A method and system for acoustic emission monitoring and damage identification of crack propagation in welded joints of high-temperature chemical pipelines.

Abstract

This invention discloses a method and system for acoustic emission monitoring and damage identification of crack propagation in welded joints of high-temperature chemical pipelines, belonging to the field of acoustic emission monitoring technology for pressure pipelines. The system includes a display terminal, a wireless data acquisition unit, a signal transmission line, an acoustic emission sensor, a waveguide rod boss, and a waveguide rod round rod. This system enables real-time acoustic emission monitoring of the propagation behavior of active defects in welded joints of pipelines under high-temperature service conditions. Furthermore, based on this system, the invention provides a method for acoustic emission monitoring and damage identification of crack propagation in welded joints of high-temperature chemical pipelines, offering a signal feature discrimination method for active defects in high-temperature pipeline welded joints, enabling real-time early warning of abnormal signals such as defect propagation. This invention is applicable to wireless real-time acoustic emission monitoring of welded joints in complex high-temperature service environments, enabling rapid and accurate monitoring and identification of active defects in welded joints, and has significant engineering application value.

Description

Technical Field

[0001] This invention belongs to the field of acoustic emission monitoring technology for pressure pipelines, and relates to an acoustic emission monitoring and damage identification method and system for crack propagation in welded joints of high-temperature chemical pipelines. Background Technology

[0002] In chemical production processes, high-temperature pipeline systems undertake critical tasks such as transporting corrosive media and high-pressure steam. Their welded joints, subjected to long-term thermal cycling stress, creep, and electrochemical corrosion, become weak points in structural integrity. Once active defects such as cracks propagate in high-temperature pipeline welded joints, they can cause leaks, and even trigger major accidents such as fires and explosions. Traditional non-destructive testing methods for welded joints, such as radiographic testing and ultrasonic testing, require system shutdown and therefore cannot monitor the propagation of cracks during pipeline service. As modern chemical plants develop towards high-temperature, high-pressure, and long-cycle operation, developing a method and system for real-time monitoring of welded joint crack damage evolution has significant engineering value for preventing catastrophic failures and ensuring intrinsic safety.

[0003] Acoustic emission (AE) monitoring technology captures the elastic waves released during the propagation of internal defects in materials using highly sensitive sensors. Combined with signal processing and analysis methods, it enables real-time dynamic assessment of defects. Compared to traditional non-destructive testing techniques (such as ultrasonic and radiographic testing), AE technology has unique advantages: it is a dynamic, online monitoring method that can directly reflect the evolution of active defects, making it particularly suitable for real-time condition monitoring of welded joints in chemical pipelines. Although AE technology has been applied to some extent in the inspection and monitoring of high-temperature and pressure-bearing equipment such as pipelines, there is still a lack of methods for monitoring and identifying AE signals related to crack propagation in welded joints under high-temperature environments. This makes it difficult for existing technologies to accurately identify AE signals related to crack propagation, leading to misjudgments of crack damage development trends. For example, patent CN 108799846 A discloses an acoustic emission detector and method for pressure pipelines in nuclear power plants, solving the technical problem that acoustic emission sensors cannot be directly attached to the surface of high-temperature pipelines in nuclear power plants for detection. However, the technical solution only discloses the detection method for acoustic emission signals in high-temperature pressure pipelines, and does not address core issues such as crack signal identification and damage assessment, which to some extent affects the accuracy and reliability of pipeline condition assessment. Patent CN 119985722 A discloses an online acoustic emission monitoring device and method for crack propagation in high-temperature pressure equipment. This method analyzes the characteristic parameters (such as amplitude and energy) of acoustic emission signals in the monitoring area through trend analysis to determine whether crack propagation exists. However, this technical solution only makes qualitative judgments based on the changing trends of amplitude and energy. In high-temperature and high-noise environments, it cannot effectively distinguish environmental noise signals with similar characteristic parameters to crack signals, which to some extent affects the accurate identification of crack propagation signals and the reliability of damage assessment in high-temperature pressure equipment.

[0004] Therefore, it is necessary to develop an acoustic emission monitoring and damage identification method and system for crack propagation in welded joints of chemical pipelines in high-temperature environments. This method and system can accurately monitor and identify acoustic emission signals generated by crack propagation in welded joints and evaluate the crack damage status accordingly. This will help avoid cracking or even leakage accidents in welded joints of high-temperature pipelines and improve the safety and reliability of chemical equipment. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of the prior art and provide a method and system for acoustic emission monitoring and damage identification of crack propagation in welded joints of high-temperature chemical pipelines, so as to solve the problem that it is difficult to accurately identify crack propagation signals in acoustic emission monitoring in high-temperature environments in the prior art.

