Wheel disc structure fatigue crack prediction method based on acoustic emission signal and in-situ image

Abstract

The invention discloses a wheel disc structure fatigue crack prediction method based on acoustic emission signals and in-situ images. The method comprises the following steps: acquiring acoustic emission signals and crack in-situ images of a wheel disc structure under a plurality of load cycle periods; fitting is carried out according to the corresponding acoustic emission signals in the multiple load cycle periods, and a first fitting function is generated; determining a first material parameter according to the first fitting function and a Forman equation; performing fitting according to the corresponding crack in-situ images under the plurality of load cycle periods to generate a second fitting function; determining a second material parameter according to the second fitting function in combination with a Forman equation; predicting the crack length according to the first material parameter and the second material parameter; determining the residual life according to a preset crack length and a predicted value of the crack length; according to the method, the material parameters can be accurately determined by combining the dual-path fitting of the acoustic emission signal and the in-situ image with the Forman equation, and a reliable basis is provided for crack length prediction, so that the prediction precision is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of fault diagnosis and prediction, and more particularly to a method for predicting fatigue cracks of a wheel disc structure based on acoustic emission signals and in-situ images. Background Art

[0002] As a key component of a gas turbine engine, the safety and reliability of the disc are of vital importance, which is mainly reflected in the complex working load, complex structural form, and serious consequences of rupture. Since disc failure may lead to catastrophic consequences for the engine, the airworthiness regulations of civil aviation gas turbine engines and military standards of gas turbine engines in various countries in the world manage discs as "life-limited parts" and "critical parts", respectively, and their health management theories and methods are particularly important. Therefore, it is urgent to carry out research on real-time monitoring methods for disc crack extension and establish a monitoring method for disc crack implementation to provide support for model disc health monitoring and damage tolerance design.

[0003] Acoustic emission technology has significant advantages over other nondestructive testing methods. It can monitor tiny damage in real time, provide early warning, and does not require destructive operations. Common AE damage identification methods include signal time domain, frequency domain, time-frequency analysis, waveform feature analysis, pattern recognition and positioning technology, which can effectively identify damage types such as cracks and corrosion. However, these methods have several limitations: signal quality and noise interference affect prediction accuracy, and signal feature extraction is complex; traditional methods have low accuracy when dealing with nonlinear and complex damage evolution; machine learning and deep learning rely on a large amount of high-quality data, and face the problems of overfitting and insufficient generalization ability.

[0004] Therefore, how to improve prediction accuracy is an urgent problem that technicians in this field need to solve. Summary of the invention

[0005] In view of this, the present invention provides a fatigue crack prediction method for a wheel disc structure based on acoustic emission signals and in-situ images, which can accurately determine material parameters through dual-path fitting of acoustic emission signals and in-situ images combined with the Forman equation, providing a reliable basis for crack length prediction, thereby improving prediction accuracy.

[0006] In order to achieve the above object, the present invention adopts the following technical solution:

[0007] A method for predicting fatigue cracks of a wheel disc structure based on acoustic emission signals and in-situ images comprises the following steps:

[0008] Acquire the acoustic emission signals and in-situ crack images of the disk structure under multiple load cycles;

[0009] Perform fitting according to the acoustic emission signals corresponding to a plurality of load cycles to generate a first fitting function; determine a first material parameter according to the first fitting function and the Forman equation;

[0010] Perform fitting according to the corresponding in-situ crack images under multiple load cycles to generate a second fitting function; determine the second material parameter according to the second fitting function combined with the Forman equation;

[0011] The crack length is predicted according to the first material parameter and the second material parameter; and the remaining life is confirmed according to the preset crack length and the predicted value of the crack length.

[0012] Preferably, the step further includes: confirming the remaining life according to a preset crack length and a predicted value of the crack length.

[0013] Preferably, generating the first fitting function specifically includes:

[0014] Extract characteristic parameters according to the acoustic emission signal and calculate the cumulative energy value to obtain a series of first associated data points consisting of cumulative energy values ​​and cycle times;

[0015] The first associated data points are fitted and derived to obtain an energy release rate curve, that is, the first fitting function.

[0016] Preferably, determining the first material parameter according to the first fitting function and the Forman equation specifically includes:

[0017] According to the correlation between energy release rate and crack growth rate and the Forman equation, the correlation between the energy release rate and the stress intensity range is confirmed:

[0018]

[0019] Where U is the cumulative released energy, N is the number of load cycles, is the energy release rate; c 1 and m 1 is the first material parameter, ΔK is the stress intensity range; R is the load ratio, K c is the fracture toughness;

[0020] The energy release rate is confirmed by the first fitting function and linearly fitted with the corresponding stress intensity range to obtain the first material parameter.

