A method, apparatus, equipment, medium, and product for detecting crack propagation state.

Abstract

This invention relates to the field of material structural performance testing technology, and discloses a method, apparatus, equipment, medium, and product for detecting crack propagation state. This invention improves signal purity and eliminates interference from other damage signals by performing noise reduction and blind separation processing on the acoustic emission signal. Furthermore, through feature classification processing, the most accurate set of acoustic emission signal feature parameters reflecting the crack propagation state can be found from numerous features, improving the effectiveness and specificity of the features. Furthermore, by using the fatigue crack propagation acoustic emission signal test method and the Paris formula to establish the relationship between the acoustic emission signal feature parameters and the crack propagation rate, the connection between the acoustic emission signal feature parameters and the crack propagation state can be quantified from a theoretical and experimental perspective. Finally, by combining the acoustic emission signal feature parameter set and its variation relationship, the crack propagation state can be accurately identified in high-noise environments, improving the accuracy and reliability of the detection results.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of material structure performance detection, and in particular to a crack propagation state detection method, device, equipment, medium and product. BACKGROUND

[0002] Crack propagation state monitoring of materials is an important industrial problem related to the safety and reliability of materials. At present, it is found through analysis of various fracture scenarios that the materials that fracture in the fan are mainly metal materials and composite materials.

[0003] At present, when crack propagation monitoring is carried out by using acoustic emission technology, the following problems exist: the acoustic emission signals generated by crack propagation may be mixed with signals generated by other types of damage (such as fiber fracture, matrix cracking, etc.) in the material, making signal analysis complex; acoustic emission signals are usually very weak and can be covered by environmental noise, mechanical noise or electronic instrument interference noise, making signal collection and identification difficult; the prediction of crack propagation rate is crucial for preventing structural failure, but how to accurately predict the crack propagation rate from acoustic emission signals is a technical challenge. Furthermore, the existing crack propagation detection method using acoustic emission technology has low accuracy and reliability. SUMMARY

[0004] Therefore, the present application provides a crack propagation state detection method, device, equipment, medium and product to solve the problem of low accuracy and reliability of the existing crack propagation detection method using acoustic emission technology.

[0005] In a first aspect, the present application provides a crack propagation state detection method for a fan component, which comprises:

[0006] An initial acoustic emission signal set of the fan is obtained; noise reduction and blind separation processing are performed on the initial acoustic emission signal set to obtain a crack propagation signal set of the fan component; feature classification processing is performed on the crack propagation signal set to obtain an acoustic fingerprint signal feature parameter set reflecting the crack propagation state; a fatigue crack propagation acoustic fingerprint signal test method and Paris formula are used to establish a change relationship between the acoustic fingerprint signal feature parameter and the crack propagation rate; based on the acoustic fingerprint signal feature parameter set and the change relationship, the crack propagation state of the fan component is detected to obtain a crack propagation state detection result.

[0007] The crack propagation state detection method provided by the application can effectively remove environmental noise, mechanical noise, electronic instrument interference noise and the like by performing noise reduction processing on the initial acoustic emission signal set, and improves the purity of the signal. Meanwhile, by blind separation processing, the crack propagation signal can be focused on, and the interference of other damage signals is excluded, and the accuracy of the crack propagation state judgment is improved. Further, by performing feature classification processing on the crack propagation signal set, the acoustic fingerprint signal feature parameter set that can most accurately reflect the crack propagation state can be found from numerous features, and redundant or irrelevant features are removed, and the effectiveness and pertinence of the features are improved. Further, the acoustic fingerprint signal feature parameter and the change relationship of the crack propagation rate are established by using the fatigue crack propagation acoustic fingerprint signal test method and the Paris formula, the relationship between the acoustic fingerprint signal feature parameter and the crack propagation state can be quantified from the perspective of combining theory and experiment, and the scientificity and accuracy of the crack propagation state judgment are improved. Finally, the acoustic fingerprint signal feature parameter set and the change relationship can be combined to accurately identify the crack propagation state in a high-noise environment, improve the accuracy and reliability of the detection result, effectively reduce the false positive rate, and provide a strong guarantee for the safe operation of the fan.

[0008] In an optional embodiment, the initial acoustic emission signal set is subjected to noise reduction and blind separation processing to obtain a crack propagation signal set of the fan component, including:

[0009] The initial acoustic emission signal set is subjected to noise reduction processing by using a preset signal noise reduction combination algorithm to obtain a target acoustic emission signal set, and the target acoustic emission signal set is subjected to blind separation by using a nonlinear principal component analysis algorithm to obtain the crack propagation signal set of the fan component.

[0010] The crack propagation state detection method provided by the application can effectively remove environmental noise, mechanical noise, electronic instrument interference noise and the like by using the preset signal noise reduction combination algorithm, and improve the purity of the signal. Further, the acoustic emission signal generated by crack propagation and the mixed signal generated by other types of damage (such as fiber fracture, matrix cracking, etc.) in the material can be separated by using the nonlinear principal component analysis algorithm for blind separation, and then the crack propagation signal can be focused on, and the interference of other damage signals is excluded, and the accuracy of the crack propagation state judgment is improved.

[0011] In an optional embodiment, the target acoustic emission signal set is subjected to blind separation by using the nonlinear principal component analysis algorithm to obtain the crack propagation signal set of the fan component, including:

[0012] The target acoustic emission signal set is subjected to time-varying linear mixing to obtain a mixed acoustic emission signal set, a target function is constructed by using a minimum mean square error approximation criterion, and the crack propagation signal set is obtained by processing through a gradient descent algorithm based on the mixed acoustic emission signal set and the target function.

[0013] The crack propagation state detection method provided by the application can recombine the original acoustic emission signals by time-varying linear mixing of the target acoustic emission signal set. Further, a target function is constructed by using a minimum mean square error approximation criterion, so that the target function can approximate the real crack propagation signal as much as possible, and the signal processing target can be converted into a mathematically optimizable problem. Finally, the mixed acoustic emission signal set and the target function are processed by using a gradient descent algorithm, so that the minimum value of the target function can be quickly and effectively found, the influence of other interference signals on the crack propagation signal is further reduced, the subsequent crack state analysis based on the separated signal is more reliable, and the accuracy of the entire crack propagation state monitoring is effectively improved.

[0014] In an optional embodiment, the crack propagation signal set is subjected to feature classification processing to obtain a voiceprint signal feature parameter set reflecting the crack propagation state, including:

[0015] The crack propagation signal set is processed by a one-dimensional convolutional neural network algorithm to obtain a fault feature set;

[0016] The fault feature set is processed by a bird swarm algorithm to obtain a voiceprint signal feature parameter set reflecting the crack propagation state.

[0017] The crack propagation signal set is processed by a one-dimensional convolutional neural network algorithm to obtain a fault feature set, which can effectively extract the fault feature, further reduces the interference of human factors, and improves the accuracy and efficiency of feature extraction. Further, the intelligent optimization characteristics of the bird swarm algorithm are used to further screen and optimize the features and select the voiceprint signal feature parameter set that can most accurately reflect the crack propagation state from the numerous features, so that redundant or irrelevant features are removed, and the effectiveness and pertinence of the features are improved.

[0018] In an optional embodiment, a fatigue crack propagation voiceprint signal test method and Paris formula are used to establish a change relationship between the voiceprint signal feature parameter and the crack propagation rate, including:

[0019] A first change relationship curve between the crack propagation rate and the stress intensity factor range and a second change relationship curve between the voiceprint signal feature parameter and the cycle number are established by using the fatigue crack propagation voiceprint signal test method; based on the first change relationship curve and the second change relationship curve, a third correlation relationship between the voiceprint signal feature parameter and the stress intensity factor range and a fourth correlation relationship between the crack propagation rate and the stress intensity factor range are determined by using the Paris formula; and the change relationship between the voiceprint signal feature parameter and the crack propagation rate is determined according to the third correlation relationship and the fourth correlation relationship.