[0006] To achieve the above objectives, the present invention employs the following technical solution: An acoustic emission monitoring and damage identification method for crack propagation in welded joints of high-temperature chemical pipelines includes the following steps: S1, collect and monitor the acoustic emission signals generated by the welded joints of the high-temperature chemical pipeline during service, and calculate multiple acoustic emission characteristic parameters from each acoustic emission signal. After normalization processing, collect and obtain a multi-dimensional acoustic emission characteristic space. The acoustic emission characteristic parameters include amplitude, rise time, root mean square value, peak frequency and centroid frequency. S2, the multidimensional acoustic emission feature space of each acoustic emission signal is input into the optimal SBOA-SVM vector machine to identify the crack propagation signal generated by the welded joint of the high-temperature chemical pipeline during service. Combined with the identification decision function, the signal classification result is obtained, and the output of the classification result is either a crack signal or a noise signal. The optimal SBOA-SVM vector machine is obtained through butterfly optimization algorithm. The specific optimization process is as follows: The butterfly population position is initialized by the multidimensional acoustic emission feature space in the training set. The classification accuracy of each set of parameters on the training set is calculated as the fitness value. Global search and local search are alternately executed to update the butterfly population position. After multiple iterations, the optimal SBOA-SVM vector machine parameters are output to obtain the optimal SBOA-SVM vector machine. S3. Based on the classification results and combined with the crack damage status alarm standard, the degree of crack damage to the welded joint is evaluated and alarmed.

[0007] A further improvement of the present invention is that Preferably, in S1, the amplitude, rise time, and root mean square value are extracted from the time domain waveform, and the peak frequency and centroid frequency are extracted from the frequency domain waveform.

[0008] Preferably, the butterfly population location includes a penalty factor and kernel function parameters.

[0009] Preferably, the formula for calculating the spatially initialized butterfly population location is:

[0010]

[0011] in, It is a calculation The base value, It is a calculation The scaling factor, and It's weight. It is the normalized mean of the training set. It is the normalized mean of the root mean square values ​​in the training set. It is a calculation The base value, It is a calculation The scaling factor, It is the normalized mean of the centroid frequencies in the training set. ε It is the smallest positive number that is prevented from being divided by zero.

[0012] Preferably, in S2, the process of alternately performing global and local searches to update the butterfly population location is as follows: (1) Define the optimization function, construct an SVM model for each multidimensional acoustic emission feature space using the training set, and calculate its fitness value; (2) Perform a global search, select the butterfly with the highest fitness value in the current population as the global optimal solution, and update the positions of other butterflies based on the global optimal solution; (3) Perform a local search, randomly select the positions of two butterflies, calculate their average position as the local optimal solution, and update the positions of other butterflies based on the local optimal solution; (4) Repeat steps (1)-(3) until the maximum number of iterations is reached.

[0013] Preferably, in S2, the optimal SBOA-SVM vector machine calculates its classification accuracy using a test set to obtain the final crack signal recognition model.

[0014] Preferably, the recognition decision function is:

[0015] in, It is the first in the training set i training samples, This is the sample to be predicted. It is a kernel function It is a Lagrange multiplier. It is a bias.

[0016] Preferably, the specific process of S3 is as follows: S31, record and plot the trend of crack propagation signal amplitude and centroid frequency over time; S32, if the signal amplitude suddenly increases to above 80 dB and the centroid frequency drops to below 240 kHz, it is preliminarily judged as severe crack damage; S33, with a step size of 2h, count the number N of crack signals that simultaneously satisfy the conditions of signal amplitude > 80 dB and centroid frequency < 240 kHz within the time window; S34. If N≥30, issue a crack alarm signal; otherwise, continue monitoring.

[0017] Preferably, in S33, the time window is 24 hours.

[0018] An acoustic emission monitoring and damage identification system for crack propagation in welded joints of high-temperature chemical pipelines includes: The acquisition module is used to acquire acoustic emission signals generated by the welded joints of the monitored high-temperature chemical pipeline during service, and to calculate multiple acoustic emission characteristic parameters from each acoustic emission signal. After normalization processing, a multi-dimensional acoustic emission characteristic space is collected. The acoustic emission characteristic parameters include amplitude, rise time, root mean square value, peak frequency and centroid frequency. The judgment module is used to input the multidimensional acoustic emission feature space of each acoustic emission signal into the optimal SBOA-SVM vector machine to identify the crack propagation signal generated by the welded joint of the high-temperature chemical pipeline during service. Combined with the recognition decision function, the signal classification result is obtained, and the output of the classification result is either a crack signal or a noise signal. The optimal SBOA-SVM vector machine is obtained through butterfly optimization algorithm. The specific optimization process is as follows: The butterfly population position is initialized by the multidimensional acoustic emission feature space in the training set. The classification accuracy of each set of parameters on the training set is calculated as the fitness value. Global search and local search are alternately executed to update the butterfly population position. After multiple iterations, the optimal SBOA-SVM vector machine parameters are output to obtain the optimal SBOA-SVM vector machine. The output module is used to evaluate and alarm the degree of crack damage in welded joints based on the classification results and the crack damage status alarm standard.