[0021] Preferably, generating the second fitting function specifically includes:

[0022] Acquire the crack length according to the in-situ image to obtain a series of second associated data points consisting of the crack length and the number of cycles;

[0023] The second associated data points are fitted and differentiated to obtain a crack length extension rate curve, that is, the second fitting function.

[0024] Preferably, determining the second material parameter according to the second fitting function in combination with the Forman equation specifically includes:

[0025] The Forman equation is:

[0026] Where a is the crack length, N is the number of load cycles, is the crack length growth rate, c 2 and m 2 is the second material parameter;

[0027] The crack length extension rate is confirmed according to the second fitting function, and a linear fitting is performed in combination with a corresponding stress intensity range to confirm the second material parameter.

[0028] Preferably, the predicted crack length specifically includes:

[0029] Obtaining a prediction target, wherein the prediction target is a target number of cycles;

[0030] Confirming a current roulette state, the current roulette state including a current crack length and a current number of cycles; predicting a crack length in a next cycle based on the current roulette state, and iterating until the target number of cycles is reached;

[0031] The predicting of the crack length in the next cycle specifically includes:

[0032] The crack growth rate of the next cycle is determined according to the first material parameter and the second material parameter:

[0033]

[0034] The crack length of the next cycle is obtained based on the crack growth rate of the next cycle and the current crack length.

[0035] Preferably, the confirming of the remaining life specifically includes: setting a crack length threshold, and when the crack length is predicted to exceed the crack length threshold, outputting the current number of cycles to obtain the remaining life.

[0036] Preferably, the steps further include: clustering according to the acoustic emission signals, and associating the clustering results with multiple crack extension stages to obtain multiple damage state labels; and comparing the damage state labels with the actual damage to confirm the damage state of the actual damage.

[0037] A wheel disc structure fatigue crack prediction system based on acoustic emission signals and in-situ images includes an acoustic emission monitoring subsystem for acquiring acoustic emission signals and in-situ crack images of a target structure, and includes an acoustic emission sensor, a piezoelectric chip, an image acquisition device and a controller.

[0038] The acoustic emission sensor, the piezoelectric and the image acquisition device are electrically connected to the controller; the acoustic emission sensor is used to monitor the acoustic emission vibration signal; the piezoelectric chip is used to convert the acoustic emission vibration signal into an electrical signal and send it to the controller; the image acquisition device is used to capture the in-situ image of the wheel disc structure and send it to the controller; the controller is used to predict the crack length of the wheel disc and confirm the remaining life according to the acoustic emission signal and the in-situ image.

[0039] It can be seen from the above technical solutions that, compared with the prior art, the present invention discloses a method for predicting fatigue cracks in a wheel structure based on acoustic emission signals and in-situ images, which can accurately determine material parameters through dual-path fitting of acoustic emission signals and in-situ images combined with the Forman equation, and provide a reliable basis for crack length prediction, thereby improving prediction accuracy; the CLCA+PCA+K-means algorithm is used to reduce the dimension and cluster the AE signal, and the clusters of the clusters are associated with the three stages of crack extension, and the physical mechanism of the crack extension process is mapped to the AE signal, which can more clearly identify the damage mechanism. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0041] Figure 1 A schematic diagram of a method for predicting fatigue cracks in a wheel disc structure based on acoustic emission signals and in-situ images provided by the present invention.

[0042] Figure 2 Schematic diagram of the experimental results for verifying the correlation between AE cumulative energy release rate and crack growth rate in an embodiment of the present invention.

[0043] Figure 3 Schematic diagram of the accuracy experimental effect of life prediction in an embodiment of the present invention.

[0044] Figure 4 It is a hierarchical tree diagram of CLCA in an embodiment of the present invention.

[0045] Figure 5Schematic diagram of the effect of K-means clustering mapped to the original feature space in an embodiment of the present invention.

[0046] Figure 6 Schematic diagram of the correlation effect between the cumulative energy of each cluster of K-means clustering and the crack propagation process in an embodiment of the present invention.