[0020] The crack propagation state detection method provided by the application can directly obtain actual crack propagation data and corresponding acoustic fingerprint signal characteristic parameters and the change of the acoustic fingerprint signal characteristic parameters with the cycle number by carrying out a fatigue crack propagation acoustic fingerprint signal test. Further, the Paris formula can be used to further study the internal relationship between the acoustic fingerprint signal characteristic parameters and the stress intensity factor range and more accurately determine the quantitative relationship between the crack propagation rate and the stress intensity factor range. Finally, the change relationship between the acoustic fingerprint signal characteristic parameters and the crack propagation rate can be finally determined according to the determined third and fourth correlation relationships. Therefore, by implementing the application, the relationship between the acoustic fingerprint signal characteristic parameters and the crack propagation state is quantified from the perspective of combining theory and experiment, and the scientificity and accuracy of crack propagation state judgment are improved.

[0021] In an optional embodiment, based on the acoustic fingerprint signal characteristic parameter set and the change relationship, the crack propagation state of the fan component is detected to obtain a crack propagation state detection result, including:

[0022] The change of the acoustic fingerprint signal characteristic parameters and a plurality of first crack propagation states are determined according to the acoustic fingerprint signal characteristic parameter set. A plurality of second crack propagation states are determined based on the change and the change relationship. The crack propagation state detection result is determined according to the plurality of first crack propagation states and the plurality of second crack propagation states.

[0023] The crack propagation state detection method provided by the application can directly obtain actual crack propagation data and corresponding acoustic fingerprint signal characteristic parameters and the change of the acoustic fingerprint signal characteristic parameters with the cycle number by carrying out a fatigue crack propagation acoustic fingerprint signal test. Further, the Paris formula can be used to further study the internal relationship between the acoustic fingerprint signal characteristic parameters and the stress intensity factor range and more accurately determine the quantitative relationship between the crack propagation rate and the stress intensity factor range. Finally, the change relationship between the acoustic fingerprint signal characteristic parameters and the crack propagation rate can be finally determined according to the determined third and fourth correlation relationships. Therefore, by implementing the application, the relationship between the acoustic fingerprint signal characteristic parameters and the crack propagation state is quantified from the perspective of combining theory and experiment, and the scientificity and accuracy of crack propagation state judgment are improved.

[0024] In a second aspect, the application provides a crack propagation state detection device for a fan component, including:

[0025] The acquisition module is configured to acquire an initial acoustic emission signal set of the fan; the first processing module is configured to perform noise reduction and blind separation processing on the initial acoustic emission signal set to obtain a crack propagation signal set of the fan component; the second processing module is configured to perform feature classification processing on the crack propagation signal set to obtain an acoustic print signal feature parameter set reflecting the crack propagation state; the establishment module is configured to establish a change relationship between the acoustic print signal feature parameter and the crack propagation rate by using a fatigue crack propagation acoustic print signal test method and a Paris formula; and the detection module is configured to detect the crack propagation state of the fan component based on the acoustic print signal feature parameter set and the change relationship to obtain a crack propagation state detection result.

[0026] In a third aspect, the present application provides a computer device, comprising a memory and a processor, the memory and the processor are communicatively connected with each other, the memory stores computer instructions, and the processor executes the computer instructions to perform the crack propagation state detection method of the first aspect or any of the corresponding embodiments thereof.

[0027] In a fourth aspect, the present application provides a computer readable storage medium, which stores computer instructions, and the computer instructions are used to make a computer execute the crack propagation state detection method of the first aspect or any of the corresponding embodiments thereof.

[0028] In a fifth aspect, the present application provides a computer program product, which comprises computer instructions, and the computer instructions are used to make a computer execute the crack propagation state detection method of the first aspect or any of the corresponding embodiments thereof. BRIEF DESCRIPTION OF DRAWINGS

[0029] In order to more clearly illustrate the specific embodiments of the present application or the technical solutions in the prior art, the drawings needed in the specific embodiments or the prior art description will be briefly introduced as follows. Obviously, the drawings in the following description are some embodiments of the present application, and other drawings can also be obtained by those skilled in the art without creative labor.

[0030] Figure 1 is a flowchart of a crack propagation state detection method according to an embodiment of the present application;

[0031] Figure 2 is a flowchart of another crack propagation state detection method according to an embodiment of the present application;

[0032] Figure 3 is a flowchart of still another crack propagation state detection method according to an embodiment of the present application;

[0033] Figure 4 is a structural block diagram of a crack propagation state detection device according to an embodiment of the present application;

[0034] Figure 5 Figure 1 is a schematic diagram of the hardware structure of a computer device according to an embodiment of the present application. DETAILED DESCRIPTION

[0035] In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are some but not all of the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work fall within the protection scope of the present application.

[0036] Crack propagation monitoring of materials is an important industrial issue related to the safety and reliability of materials. Existing common methods include:

[0037] Ultrasonic testing (UT): By emitting ultrasonic waves and receiving their reflected signals, cracks inside the material can be detected. Cracks change the propagation path and speed of sound waves, which are detected.

[0038] Radiographic testing (RT): Using X-rays or gamma rays to penetrate materials, cracks will block the rays, causing shadows on the film or digital detector, thus revealing the existence of cracks.

[0039] Magnetic particle testing (MT): Using the principle that cracks in the magnetic field will attract magnetic particles, cracks are detected by observing the distribution of magnetic particles.

[0040] Penetrant testing (PT): A penetrant containing fluorescent or dye is applied to the surface of the material, cracks will absorb the penetrant, and then a developer is used to make the penetrant in the cracks visible.

[0041] Eddy current testing (ET): By generating eddy currents on the surface of the material, cracks will change the distribution of eddy currents, and cracks are detected by detecting changes in eddy currents.

[0042] Acoustic emission (AE): When cracks propagate, acoustic waves are generated, which can be captured by sensors installed on the surface of the material, allowing real-time monitoring of crack propagation.

[0043] Optical microscope: Under laboratory conditions, an optical microscope can be used to observe cracks on the surface of the material.

[0044] Electron microscope: Higher resolution than optical microscope, can observe smaller cracks.

[0045] Digital image correlation (DIC): By comparing images before and after crack propagation, the displacement and propagation of cracks can be calculated.

[0046] Stress Wave Analysis: By analyzing the propagation characteristics of stress waves within the material, the presence and propagation of cracks can be inferred.

[0047] Optical Fiber Sensors: Embed optical fibers within the material, and the propagation of cracks will cause the optical fibers to break or the optical signal to change, thus monitoring the cracks.

[0048] Smart Materials and Structures: Use smart materials such as piezoelectric materials or shape memory alloys, which can generate electrical signals or shape changes when cracks propagate, thus being monitored.

[0049] The detection method of the crack propagation state in the material depends on the type of material, the size of the crack, the monitoring frequency, whether real-time monitoring is required, and cost factors. In practical applications, due to different usage scenarios, a combination of multiple crack detection methods may be required to optimize the combination, in order to improve the accuracy and reliability of crack monitoring.

[0050] For wind turbines, the key components include blades, towers, generators, gearboxes, bearings, etc. These components are subjected to significant mechanical stress and environmental influences during operation, and offshore wind power is even more severely tested due to the effects of salt spray, making the service life of the wind turbine a serious challenge. Some key structural components are prone to fracture, causing significant economic losses. According to statistics, cracks caused by wind turbine fractures generally occur in the locations of blades, towers, generators, gearboxes, various shafts and bearings, bolts, and connecting parts.

[0051] In the blade, the tip: due to the large bending stress and centrifugal force at the tip, it is prone to fracture; the root of the blade: the connection between the blade and the hub bears a large bending moment and torque, which is a common location for fracture; the surface of the blade: due to fatigue, corrosion, or impact damage, cracks may appear on the surface of the blade and propagate to fracture.

[0052] In the tower, the connecting flange: the flange connecting the tower to the foundation, blades, or generator may fracture due to bearing large loads; the bottom of the tower: the connection between the bottom of the tower and the foundation may fracture due to bearing the weight of the entire wind turbine and wind load; the middle section of the tower: the middle section of the tower may fracture due to corrosion, fatigue, or manufacturing defects.

[0053] In the generator, the shaft: the shaft of the generator may fracture due to fatigue or overload during long-term operation; the bearing: the bearing of the generator may be damaged due to poor lubrication, overheating, or fatigue, leading to shaft fracture.

[0054] In the gearbox, the gear: the gear in the gearbox may fracture due to fatigue, overload, or manufacturing defects; the bearing: the bearing in the gearbox may be damaged due to poor lubrication, overheating, or fatigue, leading to gear fracture.

[0055] Among various types of bearings, the rolling elements (balls or rollers) inside the bearing can break due to fatigue or impact; the bearing raceways can break due to overloading or manufacturing defects.

