more efficient, and the grayscale image is more intuitive to display defects. Compared with existing conventional algorithms, the algorithm of the present invention has the following advantages:

[0038] 1. Dynamically adjust the scanning speed: Existing conventional algorithms usually use a fixed scanning speed and cannot be adjusted according to the defect density, which may lead to low efficiency or insufficient scanning accuracy. The present invention dynamically adjusts the scanning speed to ensure that the scanning efficiency is improved when the defects are sparse and the scanning accuracy is guaranteed when the defects are dense.

[0039] 2. Higher detection accuracy: Through adaptive control, the scanning speed can be reduced in defect-intensive areas, a more detailed grayscale image can be obtained, and higher detection accuracy can be provided. This method effectively avoids the defect omission or misdetection problem caused by fixed speed in existing algorithms.

[0040] 3. Intelligent control: The algorithm of the present invention utilizes the scan data for real-time analysis and adjustment, which reflects the advantages of intelligent control and is more efficient and accurate than existing manual or semi-automatic adjustment methods.

[0041] 4. Optimized utilization of resources: By optimizing the scanning speed, the present invention reduces unnecessary scanning time and resource consumption, improves the resource utilization of the overall detection process, and is more economical and efficient than existing algorithms.

[0042] In summary, the present invention not only has significant advantages in defect visualization processing and grayscale image generation, but also significantly improves scanning efficiency and detection accuracy by adaptively controlling the scanning speed, showing significant advantages over existing conventional algorithms. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0044] Figure 1 It is a schematic diagram of the ultrasonic flaw detection process involved in the present invention.

[0045] Figure 2 The present invention is a flowchart of the golden jackal algorithm based on the composite index CI to optimize the variational mode decomposition algorithm.

[0046] Figure 3 This is a flow chart of the adaptive reconstruction algorithm adopted by the present invention.

[0047] Figure 4 Schematic diagram of an ultrasonic signal after noise reduction and reconstruction obtained by using algorithm Ra in the present invention.

[0048] Figure 5 This is a schematic diagram of the circuitous full traversal path scanning principle adopted by the present invention.

[0049] Figure 6 This is an ultrasonic grayscale image of a circular tiny defect obtained by the method of the present invention.

[0050] Figure 7 This is an ultrasonic grayscale image of an elliptical small defect obtained by the method of the present invention.

[0051] Figure 8 This is an ultrasonic grayscale image of a square medium-sized defect obtained by the method of the present invention.

[0052] Fig. 9This is an ultrasonic grayscale image of a large circular defect obtained by the method of the present invention. DETAILED DESCRIPTION

[0053] like Figure 1 As shown, a tank defect adaptive imaging method based on ultrasonic flaw detection technology in this embodiment includes the following steps:

[0054] S1. Set the driving path of the ultrasonic flaw detection trolley according to the inspection requirements of the inspected parts.

[0055] The minimum detection accuracy of the flaw detection trolley should be reasonably set according to the area, shape and detection accuracy requirements of the inspected part, and the operation path of the trolley should be reasonably planned. There should be a certain overlapping area between adjacent scanning paths to avoid undetected gaps between two scans.

[0056] S2. The ultrasonic flaw detection vehicle travels along the preset driving path, and the ultrasonic probe starts working, sending ultrasonic waves to the inspected parts. The reflected signals from the inspected parts are received and recorded, and then transmitted back to the vehicle data processing center. During this process, the ultrasonic flaw detection vehicle will automatically identify the density of defects in the inspected parts in the driving area. If the density of defects is higher than the preset value, the vehicle will slow down to increase the scanning accuracy of the defect-dense areas; if the density of defects is lower than the preset value, the vehicle will increase the driving speed to increase the scanning efficiency while ensuring the minimum scanning accuracy.

[0057] S3. The data processing center uses the Golden Jackal algorithm based on the composite index CI to optimize the variational mode decomposition algorithm to process the ultrasonic signal, and then based on the algorithm approximation index value And the correlation coefficient index value Establishing the objective function , the signal is denoised and reconstructed using an adaptive reconstruction algorithm, and the defect grayscale image is obtained under full traversal path planning, and the defect grayscale image is transmitted back to the display end using the internal communication device of the car.