[0047] Figure 7 A schematic structural diagram of a wheel disk structure fatigue crack prediction system based on acoustic emission signals and in-situ images provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0048] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0049] Example 1

[0050] like Figure 1 The embodiment of the present invention discloses a method for predicting fatigue cracks of a wheel disc structure based on acoustic emission signals and in-situ images, comprising the following steps:

[0051] S1: Acquire the acoustic emission signals and in-situ crack images of the disc structure under multiple load cycles;

[0052] S2: performing fitting according to the crack in-situ images corresponding to a plurality of load cycles to generate a first fitting function; and determining a first material parameter according to the first fitting function in combination with the Forman equation;

[0053] S3: performing fitting according to the acoustic emission signals corresponding to a plurality of load cycles to generate a second fitting function; and confirming a second material parameter according to the second fitting function and the Forman equation;

[0054] S4: Predicting a crack length according to the first material parameter and the second material parameter.

[0055] In this embodiment, the change characteristics of the acoustic emission signal can be obtained through the first fitting function, and the correlation between the acoustic emission signal and the stress intensity range can be further clarified in combination with the Forman equation, so that the fitting of the first material parameter can be achieved through the correlation of multiple data points; the change characteristics of the crack in the in-situ image are obtained through the second fitting function, and the correlation between the crack characteristics and the stress intensity range can be further clarified in combination with the Forman equation, so that the fitting of the second material parameter can be achieved through the correlation of multiple data points. The first material parameter and the second material parameter are obtained from two paths, namely, the cumulative energy of the signal and the crack length. According to the combination of the first material parameter and the second material parameter, the two characteristics can be integrated to achieve accurate prediction.

[0056] In order to further implement the above technical solution, the steps also include: confirming the remaining life according to the preset crack length and the predicted value of the crack length.

[0057] In one embodiment, the generation process of the first fitting function includes:

[0058] Characteristic parameters are extracted from the acoustic emission signal and the cumulative energy value is calculated to obtain a series of first associated data points consisting of cumulative energy values ​​and cycle numbers; the first associated data points are fitted and derived to obtain an energy release rate curve, that is, the above-mentioned first fitting function.

[0059] Furthermore, by combining the first fitting function with the Forman equation, the expression of the stress intensity range and the first material parameter for the first fitting function can be derived, so that the first fitting function can be further fitted with the stress intensity range to obtain the first material parameter.

[0060] Specifically, it is known that the energy release rate and crack growth rate in the acoustic emission process of fatigue crack growth have the following relationship:

[0061]

[0062] in, The energy release rate based on the acoustic emission signal fitting is the first fitting function value. Specifically, the cumulative energy data is obtained according to the accumulation of AE energy parameters, and the energy change rate is obtained by the difference method. is the crack growth rate, i.e., the second fitting function value. Specifically, the crack change rate is obtained by differential processing of the MTS crack growth data; a is the crack length, and N is the number of load cycles. B represents the specific material proportional constant. ΔK is the stress intensity range. Under plane stress conditions, E' is the Young's modulus E; under plane strain conditions, E'=E / (1-υ 2 ). υ is Poisson's ratio, R is the load ratio, and t is the specimen thickness.

[0063] The Forman equation can be used to describe the crack growth behavior:

[0064]

[0065] Among them, c 2 and m 2 is the second material constant; K c is the fracture toughness. The test material is Ti alloy, K c =60Mpa*m 0.5 ; ΔK is the stress intensity factor range.

[0066] Through equations (1) and (2), it can be deduced that the correlation between the energy release rate and the range of stress intensity factor is:

[0067]

[0068] Among them, c 1 and m 1 is the first material constant.

[0069] In formula (3), the load ratio and fracture toughness are known quantities, and the range of the stress intensity factor can be obtained based on the known quantities and in-situ image analysis. Therefore, when the first function value is confirmed, the fitting of the first material parameter can be achieved.

[0070] The fitting function is: 1 =ln(c 1 )+m 1 x 1 +b

[0071] Among them, the fitting function value y 1 is the logarithm of the energy release rate, the independent variable x 1 is the logarithm of the stress intensity range; b is the bias term constant.

[0072] In this embodiment, the correlation between the energy release rate and the crack growth rate is verified. Figure 6 To verify the experimental data, Figure 6 It can be seen that the two are in a logarithmic linear relationship:

[0073]

[0074] Where a is the crack length, N is the number of cycles, c is the accumulated signal of acoustic emission, and m and p are material constants that depend on the material and load conditions. Most of the data points are within the 95% prediction interval of all acoustic emission parameters. At the same time, different frequencies were used during the experiment, which can prove that the quantitative relationship between the rate of change of acoustic emission parameters and the crack growth rate is independent of frequency.