[0056] Among various types of bolts and connectors, the bolts connecting the components of the fan can break due to fatigue, overloading, or corrosion; the connectors can break due to fatigue, overloading, or corrosion.

[0057] Therefore, by analyzing various fracture scenarios, it is not difficult to conclude that the materials that fracture in the fan are mainly metal materials and composite materials.

[0058] Further, the acoustic emission signals during the fracture process of metal materials usually exhibit three stages: crack nucleation: under the action of stress, the micro-defects inside the material begin to expand, forming a small crack, which produces acoustic emission signals. Stable crack propagation: as the load continues to act, the crack gradually expands, accompanied by the continuous generation of acoustic emission signals. Rapid crack: when the crack expands to the critical size, the remaining material cannot withstand the external stress, leading to fracture, which produces a large amount of acoustic emission signals.

[0059] The fracture process of composite materials can involve multiple different failure mechanisms, including but not limited to: fiber breakage: the breakage of reinforcing fibers in composite materials releases energy and produces acoustic emission signals. Matrix breakage: the breakage of matrix materials in composite materials also produces acoustic emission signals. Interface debonding: the interface debonding between fibers and matrix also produces acoustic emission signals. Crack propagation: the formation of new surfaces during the crack propagation process in composite materials also produces acoustic emission signals.

[0060] Therefore, as long as the correspondence between the acoustic emission signal parameters and the crack propagation state in metal materials and composite materials is established, real-time monitoring and analysis of the crack propagation state of key structural components in the fan using acoustic emission technology can be achieved, and the initiation and propagation of cracks can be detected in a timely manner, thereby preventing the failure of the structure.

[0061] According to an embodiment of the present application, a crack propagation state detection method embodiment is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although the logical order is shown in the flowchart, in some cases, the steps shown or described herein can be executed in a different order.

[0062] In this embodiment, a crack propagation state detection method is provided for fan components. Figure 1 The flowchart of the crack propagation state detection method according to an embodiment of the present application is shown in FIG. 1, which includes the following steps: Figure 1 ​

[0063] Step S101, obtaining an initial acoustic emission signal set of the wind turbine.

[0064] The initial acoustic emission signal set represents a collection of original acoustic wave signals collected by acoustic emission sensors installed on the surface of key components of the wind turbine (such as blades, towers, generators, gearboxes, bearings, etc.) during operation of the wind turbine, and can include various acoustic wave information generated by the wind turbine components under various working conditions such as mechanical stress, environmental impact (such as salt spray environment for offshore wind turbines), etc.

[0065] Step S102, performing noise reduction and blind separation processing on the initial acoustic emission signal set to obtain a crack propagation signal set of the wind turbine component.

[0066] The noise reduction processing can use methods such as wavelet transform, adaptive filtering (such as least mean square algorithm and recursive least squares algorithm, etc.).

[0067] Further, the blind separation processing can use methods such as nonlinear principal component analysis (NLPCA), independent component analysis (ICA), and clustering-based blind separation.

[0068] Further, the crack propagation signal set represents a collection of acoustic emission signals directly related to the crack propagation process of the wind turbine component, extracted from the initial acoustic emission signal set after noise reduction and blind separation processing, and can reflect acoustic fingerprint information generated at various stages such as crack initiation, stable propagation, and possible unstable propagation.

[0069] Specifically, by performing noise reduction processing on the initial acoustic emission signal set, environmental noise, mechanical noise, and electronic instrument interference noise can be effectively removed, improving the purity of the signal. At the same time, by performing blind separation processing, the interference of other damage signals can be excluded, and then the crack propagation signals can be focused on and the corresponding crack propagation signal set can be obtained.

[0070] Step S103, performing feature classification processing on the crack propagation signal set to obtain an acoustic fingerprint signal feature parameter set reflecting the crack propagation state.

[0071] The feature classification processing can use methods such as deep learning, machine learning, and statistical-based methods; and the acoustic fingerprint signal feature parameter set can include time domain feature parameters (such as ring count, rise time, duration, etc.) and frequency domain feature parameters (such as center frequency, frequency bandwidth, etc.).

[0072] Specifically, by performing feature extraction and classification on the crack propagation signal set, the acoustic fingerprint signal feature parameter set that most accurately reflects the crack propagation state can be found from numerous features, removing redundant or irrelevant features and improving the effectiveness and relevance of the features.

[0073] In an optional embodiment, the crack propagation signal set is processed for feature classification by a deep neural network.

[0074] Firstly, the crack propagation signal set is converted into a form suitable for neural network input, such as digitizing and normalizing the signal.

[0075] Secondly, a neural network structure with multiple hidden layers is constructed, and the number of layers and the number of neurons in each layer are determined according to the dimension of the input data and the complexity of the problem.

[0076] Further, the network is trained using labeled training data (data with known crack propagation state), and the weights of the network are adjusted through a backpropagation algorithm to minimize the error between the predicted result and the true label.

[0077] Finally, after multiple iterations of training, the network can learn the mapping from the input signal to the crack propagation state, and can then classify new crack propagation signals and obtain the corresponding voiceprint signal feature parameter set.

[0078] Step S104, using the fatigue crack propagation voiceprint signal test method and the Paris formula, the relationship between the voiceprint signal feature parameters and the crack propagation rate is established.

[0079] Among them, the fatigue crack propagation voiceprint signal test method represents a test method for studying the change law of acoustic emission signals (voiceprint signals) during the crack propagation process of materials under fatigue load. During the test, periodic fatigue load is applied to the test sample (usually the material used by the fan key component, such as metal or composite material) to simulate the alternating stress that the component is subjected to under actual working conditions. At the same time, acoustic emission sensors can be used to monitor the acoustic emission signals generated during the crack initiation, propagation and fracture process of the material inside.

[0080] Further, the Paris formula (Paris Law) represents a formula describing the relationship between the fatigue crack propagation rate and the stress intensity factor range.

[0081] Specifically, by combining the fatigue crack propagation voiceprint signal test method and the Paris formula, the relationship between the voiceprint signal feature parameters and the crack propagation state can be quantified from the perspective of theory and experiment, thereby improving the scientificity and accuracy of crack propagation state judgment.

[0082] Step S105, based on the voiceprint signal feature parameter set and the change relationship, the crack propagation state of the fan component is detected to obtain the crack propagation state detection result.

[0083] Specifically, by analyzing the current state and the trend of change of the voiceprint signal feature parameters, and then combining the change relationship, the crack propagation state can be accurately identified in a high-noise environment, and finally the crack propagation state detection result is obtained, thereby improving the accuracy and reliability of the detection result.

[0084] The crack propagation state detection method provided in the embodiment can effectively remove environmental noise, mechanical noise, electronic instrument interference noise and the like by performing noise reduction processing on the initial acoustic emission signal set, thereby improving the purity of the signal. Meanwhile, by blind separation processing, the crack propagation signal can be focused on, and the interference of other damage signals is excluded, thereby improving the accuracy of the crack propagation state judgment. Further, by performing feature classification processing on the crack propagation signal set, the voiceprint signal feature parameter set that can most accurately reflect the crack propagation state can be found from numerous features, thereby removing redundant or irrelevant features, and improving the effectiveness and pertinence of the features. Further, the change relationship between the voiceprint signal feature parameters and the crack propagation rate is established by using the fatigue crack propagation voiceprint signal test method and the Paris formula, thereby quantifying the relationship between the voiceprint signal feature parameters and the crack propagation state from the perspective of combining theory and experiment, and improving the scientificity and accuracy of the crack propagation state judgment. Finally, by combining the voiceprint signal feature parameter set and the change relationship, the crack propagation state can be accurately identified in a high-noise environment, thereby improving the accuracy and reliability of the detection result, effectively reducing the false positive rate, and providing a strong guarantee for the safe operation of the fan.

[0085] In the embodiment, a crack propagation state detection method is provided for a fan component. Figure 2 is a flowchart of the crack propagation state detection method according to the embodiment of the application, as shown in the figure, the flowchart comprises the following steps: Figure 2

[0086] Step S201: obtaining an initial acoustic emission signal set of a fan. For details, refer to step S101 of the embodiment shown in Figure 1 and will not be described here again.

[0087] Step S202: performing noise reduction and blind separation processing on the initial acoustic emission signal set to obtain a crack propagation signal set of the fan component.