[0019] Compared with the prior art, the present invention has the following beneficial effects: This invention discloses an acoustic emission monitoring and damage identification method and system for crack propagation in welded joints of high-temperature chemical pipelines. Compared with traditional techniques that rely solely on the trend of signal characteristic parameters for defect evaluation, the method proposed in this invention utilizes a multi-dimensional acoustic emission feature space and a crack propagation signal identification model. This enables signal pattern discrimination of acoustic emission signals generated by welded joints of high-temperature pipelines during service, accurately identifying crack propagation signals, and providing damage evaluation and real-time alarms based on crack damage state alarm standards. Therefore, the crack damage evaluation method is more accurate. Furthermore, this invention discloses an acoustic emission monitoring and damage identification system for crack propagation in welded joints of high-temperature chemical pipelines. This system can wirelessly monitor the crack propagation process of welded joints under high-temperature service conditions and provide crack signal identification and evaluation results. Therefore, this invention is applicable to wireless real-time acoustic emission monitoring of welded joints in complex high-temperature service environments, and has significant engineering application value for achieving rapid and accurate monitoring, identification, and alarming of crack propagation in welded joints. Attached Figure Description

[0020] Figure 1 This is a flowchart of an acoustic emission monitoring and damage identification method for crack propagation in welded joints of high-temperature chemical pipelines according to the present invention.

[0021] Figure 2 This is a schematic diagram of the acoustic emission monitoring and damage identification system for crack propagation in welded joints of high-temperature chemical pipelines according to the present invention.

[0022] The attached diagram shows the markings and corresponding component names: 1. Display and computing terminal; 2. Wireless data acquisition device; 3. Signal transmission line; 4. Acoustic emission sensor; 5. Waveguide rod boss; 6. Waveguide rod round rod; 7. Insulation layer; 8. Pipe base material; 9. Pipe welding joint.

[0023] Figure 3 This is a graph showing the change in acoustic emission signal amplitude over time at a welded joint in a high-temperature chemical pipeline according to an embodiment of the present invention.

[0024] Figure 4 This is a graph showing the change in the centroid frequency of the acoustic emission signal of a welded joint in a high-temperature chemical pipeline over time, according to an embodiment of the present invention.

[0025] Figure 5 To utilize the acoustic emission monitoring and damage identification method for crack propagation in welded joints of high-temperature chemical pipelines obtained by this invention, the acoustic emission signals generated by the welded joints of high-temperature chemical pipelines in the embodiments were analyzed, and the amplitude of the crack propagation signal changing over time was identified.

[0026] Figure 6 To utilize the acoustic emission monitoring and damage identification method for crack propagation in welded joints of high-temperature chemical pipelines obtained by this invention, the acoustic emission signals generated by the welded joints of high-temperature chemical pipelines in the embodiments were analyzed, and the result diagram of the change of the centroid frequency of the crack propagation signal over time was identified. Detailed Implementation

[0027] Hereinafter, the terms "first," "second," "third," and "fourth" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first," "second," "third," or "fourth" may explicitly or implicitly include one or more of that feature.

[0028] The method provided in this application can be applied to mobile phones, tablets, wearable devices, in-vehicle devices, augmented reality (AR) / virtual reality (VR) devices, laptops, and ultra-mobile personal computers. In this application, the specific type of terminal device is not limited to terminal devices such as mobile personal computers (UMPCs), netbooks, and personal digital assistants (PDAs).

[0029] It should be noted that the terms "first," "second," etc., used in the specification and drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0030] This invention discloses an acoustic emission monitoring and damage identification method and system for crack propagation in welded joints of high-temperature chemical pipelines.

[0031] This invention first uses an acoustic emission monitoring system to monitor the acoustic emission signals generated by welded joints in high-temperature chemical pipelines during service in real time. It combines time-domain parameters (amplitude, rise time, root mean square value) with frequency-domain parameters (peak frequency, centroid frequency) to construct a multi-dimensional acoustic emission feature space from each signal. Next, the multi-dimensional acoustic emission feature space calculated for each signal is input into a crack propagation signal identification model to identify crack propagation signals generated by the welded joints in high-temperature chemical pipelines during service. Finally, based on the identified crack propagation signals, the degree of crack damage to the welded joint is evaluated and an alarm is issued based on the trend of signal quantity and feature parameters. Specifically, refer to... Figure 1 This includes the following steps: S1 uses an acoustic emission monitoring and damage identification system to monitor the acoustic emission signals generated by the welded joints of high-temperature chemical pipelines in real time during service. Based on time and frequency domain analysis, acoustic emission parameters are calculated from each signal waveform and normalized to construct a set of multi-dimensional acoustic emission feature spaces.

[0032] The multidimensional acoustic emission feature space in step S1 includes the normalized amplitude, rise time, root mean square value, peak frequency, and centroid frequency. The amplitude, rise time, and root mean square value are extracted from the time domain waveform, while the peak frequency and centroid frequency are extracted from the frequency domain waveform.