[0058] like Figure 2 As shown, in step s3, the specific steps of processing the ultrasonic signal by using the algorithm of optimizing the variational mode decomposition based on the golden jackal algorithm of the composite index CI include:

[0059] a. Input ultrasonic signal, initialize the parameters of Golden Jackal algorithm GJO, set \left [ {K,\alpha} \right ] The search range is set to [2, 12] for the number of modal components K and 1 for the optimization step; the optimization range of the penalty coefficient α is , the optimal step length is 50; the maximum number of iterations is 20; the population size is 30, and the golden jackal pairs and prey locations are randomly generated.

[0060] b. VMD was used to decompose the input signal under different positions of golden jackal pairs, and the value of the composite index CI of the fitness function was calculated. The optimal solution was the position of the male jackal, and the suboptimal solution was the position of the female jackal.

[0061] Calculate Levy motion random numbers and prey escape energy ,like , then use Formula I, Formula II, and Formula III to update the prey position, and the golden jackal conducts a global search; otherwise, use Formula III, Formula IV, and Formula V to update the prey position, and the golden jackal conducts a local search and attack.

[0062] The formula I is:

[0063] .

[0064] The formula II is:

[0065] .

[0066] The formula III is:

[0067] .

[0068] The formula IV is:

[0069] .

[0070] The formula V is:

[0071] .

[0072] c. Update the position of the golden jackal pair according to the location of the prey.

[0073] d. Iterate and return to step b until the optimal \left [ {K,\alpha} \right ] is output after the maximum number of iterations and the minimum composite index.

[0074] e. The best \left [ {K,\alpha} \right ] obtained in step d Results The signal is decomposed by VMD to obtain K IMF components.

[0075] like Figure 3 As shown, in a possible implementation, in step s3, the specific steps of obtaining the defect grayscale image include:

[0076] (1) Ultrasonic signal After VMD decomposition, the reconstructed signal is expressed as:

[0077] .

[0078] In the formula, is the intrinsic mode function; is the number of natural mode components; For the remainder.

[0079] (2) The ultrasonic signal is expressed as: , define any algorithm The root mean square error of the denoising deviation is:

[0080] .

[0081] In the formula, For the The ultrasonic signal value at the time, ; is the total number of sampled signals; For the algorithm right Noise reduction results, ; Used to evaluate the error between the reconstructed signal and the original signal.

[0082] definition For the algorithm Correlation coefficient:

[0083] .

[0084] In the formula, For the The ultrasonic signal value at the time, ; For the algorithm right Noise reduction results, ; <·> represents inner product operation; Used to evaluate the similarity between the reconstructed signal and the original signal.

[0085] (3) According to the above approximation index And the correlation coefficient index , the following good noise reduction criteria are established:

[0086] Constraints:

[0087] .

[0088] Denoising and reconstruction algorithm The objective function is:

[0089] .

[0090] (4) Iterative operation, the noise-reduced and reconstructed ultrasonic signal is obtained through the algorithm Ra, and the measured thickness value is extracted, such as Figure 4 The ultrasonic signal obtained after noise reduction and reconstruction is shown.

[0091] (5) The ultrasonic grayscale image of the defect is obtained by scanning the entire path of the ultrasonic flaw detection vehicle; Figure 5 This is a schematic diagram of the circuitous full traversal path scanning principle used in this embodiment. Figure 6-Figure 9 This is the defect grayscale image obtained in this embodiment.

[0092] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention may have various changes and modifications. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention are included in the protection scope of the present invention.