[0088] Specifically, the above step S202 comprises:

[0089] Step S2021: performing noise reduction processing on the initial acoustic emission signal set by using a preset signal noise reduction combination algorithm to obtain a target acoustic emission signal set.

[0090] ​The preset signal denoising combination algorithm represents a method that integrates multiple denoising technologies, and is used for removing noise in the initial acoustic emission signal set to improve signal quality and obtain a purer target acoustic emission signal set. The preset signal denoising combination algorithm in this embodiment is a Bayesian optimization-wavelet thresholding-variational modal decomposition (BO-WT-VMD) algorithm.

[0091] First, the wavelet layer number is determined and denoising is performed by using wavelet transform. The wavelet transform represents a time-frequency analysis method, which can decompose a signal into wavelet coefficients of different frequencies.

[0092] Specifically, the characteristics of the original signal are analyzed, and the appropriate wavelet layer number is determined, and then the original signal is denoised according to the wavelet thresholding method.

[0093] The wavelet thresholding denoising means that the coefficients with small absolute values (which are generally considered to be caused by noise) in the wavelet coefficients are set to zero or are shrunk, while the coefficients with large absolute values (which are considered to be the main components of the signal) are retained, so as to achieve the purpose of removing noise.

[0094] Second, the parameters K and a of VMD are determined by using Bayesian optimization. The VMD (variational modal decomposition) represents an adaptive signal decomposition method; K represents the number of modalities of decomposition; and a represents a parameter related to the bandwidth constraint.

[0095] Specifically, the hyperparameter combination and the prior data set are randomly initialized. Further, the sampling points can be increased by using a Gaussian process (GP) to obtain a posterior data set.

[0096] Further, if the posterior data set satisfies a termination condition (for example, a certain accuracy requirement or a number of iterations, etc.), the optimized parameter combination is determined to obtain the parameters of the wavelet transform; if the termination condition is not satisfied, the loop Gaussian process and the previous steps are performed until the posterior data set satisfies the termination condition, and finally the optimized parameter combination is determined to obtain the parameters K and a of VMD.

[0097] Finally, the parameters K and a obtained are substituted into the VMD algorithm to obtain the decomposed modalities and select the optimal modality.

[0098] Specifically, the parameters K and a obtained are substituted into the VMD algorithm to decompose the signal and obtain K modalities. Further, the optimal modality can be selected by using the Pearson coefficient, and then the remaining modalities are removed, and finally the reconstructed data after denoising, i.e., the target acoustic emission signal set, is obtained.

[0099] The Pearson coefficient represents an index for measuring the linear correlation between two variables.

[0100] Step S2022, performing blind separation on the target acoustic emission signal set by using a nonlinear principal component analysis algorithm to obtain a crack propagation signal set of the fan component.

[0101] The nonlinear principal component analysis (NLPCA) algorithm represents a principal component analysis method for processing nonlinear data, which can map high-dimensional nonlinear data to a low-dimensional space while preserving the main information and structure of the data as much as possible.

[0102] In some optional embodiments, the step S2022 includes:

[0103] Step a1, performing time-varying linear mixing on the target acoustic emission signal set to obtain a mixed acoustic emission signal set.

[0104] Step a2, constructing a target function by using a minimum mean square error approximation criterion.

[0105] Step a3, obtaining the crack propagation signal set by using a gradient descent algorithm based on the mixed acoustic emission signal set and the target function.

[0106] Specifically, assuming that the target acoustic emission signal set is s(t), the time-varying system linear mixing obtains the mixed acoustic emission signal set x(t), that is,

[0107] wherein the time-varying linear mixing represents the change of the coefficient (represented by ) in the mixing process with time, which can simulate the time-varying interference and mixing conditions that the actual signal may be subjected to in the transmission or collection process.

[0108] Further, the minimum mean square error approximation criterion can be used as an independence criterion to construct the corresponding target function J(W) as shown in the following relation (1):

[0109]

[0110] In the formula: represents the generalized inverse matrix of the time-varying system to be solved; represents a nonlinear odd function; and E represents a mathematical expectation.

[0111] Further, the purpose of the above target function is to find the optimal by minimizing the error, so that the separated signals are as close as possible to the original independent signal sources.

[0112] Further, the gradient descent algorithm is used for solving, and the iteration formula is shown in the following relation (2):

[0113]

[0114] wherein: μ represents a learning rate, used to control the step size of each iteration; LT represents a lower triangular matrix.

[0115] Further, in the iteration process, according to the mixed acoustic emission signal set x(t) and the current The gradient of the objective function is calculated, and then the relationship (2) is updated according to the above Iterate until convergence (for example, the value of the objective function no longer changes significantly or reaches a preset number of iterations), and further, the final for separating the mixed signal.

[0116] Finally, the crack propagation signal set y(t) of the fan component is obtained, that is,

[0117] Further, when y(t) is a sub-Gaussian function, When y(t) is a super-Gaussian signal,

[0118] Step S203, performing feature classification processing on the crack propagation signal set to obtain a voiceprint signal feature parameter set reflecting the crack propagation state.

[0119] Specifically, the above step S203 includes:

[0120] Step S2031, processing the crack propagation signal set through a one-dimensional convolutional neural network algorithm to obtain a fault feature set.

[0121] The fault feature set can include a plurality of key information capable of reflecting the crack propagation state.

[0122] Further, the one-dimensional convolutional neural network (1DCNN) algorithm represents a deep learning model for processing one-dimensional sequence data, which can automatically learn feature representations from input data through convolutional layers, pooling layers, and fully connected layers, and can be suitable for tasks such as feature extraction and classification of time series data.

[0123] Specifically, the crack propagation signal set can be sorted and segmented to form a fixed segment data sample set. The data sample set can include laboratory runner material fatigue test voiceprint signals, model runner experimental voiceprint signals in different water environments, etc., to ensure that the data is representative and diverse.

[0124] Further, the obtained data sample set is input into the convolutional layer of the 1DCNN, and then the convolution kernel in the convolutional layer can slide on the input data and extract local features through convolution operation.

[0125] The weights of the convolution kernel can be automatically learned through training, and various patterns and features in the crack propagation signal, such as different frequency components and trends in time series, can be adaptively captured.

[0126] Further, after the processing of the convolution layer, the extracted features can be transmitted to subsequent layers (which can include a pooling layer for feature dimension reduction and a fully connected layer for feature integration), and finally output by the fully connected layer and sent to the output layer, thereby outputting the fault feature set extracted after the one-dimensional convolution neural network algorithm processing.

[0127] In step S2032, the fault feature set is processed by the bird swarm algorithm to obtain the voiceprint signal feature parameter set reflecting the crack propagation state.

[0128] The bird swarm algorithm represents an intelligent optimization algorithm based on the group behavior of birds, which can simulate the cooperation and competition mechanism of birds in foraging and flying, and find the optimal solution by constantly updating the position and speed of each individual in the bird swarm.

[0129] Specifically, the bird swarm algorithm can be used to optimize the values of the penalty parameter c and the radial basis kernel function parameter g in the support vector machine (SVM) model.

[0130] In the bird swarm algorithm, each bird is regarded as a possible solution (i.e., a set of values of c and g), and the bird swarm flies and forages in the search space, constantly updating the position of the bird (i.e., updating the parameter values) to find the optimal parameter combination.

[0131] Further, the individual (bird) keyi1 in the bird swarm algorithm adjusts its speed and position according to its own historical optimal position, the historical optimal position of the group, and some random factors, so as to explore and utilize in the entire search space to find the parameter values that optimize the performance of the SVM model.

[0132] Further, the fault feature set is input into the SVM classifier optimized by the bird swarm.

[0133] Further, the optimized SVM classifier can classify the fault feature set according to the optimized parameters c and g, and finally obtain the voiceprint signal feature parameter set reflecting the crack propagation state.

[0134] In step S204, the relationship between the voiceprint signal feature parameter and the crack propagation rate is established using the fatigue crack propagation voiceprint signal test method and the Paris formula. For details, please refer to Figure 1 The step S104 of the embodiment shown in the figure will not be repeated here.

[0135] Step S205, based on the voiceprint signal feature parameter set and the change relationship, detecting the crack propagation state of the fan component to obtain a crack propagation state detection result. For details, please refer to Figure 1 Step S105 of the embodiment shown will not be described here.