[0033] Of the five characteristic parameters mentioned above, amplitude is closely related to crack propagation rate, and tends to increase significantly when microcracks propagate rapidly in high-temperature pipe welded joints. As a time-domain feature, amplitude is used to characterize the intensity of defect activity. Rise time is an indicator describing the dynamic response characteristics of a signal, such as the instantaneous propagation of cracks caused by brittle fracture or local stress concentration; while a longer rise time may originate from environmental noise interference. Therefore, rise time helps distinguish signals generated by different types of source mechanisms. The root mean square (RMS) value reflects the average energy level of the signal throughout its duration, and is more stable than peak amplitude, effectively suppressing the influence of transient interference, and is suitable for energy assessment of continuous acoustic emission events. These three parameters together constitute the basic characterization system of acoustic emission signals in the time dimension, effectively capturing the dynamic behavior of elastic wave propagation during crack propagation. On the other hand, peak frequency refers to the frequency component with the highest energy in the acoustic emission signal spectrum. Different types of defect sources will produce signals with specific frequency distribution characteristics. For example, due to geometric discontinuities and material heterogeneity, high-frequency components in the welded joint region may decay faster, while active crack propagation is often accompanied by energy enhancement in the mid-to-low frequency range. Therefore, peak frequency can be used to identify the physical mechanism of a signal's origin. Centroid frequency, which is the weighted average frequency of the spectral energy distribution, reflects the overall frequency band characteristics of a signal more comprehensively than a single peak, especially in multimodal or broadband signals.

[0034] The five feature parameters mentioned above constitute a five-dimensional feature vector, effectively fusing the spatiotemporal and spectral information of the signal and enhancing the ability to distinguish different types of acoustic emission sources. Normalization is performed using the min-max scaling method or Z-score standardization to eliminate dimensional differences between different parameters, improving the stability and convergence speed of the subsequent classification model. Since features are extracted based on the essential attributes of the signal in both the time and frequency domains, the problem of information distortion or misjudgment caused by mixed extraction is solved, thereby improving the effectiveness and discriminative power of the subsequently constructed multi-dimensional feature space. This lays the foundation for optimizing the input quality of the support vector machine model, ultimately improving the accuracy of crack signal identification and the reliability of damage assessment.

[0035] S2 inputs the multidimensional acoustic emission feature space calculated for each signal into the crack propagation signal identification model to identify the crack propagation signal generated by the welded joint of the high-temperature chemical pipeline during service and eliminate noise signals.

[0036] The crack propagation signal identification model used in step S2 is a Support Vector Machine (SBOA-SVM) model based on the Simplified Butterfly Optimization Algorithm. The model establishment process includes the following specific steps: S21: Conduct fatigue crack propagation tests on pipeline materials using acoustic emission technology, and collect environmental noise during the service of high-temperature pipelines using acoustic emission technology to obtain 10,000 crack propagation signals and noise signals respectively; S22: Construct a dataset where the input features for each dataset are the established multidimensional acoustic emission feature space, and the output features are signal patterns, including crack propagation signals and noise signals. Randomly divide the dataset into 80% training and 20% test sets.

[0037] S23: Train the crack propagation signal recognition model and obtain the optimal model parameters for the recognition of acoustic emission signals generated by welded joints in actual chemical high-temperature pipelines during service.

[0038] Step S23 includes: S231: Randomly generate an initial butterfly population, including... N There are individual butterflies, and the position of each butterfly represents a set of SVM parameters (penalty factor C and kernel function parameter γ). Using... Indicates the first i The location of the butterfly.

[0039]

[0040] S232: Initializing Butterfly Population Locations Based on Acoustic Emission Feature Parameters This improves the quality of the initial solution. Among them,

[0041]

[0042] in, It is a calculation The base value, It is a calculation The scaling factor, and It's weight. It is the normalized mean of the amplitude in the training set as described in step S22. It is the normalized mean of the root mean square values ​​in the training set mentioned in step S22. It is a calculation The base value, It is a calculation The scaling factor, It is the normalized mean of the centroid frequencies in the training set mentioned in step S22. ε It is the smallest positive number that is prevented from being divided by zero.

[0043] In this step, representative statistical indices from the multidimensional acoustic emission feature space are used to guide the population initialization process of SVM hyperparameters. This ensures that the starting point of the Butterfly Optimization Algorithm (SBOA) is no longer a completely random distribution, but a data-driven setting based on physical meaning. This initialization strategy effectively reduces the ineffective search space, increases the probability of high-quality solutions, accelerates global convergence, and enhances the stability of the optimization process.

[0044] S233: Define the optimization function An SVM model is built for each set of parameters using the training set, and its fitness value is calculated.

[0045]

[0046] in, Accuracy It is the classification accuracy. Is it using parameters? The constructed SVM model.