Claims

1. A tank defect adaptive imaging method based on ultrasonic flaw detection technology, characterized in that: The following steps are involved: S1. Set the driving path of the ultrasonic flaw detection vehicle according to the inspection requirements of the inspected parts; S2. The ultrasonic flaw detection vehicle moves along the preset driving path, and the ultrasonic probe starts working, emitting ultrasonic waves to the inspected object. The reflected signals from the inspected object are received and recorded, and then transmitted back to the vehicle data processing center; S3. The data processing center uses the Golden Jackal algorithm based on the composite index CI to optimize the variational mode decomposition algorithm to process the ultrasonic signal, and then based on the algorithm approximation index value And the correlation coefficient index value Establishing the objective function , the signal is denoised and reconstructed using an adaptive reconstruction algorithm, and the defect grayscale image is obtained under full traversal path planning, and the defect grayscale image is transmitted back to the display end using the internal communication device of the car.

2. The method for adaptive imaging of tank defects based on ultrasonic flaw detection technology according to claim 1 is characterized in that: In step s2, the ultrasonic flaw detection vehicle automatically identifies the defect density of the inspected parts in the driving area during driving. If the defect density is higher than a preset value, the driving speed of the vehicle is slowed down to increase the scanning accuracy of the defect-dense areas; if the defect density is lower than the preset value, the vehicle increases the driving speed while ensuring the minimum scanning accuracy to increase the scanning efficiency.

3. The method for adaptive imaging of tank defects based on ultrasonic flaw detection technology according to claim 2 is characterized in that: In step s3, the specific steps of processing the ultrasonic signal by using the algorithm of optimizing the variational mode decomposition based on the composite index CI using the golden jackal algorithm include: a. Input ultrasonic signal, initialize GJO parameters, set The search range randomly generates golden jackal pairs and prey locations; b. Using VMD to decompose the input signal at different positions of the golden jackal pair, the value of the composite index CI of the fitness function is calculated. The optimal solution is the position of the male jackal, and the suboptimal solution is the position of the female jackal; Calculate Levy motion random numbers and prey escape energy ,like , then Formula I, Formula II, and Formula III are used to update the prey position, and the golden jackal performs a global search; otherwise, Formula III, Formula IV, and Formula V are used to update the prey position, and the golden jackal performs a local search and attack; The formula I is: ； The formula II is: ； The formula III is: ； The formula IV is: ； The formula V is: ； In the formula, and Respectively The updated positions of the male and female golden jackals corresponding to the prey after the iteration; and Indicates The positions of male and female golden jackals after iterations; It represents the energy of the hunted prey; It is The position vector of the prey after iterations, Indicates the number of iterations; Represents a random number from the Levy distribution in the Levy motion; , Respectively represent the distance between the golden jackal and its prey; c. Update the position of the golden jackal pair according to the location of the prey; d. Iterate and return to step b until the optimal value is output after the maximum number of iterations. and minimum composite index; e. The best result obtained in step d Results The signal is decomposed by VMD to obtain K IMF components.

4. The method for adaptive imaging of tank defects based on ultrasonic flaw detection technology according to claim 3 is characterized in that: In step s3, the specific steps of obtaining the defect grayscale image include: (1) Ultrasonic signal After VMD decomposition, the reconstructed signal is expressed as: ； In the formula, is the intrinsic mode function; is the number of natural mode components; is the remainder; (2) The ultrasonic signal is expressed as: , define any algorithm The root mean square error of the denoising deviation is: ； In the formula, For the The ultrasonic signal value at the time, ; is the total number of sampled signals; For the algorithm right Noise reduction results, ; Used to evaluate the error between the reconstructed signal and the original signal; definition For the algorithm Correlation coefficient: ； In the formula, For the The ultrasonic signal value at the time, ; For the algorithm right Noise reduction results, ; Used to evaluate the similarity between the reconstructed signal and the original signal; It is an inner product operation; (3) According to the approximation index value And the correlation coefficient index value , the following good noise reduction criteria are established: Constraints: ； Denoising and reconstruction algorithm The objective function is: ； (4) Iterative operation, obtaining the denoised and reconstructed ultrasonic signal through the algorithm Ra, and extracting the measured thickness value; (5) The ultrasonic flaw detection vehicle scans the entire path to obtain the ultrasonic grayscale image of the defect.

Images