[0136] The crack propagation state detection method provided in this embodiment can effectively remove environmental noise, mechanical noise, electronic instrument interference noise and the like by using the preset signal noise reduction combination algorithm, thereby improving the purity of the signal. Further, the original acoustic emission signal set is recombined by time-varying linear mixing. Further, the objective function is constructed by using the minimum mean square error approximation criterion, so that the objective function can approximate the real crack propagation signal as much as possible, and the signal processing target can be converted into a mathematically optimizable problem. Finally, the mixed acoustic emission signal set and the objective function are processed by using the gradient descent algorithm, so that the minimum value of the objective function can be quickly and effectively found, and the influence of other interference signals on the crack propagation signal is further reduced. Further, the crack propagation signal set is processed by using the one-dimensional convolutional neural network algorithm, so that the fault features can be effectively extracted, and the interference of human factors is further reduced, and the accuracy and efficiency of feature extraction are improved. Further, the intelligent optimization characteristics of the bird swarm algorithm are used to further screen and optimize the features and select the voiceprint signal feature parameter set that can most accurately reflect the crack propagation state from the numerous features, so that redundant or irrelevant features are removed, and the effectiveness and pertinence of the features are improved.

[0137] In this embodiment, a crack propagation state detection method is provided for a fan component. Figure 3 is a flowchart of the crack propagation state detection method according to the embodiment of the application, as shown in Figure 3 The flowchart includes the following steps:

[0138] Step S301, obtaining an initial acoustic emission signal set of the fan. For details, please refer to Figure 1 Step S101 of the embodiment shown will not be described here.

[0139] Step S302, performing noise reduction and blind separation processing on the initial acoustic emission signal set to obtain a crack propagation signal set of the fan component. For details, please refer to Figure 2 Step S202 of the embodiment shown will not be described here.

[0140] Step S303, performing feature classification processing on the crack propagation signal set to obtain a voiceprint signal feature parameter set reflecting the crack propagation state. For details, please refer to Figure 2 Step S203 of the embodiment shown will not be described here.

[0141] Step S304, using the fatigue crack propagation acoustic fingerprint signal test method and the Paris formula, the relationship between the acoustic fingerprint signal characteristic parameters and the crack propagation rate is established.

[0142] Specifically, the step S304 includes:

[0143] Step S3041, using the fatigue crack propagation acoustic fingerprint signal test method to establish the first variation relationship curve between the crack propagation rate and the stress intensity factor range, and the second variation relationship curve between the acoustic fingerprint signal characteristic parameters and the cycle number.

[0144] Specifically, the conventional metal material for the fan can be selected and the metal material fatigue crack propagation acoustic fingerprint signal test can be carried out.

[0145] During the test, a plurality of key parameters can be recorded and measured, such as the fatigue crack length a, the stress intensity factor range ΔK, the cycle number N, the fatigue crack propagation rate da / dN, and the acoustic emission characteristic parameters (such as ringing count, amplitude, energy count, etc.).

[0146] Further, the experimental data can be used to draw the variation curve of the fatigue crack propagation rate da / dN and the stress intensity factor range ΔK, i.e. the first variation relationship curve, which can reflect the law of the crack propagation rate changing with the stress intensity factor range.

[0147] At the same time, according to the data of the acoustic emission characteristic parameters (such as ringing count, amplitude, energy count, etc.) changing with the cycle number N recorded in the test, the second variation relationship curve between the acoustic fingerprint signal characteristic parameters and the cycle number N can be drawn, which can show the evolution of the acoustic fingerprint signal characteristic parameters with the cycle number in the fatigue process. For example, some characteristic parameters may change little in the early stage of crack initiation, and their values will gradually increase or show a specific change pattern with the increase of the crack propagation and the cycle number.

[0148] Step S3042, based on the first variation relationship curve and the second variation relationship curve, the Paris formula is used to determine the third correlation relationship between the acoustic fingerprint signal characteristic parameters and the stress intensity factor range, and the fourth correlation relationship between the crack propagation rate and the stress intensity factor range.

[0149] Specifically, according to the first variation relationship curve, the Paris formula can be directly used to determine the fourth correlation relationship between the crack propagation rate and the stress intensity factor range, as shown in the following relationship (3):

[0150]

[0151] In the formula, C and n represent material constants, which can be obtained by fitting the test data.

[0152] Further, according to the second change relationship curve, the cumulative value H of each characteristic parameter (such as the ringing count, amplitude, energy count, etc.) of the acoustic fingerprint signal can be obtained. Meanwhile, the Paris formula can be introduced to determine the third correlation between the acoustic fingerprint signal characteristic parameter and the stress intensity factor range, as shown in the following relationship (4):

[0153]

[0154] In the formula, C1 and n1 represent material constants, which can be obtained by fitting the test data.

[0155] In step S3043, the change relationship between the acoustic fingerprint signal characteristic parameter and the crack propagation rate is determined according to the third correlation and the fourth correlation.

[0156] Specifically, the above relationship (3) and (4) can be taken logarithm and further deduced to obtain the quantitative relationship between the crack propagation rate and the change rate of the cumulative value of the acoustic fingerprint characteristic parameter with the cycle number, as shown in the following relationship (5):

[0157]

[0158] Further, the activity parameter of the acoustic fingerprint signal characteristic parameter with the cycle number reflects the crack propagation rate and the change relationship between the acoustic fingerprint signal characteristic parameter and the crack propagation rate, i.e., the above relationship (5), can be established.

[0159] In step S305, the crack propagation state of the fan component is detected based on the acoustic fingerprint signal characteristic parameter set and the change relationship, and a crack propagation state detection result is obtained.

[0160] Specifically, the above step S305 includes:

[0161] In step S3051, the change of the acoustic fingerprint signal characteristic parameter and a plurality of first crack propagation states are determined according to the acoustic fingerprint signal characteristic parameter set.

[0162] The plurality of first crack propagation states can include crack initiation, crack propagation, and crack instability.

[0163] Specifically, the change of the acoustic fingerprint signal characteristic parameter can be obtained by time domain analysis and frequency domain analysis on the acoustic fingerprint signal characteristic parameter set.

[0164] Firstly, the acoustic fingerprint signal characteristic parameters can be arranged in time sequence to form time series data.

[0165] ​Secondly, the difference between each feature parameter at adjacent time points can be calculated, and then the instantaneous change amount can be observed. For example, for the ring count, by calculating the difference between the two adjacent monitoring, the change range of the ring count can be determined.

[0166] Further, the curve of each feature parameter changing with time can be drawn to intuitively show its change trend, including rising, falling, stable or fluctuating, etc. Further, through the curve, the overall change trend of the feature parameter in the whole monitoring time period can be clearly seen.

[0167] Further, by calculating the statistical indicators of each feature parameter, such as mean, variance, standard deviation, etc., the central tendency and dispersion degree of the feature parameter in the whole data set can be determined. Further, the change rate of the feature parameter can be calculated.

[0168] Further, the correlation coefficient (such as Pearson correlation coefficient) and other methods can be used to analyze the correlation between different feature parameters and determine the association between the feature parameters.

[0169] Finally, by combining the time domain analysis result and the frequency domain analysis result, the change of the voiceprint signal feature parameter can be finally determined.

[0170] Further, the following methods can be used to judge and determine the multiple first crack propagation states:

[0171] (1) If the changes of most voiceprint signal feature parameters are relatively gentle, the change rate is small and close to zero, and the values of the feature parameters fluctuate in a small range, it can be preliminarily judged that the crack is in the initiation stage;

[0172] (2) If some voiceprint signal feature parameters show a relatively stable linear or approximately linear change trend, the change rate remains near a relatively stable value, and the cooperative change between different feature parameters is relatively regular, it can be preliminarily judged that the crack is in the expansion stage.

[0173] (3) If the change of some voiceprint signal feature parameters becomes violent, such as the change rate suddenly increases greatly, the value of the feature parameter increases sharply or the fluctuation range increases significantly, it can be preliminarily judged that the crack is in the unstable state.

[0174] Step S3052, based on the change and the change relationship, determine multiple second crack propagation states.

[0175] Among them, the multiple second crack propagation states can include three main states of crack initiation, crack expansion and crack instability.

[0176] Specifically, according to the change of the obtained voiceprint signal feature parameters, and in combination with the change relationship between the voiceprint signal feature parameters and the crack propagation rate, the change of the crack propagation rate corresponding to the change of each voiceprint signal feature parameter can be further determined.