[0047] S234: Begin global search, select the butterfly with the highest fitness value in the current population as the global optimum, and update the positions of other butterflies based on the global optimum:

[0048] in, It is the best individual at present. It is the step size factor. It is the perceived intensity, and its calculation formula is:

[0049] S235: Begin local search, randomly select the positions of two butterflies. and Calculate their average position as a local optimum. Update the positions of other butterflies based on the local optimum:

[0050] S236: Repeat steps S233-S235 until the maximum number of iterations is reached, and output the optimal SVM parameters after SBOA optimization. .

[0051] S237: Use optimal parameters The SVM model is trained on the training set to obtain the optimal model.

[0052]

[0053] S238: On the test set, utilize the optimal model The classification accuracy is calculated, the model's accuracy and generalization ability are evaluated, and a crack propagation signal recognition model is obtained.

[0054] S24: Using a crack propagation signal identification model, the acoustic emission signal patterns generated by welded joints in actual high-temperature chemical pipelines are identified online. The output is +1 to represent a crack signal and -1 to represent a noise signal. The identification decision function in the crack propagation signal identification model is:

[0055] in, It is the first in the training set i training samples, It is the sample to be predicted (the signal data sample generated by actual monitoring). It is a kernel function It is a Lagrange multiplier. It is a bias term. The kernel function maps the nonlinear problem in the original low-dimensional feature space to a high-dimensional space, making it linearly separable in the high-dimensional space; the Lagrange multiplier automatically selects the most discriminative support vectors, reducing the model complexity; the bias term adjusts the position of the classification surface to adapt to the data distribution; finally, fast inference is completed through weighted summation and sign judgment.

[0056] In this step, by explicitly using the penalty factor C and the kernel function parameter γ as encoding variables for the butterfly population position, the optimization process has a clear objective function input dimension, avoiding the blindness and arbitrariness of parameter selection. This parameterized modeling approach helps improve the convergence efficiency and stability of the optimization algorithm, ensuring that the final optimal SBOA-SVM vector machine has a stronger crack signal recognition capability, thereby improving the accuracy and reliability of the acoustic emission monitoring system for welded joints in high-temperature chemical pipelines under service conditions.

[0057] S3, based on the crack propagation signal identified in step S2, evaluate and alarm the degree of crack damage to the welded joint according to the crack damage state alarm standard, specifically including the following steps: S31: Using the arrival time of the crack propagation signal as a sequence, record the changes in the amplitude and centroid frequency of the crack signal and display them as images; S32: Analyze the changing trends of signal amplitude and centroid frequency. If the signal amplitude suddenly and rapidly increases to more than 80 dB compared to the noise signal, and the centroid frequency suddenly drops to below 240 kHz at the same time, it is preliminarily judged that there is relatively serious crack damage. S33: Further, within a sliding time window with a step size of 2 hours. Within the range, count the number N of crack signals that simultaneously satisfy amplitude A > 80 dB and centroid frequency CF < 240 kHz; S34: When N≥30, a crack damage status alarm signal is issued.

[0058] By jointly monitoring the amplitude of crack signals and the frequency of the centroid, combined with a cumulative counting mechanism over time, the accuracy and anti-interference capability of damage assessment are effectively improved. This method not only relies on instantaneous characteristics but also focuses on evolutionary trends, avoiding false alarms caused by individual abnormal signals and preventing the omission of continuous deterioration processes. This significantly enhances the intelligence level and engineering practicality of the online monitoring system for high-temperature chemical pipelines. The alarm mechanism comprehensively considers signal strength, frequency characteristics, and occurrence frequency, avoiding false alarms caused by fluctuations in a single indicator, and improving the robustness and reliability of damage assessment.

[0059] See Figure 2 The acoustic emission monitoring and damage identification system in step S1 includes two parts: an acoustic emission monitoring device and a waveguide rod device. The acoustic emission monitoring device includes a display and computing terminal 1, a wireless data acquisition unit 2, a signal transmission line 3, and an acoustic emission sensor 4; the waveguide rod device includes a waveguide rod boss 5 and a waveguide rod round rod 6.

[0060] One end of the waveguide rod 6 is connected to the surface of the high-temperature pipe base material 8, and the other end of the waveguide rod 6 is connected to the waveguide rod boss 5. An acoustic emission sensor 4 is installed on the waveguide rod boss 5 to conduct acoustic emission signals generated by the expansion of weld joint defects. The distance between the connection position of the waveguide rod 6 and the pipe base material 8 and the weld joint along the radial direction is less than or equal to 5 cm. The acoustic emission sensor 4 is connected to the wireless data acquisition device 2 through the signal transmission line 3 to collect and analyze acoustic emission signal data in real time. The collected data is transmitted to the display terminal 1 through the network to realize the remote monitoring function.