[0075] In order to further implement the above technical solution, the fitting step of the second fitting function includes:

[0076] The crack length is obtained according to the in-situ image to obtain a series of second associated data points consisting of crack length and cycle number; the second associated data points are fitted and derived to obtain a crack length extension rate curve, that is, a second fitting function.

[0077] Furthermore, by combining the second fitting function with the Forman equation, the range of the stress intensity factor and the second material expression of the second fitting function can be deduced, so that the first fitting function can be further fitted with the stress intensity range to obtain the second material parameters.

[0078] Specifically, according to the Forman equation, the following fitting function can be obtained:

[0079] y 2 =ln(c 2 )+m 2 x 2 +b

[0080] Among them, the fitting function value y 2 is the logarithm of the crack growth rate. Multiple sets of data can be obtained through the second fitting function. 2 is the logarithm of the stress intensity range; b is the bias term constant.

[0081] In order to further implement the above technical solution, the crack length is predicted according to the first material parameter and the second material parameter, specifically including:

[0082] Get the predicted target, which is the target number of cycles.

[0083] Confirm the current roulette state, which includes the current crack length and the current number of cycles; predict the crack length in the next cycle based on the current roulette state, and iterate until the target number of cycles is reached; wherein the prediction cycle can be one or more load cycles.

[0084] Regarding the prediction of the crack length of the next cycle, it specifically includes:

[0085] The crack growth rate of the next cycle is determined according to the first material parameter and the second material parameter:

[0086]

[0087] The crack length of the next cycle is obtained based on the crack growth rate of the next cycle and the current crack length.

[0088] The specific iteration process is:

[0089] a. Initialize the parameters, let x 0 =a i 、x 1 =a i+1 .

[0090] b. Convert to implicit function to solve the equation:

[0091]

[0092] c. Newton-Simpson formula iteration to find x 1 :

[0093] x n+1 =x n -ψ(x n ) / ψ'(x n )

[0094] ψ'(x n )=dψ(x n ) / dx n

[0095] d、a i =(a i +a i+1 ) / 2,x 1 =a i+1 , x 0 =x 1 ;

[0096] e. When the crack length reaches 17.54 mm, terminate the process; otherwise, return to step b.

[0097] Furthermore, when confirming the remaining life, a crack length threshold may be set, and when the crack length is predicted to exceed the crack length threshold, the current number of cycles is output to obtain the remaining life.

[0098] The specific life prediction process is as follows:

[0099] a. Initialize the parameters, let x 0 =N i .

[0100] b. Calculate ΔN i+1 :

[0101]

[0102] c. Iteration format: x 1 =x 0 +ΔN i

[0103] d, x 0 =x 1 , a=(a i +a i+1) / 2;

[0104] e. When the crack length reaches 17.54 mm, terminate the process; otherwise, return to step b.

[0105] In this embodiment, the prediction of different types of cracks can be achieved.

[0106] For surface cracks, the stress intensity factor range is calculated as:

[0107]

[0108] Where Δσ is the applied stress range, a is the crack depth, c is the surface crack length, t is the specimen thickness, W is the specimen half width, θ is the angle between the calculated crack and the surface crack, and Q is the shape factor. Correction factor F for stress intensity factor 1 The calculation formula is as follows:

[0109]

[0110]

[0111] Where a is the crack depth, c is the surface crack length, t is the specimen thickness, W is the specimen width, and θ is the angle between the calculated crack and the surface crack.

[0112] For corner cracks, the stress intensity factor range is calculated as:

[0113]

[0114] Where Δσ is the applied stress range, a is the crack depth, c is the surface crack length, t is the specimen thickness, W is the specimen width, θ is the angle between the calculated crack and the surface crack, Q is the shape factor, and the correction factor F of the stress intensity factor is 2 The calculation formula is as follows:

[0115]

[0116]

[0117] Where a is the crack depth, c is the surface crack length, t is the specimen thickness, W is the specimen width, and θ is the angle between the calculated crack and the surface crack.