[0177] Further, the state of the crack, i.e., the plurality of second crack propagation states, can be determined according to the change of the crack propagation rate.

[0178] In step S3053, the crack propagation state detection result is determined according to the plurality of first crack propagation states and the plurality of second crack propagation states.

[0179] Specifically, by comprehensively considering the plurality of first crack propagation states and the plurality of second crack propagation states, the final crack propagation state detection result can be further determined.

[0180] For example, when the first crack propagation state is completely consistent with the second crack propagation state and the characteristics of the state are very significant, the final crack propagation state detection result can be directly determined as the state. Through comprehensive analysis, a comprehensive and accurate crack propagation state detection result can be obtained, and the accuracy and reliability of the detection result are improved.

[0181] The crack propagation state detection method provided in this embodiment can directly obtain the actual crack propagation data and the change of the corresponding voiceprint signal feature parameters with the cycle number by carrying out a fatigue crack propagation voiceprint signal test. Further, the Paris formula can be used to further study the internal relationship between the voiceprint signal feature parameters and the stress intensity factor range, and more accurately determine the quantitative relationship between the crack propagation rate and the stress intensity factor range. Finally, according to the determined third and fourth correlation relationships, the change relationship between the voiceprint signal feature parameters and the crack propagation rate can be finally determined. From the perspective of combining theory and experiment, the relationship between the voiceprint signal feature parameters and the crack propagation state is quantified. Further, by analyzing the voiceprint signal feature parameter set, the change trend of the voiceprint signal feature parameters can be obtained, and the plurality of first crack propagation states can be preliminarily identified and detected. Further, in combination with the change of the voiceprint signal feature parameters and the change relationship between the voiceprint signal feature parameters and the crack propagation rate, the change of the voiceprint signal can be related to the actual crack propagation state, and the scientificity and accuracy of the crack propagation state judgment are improved. Compared with the traditional single judgment mode, the characteristics of the crack in different stages can be more comprehensively and accurately identified. Finally, by comprehensively considering the plurality of second crack propagation states and the plurality of first crack propagation states determined in the previous steps, a comprehensive and accurate crack propagation state detection result can be obtained, and the accuracy and reliability of the detection result are improved.

[0182] In an example, a non-destructive testing optimization combination method for monitoring the crack propagation state of a key component of a wind turbine is provided. A new method of acoustic emission signal denoising, separation, feature extraction and pattern recognition is proposed from the time domain and frequency domain of acoustic emission signal. From the perspective of characteristic parameters of acoustic emission signal, the correlation between the change of characteristic parameters and the change of crack stress intensity factor is established. Combined with the above two methods, the change of acoustic emission signal is analyzed, and the corresponding relationship between acoustic emission signal and crack propagation state is obtained, which improves the accuracy and reliability of acoustic emission signal monitoring crack propagation state. Specifically, it includes:

[0183] Step one: signal denoising.

[0184] Specifically, the working environment of the fan is complex, the acoustic emission signals of various types are strong, and the background noise is strong. When evaluating the crack state using acoustic emission signals, it will be disturbed by other types of acoustic emission signals and noise. The BO-WT-VMD combination algorithm is used for signal denoising, which can effectively improve the signal-to-noise ratio and solve the interference of noise.

[0185] Firstly, the wavelet transform is used to determine the wavelet layer number, and the original signal is denoised.

[0186] Secondly, the parameters K and a of VMD are determined by Bayesian optimization, that is, firstly, the hyperparameter combination and the prior data set are randomly initialized, the sampling points are increased through the Gaussian model GP, the posterior data set is obtained, if the posterior data set meets the termination condition, the optimized parameter combination is determined, and the parameters of wavelet transform are obtained; if the posterior data set does not meet the termination condition, the loop Gaussian model and the previous steps are used until the posterior data set meets the termination condition, the optimized parameter combination is determined, and the parameters K and a of VMD are obtained.

[0187] Finally, the values of the parameters K and a are substituted into the VMD algorithm to obtain the decomposed K modes, and then the optimal mode is selected by the Pearson coefficient, and the remaining modes are removed, and finally the denoised reconstructed data is obtained.

[0188] Step two: signal blind separation.

[0189] Specifically, when the key structural part of the wind turbine produces cracks and breaks, acoustic emission signals such as impact and friction are often generated, and the crack propagation signal is accurately separated from the mixed signal for detailed analysis, which is of great significance for accurately judging the crack propagation state in the fan. The NLPCA algorithm is used to separate the denoised signal, which can effectively obtain the required signal.

[0190] Firstly, the denoised signal s(t) is linearly mixed to obtain x(t) through the time-varying system The separation process is simplified to solve the time-varying system generalized inverse matrix The process is shown in the following relation (6):

[0191]

[0192] Secondly, the objective function is constructed according to the minimum mean square error approximation criterion as the independence criterion, as shown in the above relation (1).

[0193] Finally, the gradient descent algorithm is used to solve the problem, as shown in equation (2) above.

[0194] Step 3: Signal Feature Recognition. From the time and frequency domain perspectives of acoustic emission signals, an effective method for pattern recognition of acoustic emission signals during crack propagation is established. The Bird Flock Optimization-One-Dimensional Convolutional Neural Network (BS-1DCNN) algorithm is used for feature recognition.

[0195] Specifically, in the BS-1DCNN fault diagnosis model, the BS-1DCNN algorithm utilizes the complementary characteristics of 1DCNN and SVM, combining the advantages of both while improving their respective limitations. The bird flocking algorithm is used to optimize relevant parameters in the SVM, effectively solving the local optima problem and improving classification accuracy. Finally, a bird flocking optimized support vector machine classifier is used for refined identification of the extracted fault features.

[0196] The main diagnostic process is as follows:

[0197] First, the acoustic signals of the turbine cracks that appeared in this experiment were sorted and segmented, including acoustic signals from laboratory turbine material fatigue experiments and acoustic signals from model turbine experiments under different water conditions, forming a data sample set of fixed segments.

[0198] Secondly, the data sample set is input into the 1CNN convolutional layer, and the fault features are extracted adaptively using the convolutional kernel.

[0199] Furthermore, after passing through the convolutional layer, the extracted features are output by the fully connected layer and fed into the output layer.

[0200] Furthermore, the bird flocking algorithm is used to optimize the values ​​of the penalty parameter c and the radial basis kernel parameter g in the SVM model.

[0201] Finally, a bird flock-optimized SVM classifier is used to replace the conventional softmax classifier in the output layer of the 1DCNN to further improve classification accuracy and complete the fault diagnosis process.

[0202] Step 4: Construct a numerical model and establish the correspondence between the changes in acoustic emission signal characteristic parameters and the crack propagation state.

[0203] Specifically, the crack propagation rate da / dN is reflected by the activity parameter dH / dN of the acoustic emission signal characteristic parameter with the cycle number, and a corresponding relationship between the acoustic emission signal spectrum characteristics and the macroscopic crack propagation state is constructed.

[0204] Firstly, the conventional metal material for the fan is selected, the acoustic fingerprint signal test of the metal material fatigue crack propagation is carried out, the change relationship curves of the fatigue crack length a, the stress intensity factor range ΔK with the cycle number N, the change relationship curves of the fatigue crack propagation rate da / dN and ΔK, and the acoustic emission characteristic parameter with the cycle number N are obtained.

[0205] Secondly, the Paris formula is cited, and the relationship between the acoustic fingerprint signal characteristic parameter and ΔK is established, as shown in the above relationship formula (4).

[0206] Further, it is known from the Paris formula that the relationship between the crack propagation rate and ΔK is shown in the above relationship formula (3).

[0207] Further, the above relationship formula (3) and (4) are taken logarithm and further deduced, and the quantitative relationship between the change rate of the crack propagation rate and the cumulative value of the acoustic fingerprint characteristic parameter with the cycle number is obtained, as shown in the above relationship formula (5).

[0208] Finally, the activity parameter dH / dN of the acoustic fingerprint signal characteristic parameter with the cycle number reflects the crack propagation rate da / dN. Further, the relationship expression between the acoustic fingerprint signal characteristic parameter and the crack propagation rate can be established.

[0209] Step five: comprehensive analysis of the crack propagation state.