[0061] Furthermore, the waveguide rod 6 is made of stainless steel, whose low thermal conductivity can quickly reduce the heat conduction from the high-temperature pipeline to the sensor; the waveguide rod boss 5 is made of carbon steel, whose ferromagnetism can be used to firmly fix the sensor on the waveguide rod boss 5 with a magnetic clamp.

[0062] The diameter and length of the waveguide rod 6 can be adjusted according to the different service temperatures of the actual chemical pipeline to adapt to the temperature decay, but it must be greater than the thickness of the pipeline insulation layer 7.

[0063] The waveguide rod boss 5 is square with a thickness of 6 mm, and its side length must be larger than the size of the sensor magnetic suction holder.

[0064] Furthermore, the acoustic emission sensor 4 is a narrowband explosion-proof sensor with a built-in preamplifier, used to meet the explosion-proof requirements of chemical sites.

[0065] The acoustic emission sensor 4 is tightly coupled to the waveguide rod boss 5 via a coupling agent to ensure effective transmission of acoustic signals. To ensure stable operation of the sensor in high-temperature environments, a 360° annular sealant is applied to its outer periphery. A magnetic adsorption clamp is used to rigidly fix the sensor to the plane of the waveguide rod boss 5, ensuring stable contact of the sensor under pipeline vibration conditions.

[0066] Furthermore, the connection between the end of the waveguide rod 6 and the surface of the high-temperature pipe material is by arc welding. To ensure the effective propagation of the acoustic emission signal in the welded structure, the weld should be continuous, uniform, without interruption or local abrupt changes, and there should be no defects such as cracks or pores on the surface and inside the weld. After welding, the weld needs to be ground to ensure a smooth and flat surface.

[0067] In some embodiments of the present invention, the distance between the connection position of the waveguide rod 6 and the fusion line of the welded joint 9 of the pipe under test is ≤5 cm, which is used to conduct acoustic emission signals generated by the expansion of defects in the welded joint.

[0068] The wireless data acquisition device 2 is installed in an explosion-proof box to meet the explosion-proof requirements of chemical sites.

[0069] The display and computing terminal 1 is located in the control room. The wireless data acquisition device 2 transmits the collected data to the display and computing terminal 1 in real time through the mobile communication network. This data is used to display the monitoring data and execute steps S2 and S3 to identify crack propagation signals and evaluate crack damage. This enables engineering technicians to remotely monitor and assess crack damage in pipe welded joints 9, which are located in high-temperature and harsh environments, thus avoiding the risk of personnel being directly exposed to hazardous working environments.

[0070] The present invention provides a method and system for acoustic emission monitoring and damage identification of crack propagation in welded joints of high-temperature chemical pipelines. This method and system can effectively solve the problem of online acoustic emission monitoring of crack propagation in welded joints of high-temperature pipelines and accurate evaluation of crack damage, and has broad engineering application value.

[0071] The following description, in conjunction with specific embodiments, provides further details.

[0072] Example 1 In this embodiment, the monitored high-temperature chemical pipeline is made of SA106GR.B steel, and the operating temperature inside the pipeline is 200℃. The crack propagation signal used to construct the crack propagation signal identification model was obtained by conducting acoustic emission monitoring tests on fatigue crack propagation of SA106GR.B steel using an MTS universal testing machine. The noise signal used to construct the crack propagation signal identification model was obtained by collecting ambient noise at the high-temperature chemical pipeline site.

[0073] The specific steps for acoustic emission monitoring and damage identification of crack propagation in welded joints of high-temperature chemical pipelines using the method and system of this invention are as follows: S1 monitors the acoustic emission signals generated by the welded joints of high-temperature chemical pipelines in real time during service through an acoustic emission monitoring and damage identification system. Based on time and frequency domain analysis, acoustic emission parameters are calculated from each signal waveform and normalized to construct a set of multi-dimensional acoustic emission feature spaces, including amplitude, rise time, root mean square value, peak frequency and centroid frequency.

[0074] See Figure 2 Before monitoring, the waveguide rod round rod 6 and the waveguide rod boss 5 are connected by welding. The waveguide rod round rod 6 is welded on the base material surface 8 near the high-temperature pipe welding joint 9. The welding position is arranged at the pipe base material 8 near the fusion line of the pipe welding joint 9 to be tested, about 5 cm away.

[0075] An acoustic emission sensor 4 is installed on the waveguide rod boss 5 and connected to a wireless data acquisition device 2 via a signal transmission line 3. The wireless data acquisition device 2 is placed in an explosion-proof box to meet the explosion-proof requirements of the chemical site. The acoustic emission sensor 4 and the waveguide rod boss 5 are tightly coupled by a coupling agent, and a 360° annular sealant is used to seal the outer periphery of the sensor to ensure stable operation of the sensor in a high-temperature environment. A magnetic adsorption clamp is used to rigidly fix the acoustic emission sensor 4 on the waveguide rod boss 5 to ensure that the sensor can maintain stable contact in the environment of pipeline vibration.