[0118] Finally, if Figure 3 , Figure 3In order to obtain the prediction results of this embodiment in the experimental environment, standard (CT) specimens were designed for the original acoustic emission sensor and piezoelectric chip respectively during the experiment. Non-standard (FB) samples were designed to ensure that the stress state of the simulation part is as similar as possible to that of the full-size wheel disc. The sizes of standard specimens are designed according to GB. The non-standard specimens do not consider the influence of the service environment for the time being. The main focus is on the service conditions and constraints. A finite element model is established to analyze the stress state and geometric constraints under the service conditions of the full-size wheel disc, with a focus on the analysis of the prefabricated crack area. According to the optimization design results, the structural simulation parts are processed, and electric sparks are used to prefabricate 1 / 4 elliptical angle cracks or 1 / 2 elliptical surface cracks of different sizes on the simulation parts. Starting from the standard crack extension specimen, the acoustic emission characteristics of the crack extension of the engine alloy wheel disc are studied, and a database of acoustic emission signal characteristics at different frequencies and extension rates is established. Based on the existing commercial detection equipment, the feasibility and reliability of using small-size piezoelectric chips as acoustic emission signal acquisition sensors are studied to provide a solution for real-time monitoring of crack extension of structural simulation parts and full-size components. According to Figure 3 It can be found that the prediction results are all within the 2-fold error band, with good accuracy and generalization ability.

[0119] In addition, during the experiment, the experimental conditions and fitting parameters of each sample are shown in Table 1 below:

[0120] Table 1

[0121]

[0122] In order to further implement the above technical solution, the steps also include: based on the prediction results, comparing the prediction accuracy of the crack length and cycle number of the AE sensor and the piezoelectric chip, as well as the prediction effects of different types of samples.

[0123] Use goodness of fit R 2 Indicates the prediction accuracy, and the goodness of fit refers to the degree of fit of the regression line to the observed value. The calculation formula is as follows:

[0124]

[0125] In the formula, is the prediction result, x i For the test results, is the mean of the test results.

[0126] The experimental results of accuracy prediction are shown in Table 2:

[0127] Table 2

[0128]

[0129] According to Table 2, the prediction accuracy of CT-0.1-0.05-12 and CT-0.1-0.05-9 samples is relatively high. The fitting results of CT-0.1-0.05-9 samples are applied to CT-0.1-0.05-14 and CT-0.1-0.05-16 samples, which still have good generalization ability.

[0130] In order to further implement the above technical solution, the present invention can also predict the damage state. Clustering is performed according to the acoustic emission signals, and the clustering results are associated with multiple crack extension stages to obtain multiple damage state labels; the damage state labels are compared with the actual damage to confirm the damage state of the actual damage.

[0131] Specifically, first, the acoustic emission signals were preprocessed, the data groups with signal parameters less than 0 were removed, and 11 signals including energy, signal intensity, duration, absolute energy, RMS, ASL, count, rise time, peak frequency, amplitude and average frequency were selected as preprocessed data.

[0132] Secondly, the complete link hierarchical clustering algorithm (CLCA) is used on the preprocessed AE data set. The algorithm results are shown in Figure 4 As shown, eight signals of energy, duration, absolute energy, RMS, count, rise time, peak frequency and amplitude whose cluster distance values ​​are less than 0.2 are selected as cluster components.

[0133] Then, after excluding highly correlated components, the remaining components are normalized in the range of (0,1), and the PCA algorithm is used to reduce the dimension of the feature vector. The number of principal components is selected according to the DB and SI index evaluation. The evaluation indicators are shown in Table 3. The k value corresponding to the minimum DB value and the maximum SI value is selected as the optimal clustering number. The clustering effect is shown in Figure 5 As shown, it can be seen from the figure that cluster 1 has the characteristics of low energy and high count value, cluster 2 has the characteristics of high energy and high count value, and cluster 3 has the characteristics of low energy and low count value.

[0134] Table 3

[0135] k value DB Index SI Index 2 0.79 0.72 3 0.67 0.76 4 0.79 0.67 5 0.83 0.63 6 0.82 0.61 7 0.78 0.62 8 0.82 0.59

[0136] Finally, the extracted principal components are clustered using K-means, and the clustering results are mapped to a high-dimensional feature space to obtain the cumulative energy changes of each cluster during the crack propagation process, such as Figure 5As shown in the figure, it can be seen that 8% AE data corresponds to cluster 1, 20.2% AE data corresponds to cluster 2, and 71.8% AE data corresponds to cluster 3. In the crack extension stage, the cumulative energy corresponding to cluster 3 has been at a low level, and the cumulative energy corresponding to cluster 2 has been steadily increasing. The cumulative energy corresponding to cluster 1 increases relatively slowly in the initial stage, and the rate of increase significantly increases in the later stage of crack extension. Combined with the results of step S23, it can be seen that cluster 1 represents the initial stage of crack extension, the crack extension rate is slow, and the energy increases slowly. Cluster 2 represents the stable crack extension stage, and the energy during the crack extension process increases steadily. Cluster 3 represents the later stage of crack extension, and the crack extension rate is significantly accelerated, resulting in a sharp increase in cumulative energy.