[0210] Specifically, through steps one to three, the crack propagation state can be identified by using the intelligent algorithm, and finally three main state labels of crack initiation, crack propagation and crack instability can be obtained. Through step four, the relationship between the acoustic emission signal characteristic parameter change and the crack propagation rate is analyzed, the crack propagation rate in a specific period can be obtained, the crack state corresponding to the crack propagation rate curve can be obtained, and three main states of crack initiation, crack stable propagation and crack unstable propagation are output. Corresponding to the labels obtained by the intelligent algorithm, the crack propagation state can be judged together, and the accuracy of the crack propagation state identification is improved.

[0211] The nondestructive testing optimization combination method for monitoring the crack propagation state of the key components of the fan provided in the present example has the following effects:

[0212] 1. This example proposes a novel signal denoising, blind signal separation, and signal feature recognition method. Using the new method, a complete set of processing for acoustic emission signals—signal denoising, separation, and recognition—is performed, which can effectively analyze acoustic emission signals from the time and frequency domains to determine the state of the crack.

[0213] 2. This example proposes a method to establish the relationship between acoustic emission signal characteristic parameters and crack propagation state, which can realize the judgment of crack propagation state from the perspective of acoustic emission signal parameter analysis.

[0214] 3. This example proposes to combine time-domain and frequency-domain methods with parametric methods, and use intelligent algorithms and crack propagation state relationships to jointly determine the crack state, rather than using a single method for crack propagation state assessment, which can improve the accuracy of crack propagation state determination.

[0215] Therefore, the innovative crack propagation state identification method proposed in this example can ensure the accuracy of crack propagation state identification in high-noise environments, reduce the false alarm rate, and improve the accuracy of wind turbine crack monitoring.

[0216] This embodiment also provides a crack propagation state detection device, which is used to implement the above embodiments and preferred embodiments, and will not be repeated as already described. As used below, the term "module" can be a combination of software and / or hardware that implements a predetermined function. Although the device described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0217] This embodiment provides a crack propagation state detection device for wind turbine components; such as Figure 4 As shown, the device includes:

[0218] The acquisition module 401 is used to acquire the initial acoustic emission signal set of the wind turbine.

[0219] The first processing module 402 is used to perform noise reduction and blind separation processing on the initial acoustic emission signal set to obtain the crack propagation signal set of the wind turbine component.

[0220] The second processing module 403 is used to perform feature classification processing on the crack propagation signal set to obtain a set of acoustic signal feature parameters reflecting the crack propagation state.

[0221] Module 404 is established to establish the relationship between the characteristic parameters of the acoustic signal and the crack propagation rate using the fatigue crack propagation acoustic signal test method and the Paris formula.

[0222] The detection module 405 is used to detect the crack propagation state of the fan component based on the characteristic parameter set and variation relationship of the acoustic signature signal, and obtain the crack propagation state detection result.

[0223] In some alternative implementations, the first processing module 402 includes:

[0224] The first processing submodule is used to perform noise reduction processing on the initial acoustic emission signal set using a preset signal noise reduction combination algorithm to obtain the target acoustic emission signal set.

[0225] The separation submodule is used to perform blind separation of the target acoustic emission signal set using a nonlinear principal component analysis algorithm to obtain the crack propagation signal set of the wind turbine component.

[0226] In some alternative implementations, the separation submodule includes:

[0227] The mixing unit is used to perform time-varying linear mixing on the target acoustic emission signal set to obtain a mixed acoustic emission signal set.

[0228] The construction unit is used to construct the objective function using the minimum mean square error approximation criterion.

[0229] The processing unit is used to obtain the crack propagation signal set based on the mixed acoustic emission signal set and the objective function through the gradient descent algorithm.

[0230] In some alternative implementations, the second processing module 403 includes:

[0231] The second processing submodule is used to process the crack propagation signal set through a one-dimensional convolutional neural network algorithm to obtain the fault feature set.

[0232] The third processing submodule is used to process the fault feature set through the bird flocking algorithm to obtain the acoustic signature signal feature parameter set reflecting the crack propagation state.

[0233] In some alternative implementations, the establishment module 404 includes:

[0234] A submodule is established to use the fatigue crack propagation acoustic signal test method to establish the first variation curve between crack propagation rate and stress intensity factor range, and the second variation curve between acoustic signal characteristic parameters and cycle number.

[0235] The first determining submodule is used to determine, based on the first and second variation relationship curves, the third correlation between acoustic signature signal characteristic parameters and stress intensity factor range, and the fourth correlation between crack propagation rate and stress intensity factor range using the Paris formula.

[0236] The second determination submodule is used to determine the relationship between the acoustic signature signal characteristic parameters and the crack propagation rate based on the third and fourth correlation relationships.

[0237] In some alternative implementations, the detection module 405 includes:

[0238] The third determining submodule is used to determine the changes in the acoustic signature signal characteristic parameters and multiple first crack propagation states based on the acoustic signature signal characteristic parameter set.

[0239] The fourth determination submodule is used to determine multiple second crack propagation states based on changes and relationships.

[0240] The fifth determination submodule is used to determine the crack propagation state detection result based on multiple first crack propagation states and multiple second crack propagation states.

[0241] Further functional descriptions of the above modules and units are the same as those in the corresponding embodiments described above, and will not be repeated here.

[0242] The crack propagation state detection device in this embodiment is presented in the form of a functional unit. Here, a unit refers to an ASIC (Application Specific Integrated Circuit) circuit, a processor and memory that execute one or more software or fixed programs, and / or other devices that can provide the above functions.

[0243] This invention also provides a computer device having the above-described features. Figure 4 The crack propagation state detection device shown.

[0244] Please see Figure 5 , Figure 5 This is a schematic diagram of the structure of a computer device provided in an optional embodiment of the present invention, such as... Figure 5 As shown, the computer device includes one or more processors 10, memory 20, and interfaces for connecting the components, including high-speed interfaces and low-speed interfaces. The components communicate with each other via different buses and can be mounted on a common motherboard or otherwise installed as needed. The processors can process instructions executed within the computer device, including instructions stored in or on memory to display graphical information of a GUI on external input / output devices (such as display devices coupled to the interfaces). In some alternative implementations, multiple processors and / or multiple buses can be used with multiple memories, if desired. Similarly, multiple computer devices can be connected, each providing some of the necessary operations (e.g., as a server array, a group of blade servers, or a multiprocessor system). Figure 5 Take a processor 10 as an example.

[0245] Processor 10 may be a central processing unit, a network processor, or a combination thereof. Processor 10 may further include a hardware chip. The hardware chip may be an application-specific integrated circuit (ASIC), a programmable logic device (PLD), or a combination thereof. The programmable logic device may be a complex programmable logic device (CAMP), a field-programmable gate array (FPGA), a general-purpose array logic (GDA), or any combination thereof.

[0246] The memory 20 stores instructions executable by at least one processor 10 to cause at least one processor 10 to perform the method shown in the above embodiments.

[0247] The memory 20 may include a program storage area and a data storage area. The program storage area may store the operating system and applications required for at least one function; the data storage area may store data created based on the use of the computer device. Furthermore, the memory 20 may include high-speed random access memory and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some alternative embodiments, the memory 20 may optionally include memory remotely located relative to the processor 10, and these remote memories may be connected to the computer device via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0248] The memory 20 may include volatile memory, such as random access memory; the memory may also include non-volatile memory, such as flash memory, hard disk or solid-state drive; the memory 20 may also include a combination of the above types of memory.

[0249] The computer device also includes a communication interface 30 for communicating with other devices or communication networks.

[0250] This invention also provides a computer-readable storage medium. The methods described above according to embodiments of the invention can be implemented in hardware or firmware, or implemented as computer code that can be recorded on a storage medium, or implemented as computer code downloaded via a network and originally stored on a remote storage medium or a non-transitory machine-readable storage medium and then stored on a local storage medium. Thus, the methods described herein can be processed by software stored on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. The storage medium can be a magnetic disk, optical disk, read-only memory, random access memory, flash memory, hard disk, or solid-state drive, etc.; further, the storage medium can also include combinations of the above types of memory. It is understood that computers, processors, microprocessor controllers, or programmable hardware include storage components capable of storing or receiving software or computer code, which, when accessed and executed by the computer, processor, or hardware, implements the methods shown in the above embodiments.