[0076] During on-site monitoring, acoustic emission signals generated during the service of high-temperature pipeline welded joint 9 were collected and transmitted to display and computing terminal 1 via mobile network. Multiple acoustic emission characteristic parameters, including amplitude, rise time, root mean square value, peak frequency and centroid frequency, were calculated from the signals. Figure 3 The graph shows the result of the acoustic emission signal amplitude changing over time at the welded joint of a high-temperature chemical pipeline. Figure 4 The graph shows the result of the change of the centroid frequency of the acoustic emission signal of the welded joint of a high-temperature chemical pipeline over time.

[0077] S2 inputs the multidimensional acoustic emission feature space calculated for each signal into the crack propagation signal identification model to identify the crack propagation signal generated by the welded joint of the high-temperature chemical pipeline during service and eliminate noise signals.

[0078] The crack propagation signal identification model used in step S2 is a Support Vector Machine (SBOA-SVM) model based on the Simplified Butterfly Optimization Algorithm. The butterfly population location is initialized based on acoustic emission feature parameters. In the operation, set the base value Scaling factor is 1. The base value is 1. It is 10. It is 0.001. ε 10 -8 The maximum number of iterations is set to 50. Once the maximum number of iterations is reached, iteration stops, and the optimal SVM parameters optimized by SBOA are output. In this embodiment, the optimal SVM parameters are: The Support Vector Machine (SBOA-SVM) model based on the Simplified Butterfly Optimization algorithm achieved a classification accuracy of 99.6% on the training set and 99.0% on the test set, indicating that the model has high accuracy and generalization ability.

[0079] The constructed crack propagation signal recognition model was used to perform pattern recognition on a total of 6,183 acoustic emission signals generated by the welded joint of the chemical pipeline in the example between November 7 and 14, 2024. A total of 188 crack propagation signals and 5,995 noise signals were identified.

[0080] S3, based on the crack propagation signal identified in step S2, evaluate the degree of crack damage to the welded joint according to the crack damage state alarm standard, specifically including the following steps: S31: Using the arrival time of the crack propagation signal as a sequence, record the changes in the amplitude and centroid frequency of the crack signal and display the results as follows: Figure 5 and Figure 6 As shown; S32: Analyze the changing trends of signal amplitude and centroid frequency. If the signal amplitude suddenly and rapidly increases to more than 80 dB compared to the noise signal, and the centroid frequency suddenly drops to below 240 kHz at the same time, it is preliminarily judged that a relatively serious crack damage has occurred.

[0081] like Figure 5 and Figure 6 As shown, between November 8th and 9th, 2024, and on November 13th, 2024, a large number of crack propagation signals appeared. The signal amplitude suddenly increased and exceeded 80 dB, and at the same time, the signal centroid frequency suddenly decreased to less than 240 kHz. This indicates that the system monitored and identified crack propagation signals of the high-temperature pipeline welded joint at these two times, resulting in relatively serious crack damage.

[0082] S33: Further, within a sliding time window with a step size of 2 hours. Within the range, count the number N of crack signals that simultaneously satisfy amplitude A > 80 dB and centroid frequency CF < 240 kHz; S34: When N≥30, a crack damage status alarm signal is issued.

[0083] Statistical analysis revealed that around November 13th, the number N of crack propagation signals with amplitude A > 80 dB and centroid frequency CF < 240 kHz within a 24-hour time window reached 31, meeting the alarm criteria. The system then triggered an alarm and prompted engineering technicians to go to the site for further processing.

[0084] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0085] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0086] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0087] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A method for acoustic emission monitoring and damage identification of crack propagation in welded joints of high-temperature chemical pipelines, characterized in that, Includes the following steps: S1, collect and monitor the acoustic emission signals generated by the welded joints of the high-temperature chemical pipeline during service, and calculate multiple acoustic emission characteristic parameters from each acoustic emission signal. After normalization processing, collect and obtain a multi-dimensional acoustic emission characteristic space. The acoustic emission characteristic parameters include amplitude, rise time, root mean square value, peak frequency and centroid frequency. S2, the multidimensional acoustic emission feature space of each acoustic emission signal is input into the optimal SBOA-SVM vector machine to identify the crack propagation signal generated by the welded joint of the high-temperature chemical pipeline during service. Combined with the identification decision function, the signal classification result is obtained, and the output of the classification result is either a crack signal or a noise signal. The optimal SBOA-SVM vector machine is obtained through butterfly optimization algorithm. The specific optimization process is as follows: The butterfly population position is initialized by the multidimensional acoustic emission feature space in the training set. The classification accuracy of each set of parameters on the training set is calculated as the fitness value. Global search and local search are alternately executed to update the butterfly population position. After multiple iterations, the optimal SBOA-SVM vector machine parameters are output to obtain the optimal SBOA-SVM vector machine. S3. Based on the classification results and combined with the crack damage status alarm standard, evaluate and alarm the degree of crack damage of the welded joint; The specific process of S3 is as follows: S31, record and plot the trend of crack propagation signal amplitude and centroid frequency over time; S32, if the signal amplitude suddenly increases to above 80 dB and the centroid frequency drops to below 240 kHz, it is preliminarily judged as severe crack damage; S33, with a step size of 2h, count the number N of crack signals that simultaneously satisfy the conditions of signal amplitude > 80 dB and centroid frequency < 240 kHz within the time window; S34. If N≥30, issue a crack alarm signal; otherwise, continue monitoring.