[0251] A portion of this invention can be applied as a computer program product, such as computer program instructions, which, when executed by a computer, can invoke or provide the methods and / or technical solutions according to the invention through the operation of the computer. Those skilled in the art will understand that the forms in which computer program instructions exist in a computer-readable medium include, but are not limited to, source files, executable files, installation package files, etc. Correspondingly, the ways in which computer program instructions are executed by a computer include, but are not limited to: the computer directly executing the instructions, or the computer compiling the instructions and then executing the corresponding compiled program, or the computer reading and executing the instructions, or the computer reading and installing the instructions and then executing the corresponding installed program. Here, the computer-readable medium can be any available computer-readable storage medium or communication medium accessible to a computer.

[0252] Although embodiments of the invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the invention, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. A crack propagation state detection method for a fan component, characterized by, The method comprises: acquiring an initial acoustic emission signal set of a fan; performing noise reduction and blind separation processing on the initial acoustic emission signal set to obtain a crack propagation signal set of a fan component; performing feature classification processing on the crack propagation signal set to obtain an acoustic fingerprint signal feature parameter set reflecting a crack propagation state; establishing a change relationship between the acoustic fingerprint signal feature parameter and the crack propagation rate by using a fatigue crack propagation acoustic fingerprint signal test method and a Paris formula; detecting the crack propagation state of the fan component based on the acoustic fingerprint signal feature parameter set and the change relationship to obtain a crack propagation state detection result; performing feature classification processing on the crack propagation signal set to obtain an acoustic fingerprint signal feature parameter set reflecting a crack propagation state, comprising: processing the crack propagation signal set through a one-dimensional convolutional neural network algorithm to obtain a fault feature set; processing the fault feature set through a bird swarm algorithm to obtain the acoustic fingerprint signal feature parameter set reflecting the crack propagation state; wherein the change relationship between the acoustic fingerprint signal feature parameter and the crack propagation rate is established by using the fatigue crack propagation acoustic fingerprint signal test method and the Paris formula, comprising: establishing a first change relationship curve between the crack propagation rate and the stress intensity factor range, and a second change relationship curve between the acoustic fingerprint signal feature parameter and the cycle number by using the fatigue crack propagation acoustic fingerprint signal test method; determining a third correlation relationship between the acoustic fingerprint signal feature parameter and the stress intensity factor range, and a fourth correlation relationship between the crack propagation rate and the stress intensity factor range based on the first change relationship curve and the second change relationship curve by using the Paris formula; determining the change relationship between the acoustic fingerprint signal feature parameter and the crack propagation rate according to the third correlation relationship and the fourth correlation relationship, wherein the change relationship is represented by the following relationship: In the formula: represents the crack propagation rate; , , , represents a material constant, obtained by fitting test data; the activity parameter of the voiceprint signal feature parameter with the cycle number; represents the fatigue crack length; represents the cycle number; represents the cumulative value of each feature parameter of the voiceprint signal; wherein the crack propagation state of the fan component is detected based on the acoustic fingerprint signal feature parameter set and the change relationship to obtain a crack propagation state detection result, comprising: determining the change of the acoustic fingerprint signal feature parameter and a plurality of first crack propagation states according to the acoustic fingerprint signal feature parameter set, wherein the change of the acoustic fingerprint signal feature parameter is obtained by time domain analysis and frequency domain analysis on the acoustic fingerprint signal feature parameter set; determining a plurality of second crack propagation states based on the change and the change relationship, wherein the change of the crack propagation rate corresponding to each acoustic fingerprint signal feature parameter is determined according to the change of the acoustic fingerprint signal feature parameter obtained, combined with the change relationship between the acoustic fingerprint signal feature parameter and the crack propagation rate, and the plurality of second crack propagation states are determined according to the change of the crack propagation rate; determining the crack propagation state detection result according to the plurality of first crack propagation states and the plurality of second crack propagation states.

2. The method of claim 1, wherein, performing noise reduction and blind separation processing on the initial acoustic emission signal set to obtain a crack propagation signal set of a fan component, comprising: The initial acoustic emission signal set is denoised by using a preset signal denoising combination algorithm to obtain a target acoustic emission signal set, and the preset signal denoising combination algorithm is a Bayesian optimization-wavelet threshold-variational modal decomposition algorithm; The target acoustic emission signal set is blindly separated by using a nonlinear principal component analysis algorithm to obtain the crack propagation signal set of the fan component; The initial acoustic emission signal set is denoised by using a preset signal denoising combination algorithm to obtain a target acoustic emission signal set, and the preset signal denoising combination algorithm is a Bayesian optimization-wavelet threshold-variational modal decomposition algorithm; 3. The method of claim 2, wherein, The target acoustic emission signal set is blindly separated by using a nonlinear principal component analysis algorithm to obtain the crack propagation signal set of the fan component, including: The target acoustic emission signal set is time-varying linearly mixed to obtain a mixed acoustic emission signal set; A target function is constructed by using a minimum mean square error approximation criterion; Based on the mixed acoustic emission signal set and the target function, the crack propagation signal set is obtained through a gradient descent algorithm.

4. A crack propagation state detection device for a fan component, characterized by, The device comprises: An acquisition module is configured to acquire an initial acoustic emission signal set of a fan; A first processing module is configured to perform denoising and blind separation processing on the initial acoustic emission signal set to obtain a crack propagation signal set of a fan component; A second processing module is configured to perform feature classification processing on the crack propagation signal set to obtain an acoustic print signal feature parameter set reflecting a crack propagation state; An establishment module is configured to establish a change relationship between an acoustic print signal feature parameter and a crack propagation rate by using a fatigue crack propagation acoustic print signal test method and a Paris formula; A detection module is configured to detect a crack propagation state of the fan component based on the acoustic print signal feature parameter set and the change relationship to obtain a crack propagation state detection result; The second processing module comprises: A second processing submodule is configured to perform one-dimensional convolutional neural network algorithm processing on the crack propagation signal set to obtain a fault feature set; A third processing submodule is configured to perform bird swarm algorithm processing on the fault feature set to obtain the acoustic print signal feature parameter set reflecting the crack propagation state; The establishment module comprises: An establishment submodule is configured to establish a first change relationship curve between the crack propagation rate and a stress intensity factor range and a second change relationship curve between the acoustic print signal feature parameter and a cycle number by using the fatigue crack propagation acoustic print signal test method; A first determination submodule is configured to determine a third correlation relationship between the acoustic print signal feature parameter and the stress intensity factor range and a fourth correlation relationship between the crack propagation rate and the stress intensity factor range by using the Paris formula based on the first change relationship curve and the second change relationship curve. A second determining sub-module is configured to determine the change relationship between the voiceprint signal feature parameter and the crack propagation rate according to the third correlation relationship and the fourth correlation relationship, where the change relationship is represented by a relationship formula as follows: In the formula: represents the crack propagation rate; , , , represents a material constant, obtained by fitting test data; the activity parameter of the voiceprint signal feature parameter with the cycle number; represents the fatigue crack length; represents the cycle number; represents the cumulative value of each feature parameter of the voiceprint signal; The detection module comprises: A third determining sub-module is configured to determine a change of the voiceprint signal feature parameter and a plurality of first crack propagation states according to the set of voiceprint signal feature parameters, where the change of the voiceprint signal feature parameter is obtained by time domain analysis and frequency domain analysis on the set of voiceprint signal feature parameters; A fourth determining sub-module is configured to determine a plurality of second crack propagation states based on the change and the change relationship, where the change of the crack propagation rate corresponding to each voiceprint signal feature parameter is determined according to the obtained change of the voiceprint signal feature parameter and the change relationship between the voiceprint signal feature parameter and the crack propagation rate, and the plurality of second crack propagation states are determined according to the change of the crack propagation rate; A fifth determining sub-module is configured to determine the crack propagation state detection result according to the plurality of first crack propagation states and the plurality of second crack propagation states.

5. A computer device, comprising: Comprise: A memory and a processor, which are in communication connection with each other, and the memory stores computer instructions, and the processor executes the computer instructions to perform the crack propagation state detection method in any one of claims 1 to 3.

6. A computer-readable storage medium, characterized in that, The computer readable storage medium stores computer instructions, and the computer instructions are used to make the computer execute the crack propagation state detection method in any one of claims 1 to 3.

7. A computer program product, characterised in that, The computer instructions are used to make the computer execute the crack propagation state detection method in any one of claims 1 to 3.

Images