2. The method for acoustic emission monitoring and damage identification of crack propagation in welded joints of high-temperature chemical pipelines according to claim 1, characterized in that, In S1, the amplitude, rise time, and root mean square value are extracted from the time domain waveform, while the peak frequency and centroid frequency are extracted from the frequency domain waveform.

3. The method for acoustic emission monitoring and damage identification of crack propagation in welded joints of high-temperature chemical pipelines according to claim 1, characterized in that, The butterfly population location includes a penalty factor and kernel function parameters.

4. The acoustic emission monitoring and damage identification method for crack propagation in welded joints of high-temperature chemical pipelines according to claim 3, characterized in that, The spatial initialization of butterfly population location The calculation formula is: in, It is a calculation The base value, It is a calculation The scaling factor, and It's weight. It is the normalized mean of the training set. It is the normalized mean of the root mean square values ​​in the training set. It is a calculation The base value, It is a calculation The scaling factor, It is the normalized mean of the centroid frequencies in the training set. ε It is the smallest positive number that is prevented from being divided by zero.

5. The method for acoustic emission monitoring and damage identification of crack propagation in welded joints of high-temperature chemical pipelines according to claim 1, characterized in that, In S2, the process of alternately performing global and local searches to update the butterfly population position is as follows: (1) Define the optimization function, construct an SVM model for each multidimensional acoustic emission feature space using the training set, and calculate its fitness value; (2) Perform a global search, select the butterfly with the highest fitness value in the current population as the global optimal solution, and update the positions of other butterflies based on the global optimal solution; (3) Perform a local search, randomly select the positions of two butterflies, calculate their average position as the local optimal solution, and update the positions of other butterflies based on the local optimal solution; (4) Repeat steps (1)-(3) until the maximum number of iterations is reached.

6. The acoustic emission monitoring and damage identification method for crack propagation in welded joints of high-temperature chemical pipelines according to claim 1, characterized in that, In S2, the optimal SBOA-SVM vector machine calculates its classification accuracy using a test set to obtain the final crack signal recognition model.

7. The acoustic emission monitoring and damage identification method for crack propagation in welded joints of high-temperature chemical pipelines according to claim 1, characterized in that, The identification decision function is: in, It is the first in the training set i training samples, This is the sample to be predicted. It is a kernel function It is a Lagrange multiplier. It is a bias.

8. The method for acoustic emission monitoring and damage identification of crack propagation in welded joints of high-temperature chemical pipelines according to claim 1, characterized in that, In S33, the time window is 24 hours.

9. An acoustic emission monitoring and damage identification system for crack propagation in welded joints of high-temperature chemical pipelines, characterized in that, include: The acquisition module is used to acquire acoustic emission signals generated by the welded joints of the monitored high-temperature chemical pipeline during service, and to calculate multiple acoustic emission characteristic parameters from each acoustic emission signal. After normalization processing, a multi-dimensional acoustic emission characteristic space is collected. The acoustic emission characteristic parameters include amplitude, rise time, root mean square value, peak frequency and centroid frequency. The judgment module is used to input the multidimensional acoustic emission feature space of each acoustic emission signal into the optimal SBOA-SVM vector machine to identify the crack propagation signal generated by the welded joint of the high-temperature chemical pipeline during service. Combined with the recognition decision function, the signal classification result is obtained, and the output of the classification result is either a crack signal or a noise signal. The optimal SBOA-SVM vector machine is obtained through butterfly optimization algorithm. The specific optimization process is as follows: The butterfly population position is initialized by the multidimensional acoustic emission feature space in the training set. The classification accuracy of each set of parameters on the training set is calculated as the fitness value. Global search and local search are alternately executed to update the butterfly population position. After multiple iterations, the optimal SBOA-SVM vector machine parameters are output to obtain the optimal SBOA-SVM vector machine. The output module is used to evaluate and alarm the degree of crack damage in welded joints based on the classification results and the crack damage status alarm standard. The output module performs the following steps: S31, record and plot the trend of crack propagation signal amplitude and centroid frequency over time; S32, if the signal amplitude suddenly increases to above 80 dB and the centroid frequency drops to below 240 kHz, it is preliminarily judged as severe crack damage; S33, with a step size of 2h, count the number N of crack signals that simultaneously satisfy the conditions of signal amplitude > 80 dB and centroid frequency < 240 kHz within the time window; S34. If N≥30, issue a crack alarm signal; otherwise, continue monitoring.

Images