[0137] In this embodiment, the damage state of the roulette wheel can be identified by training the prediction model, and the signal features corresponding to different states are obtained after the training set is processed by the above processing method. The corresponding features are identified by the classifier to achieve prediction.

[0138] In addition, the damage state corresponding to the actual signal can be determined by calculating the characteristic distance between the actual signal and each cluster signal and combining the distance threshold.

[0139] Example 2

[0140] like Figure 7 Based on the same inventive concept, the embodiment of the present invention discloses a wheel structure fatigue crack prediction system based on acoustic emission signals and in-situ images, including:

[0141] Acoustic emission monitoring subsystem: The acoustic emission monitoring subsystem is used to obtain the acoustic emission signal and in-situ crack image of the target structure, including: acoustic emission sensor, piezoelectric chip, image acquisition equipment and controller.

[0142] The acoustic emission sensor, piezoelectric and image acquisition equipment are electrically connected to the controller; the acoustic emission sensor is used to monitor the acoustic emission vibration signal; the piezoelectric chip is used to convert the acoustic emission vibration signal into an electrical signal and send it to the controller; the image acquisition device is used to collect the in-situ image of the wheel structure and send it to the controller; the controller is used to predict the crack length of the wheel and confirm the remaining life according to the acoustic emission signal and the in-situ image.

[0143] In this embodiment, the upper and lower clamps are used to fix the simulated parts for the experiment, and the simulated parts are subjected to mechanical tests by applying loads. The simulated parts are the experimental objects, and are subjected to stress under the action of the upper and lower clamps. The test machine control system is responsible for controlling the loading process of the upper and lower clamps, and accurately controlling the load size and loading rate.

[0144] Sensors and piezoelectric chips are used to monitor the acoustic emission signals of the simulated parts during the stress process. These signals are transmitted through lines to the "Acoustic Emission Control System" for analysis.

[0145] The present invention uses a crack opening displacement gauge (COD) or a differential interference contrast microscope (DIC) to monitor the crack length. The lens and camera cooperate with "lighting lamp 1" and "lighting lamp 2" to perform in-situ imaging of the simulated part, and the image data is transmitted to the "in-situ camera control system" for processing and analysis.

[0146] The present invention causes deformation or damage to the simulated part through mechanical loading, and uses sensors and acoustic emission systems to monitor changes inside the material, and records surface changes through an optical imaging system. The control systems work together to comprehensively study the performance and behavior of the material during the stress process.

[0147] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.

[0148] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for predicting fatigue cracks of a wheel disc structure based on acoustic emission signals and in-situ images, characterized in that: The following steps are involved: Acquire the acoustic emission signals and in-situ crack images of the disk structure under multiple load cycles; Perform fitting according to the acoustic emission signals corresponding to a plurality of load cycles to generate a first fitting function; determine a first material parameter according to the first fitting function and the Forman equation; Perform fitting according to the corresponding in-situ crack images under multiple load cycles to generate a second fitting function; determine the second material parameter according to the second fitting function combined with the Forman equation; A crack length is predicted based on the first material parameter and the second material parameter.

2. The method for predicting fatigue cracks of a wheel disc structure based on acoustic emission signals and in-situ images according to claim 1 is characterized in that: The step also includes: confirming the remaining life according to a preset crack length and a predicted value of the crack length.

3. A method for predicting fatigue cracks of a wheel disc structure based on acoustic emission signals and in-situ images according to claim 1 or 2, characterized in that: Generating the first fitting function specifically includes: Extract characteristic parameters according to the acoustic emission signal and calculate the cumulative energy value to obtain a series of first associated data points consisting of cumulative energy values ​​and cycle numbers; The first associated data points are fitted and derived to obtain an energy release rate curve, namely, the first fitting function.

4. The method for predicting fatigue cracks of a wheel disc structure based on acoustic emission signals and in-situ images according to claim 3 is characterized in that: Confirming a first material parameter according to the first fitting function and the Forman equation specifically includes: According to the correlation between energy release rate and crack growth rate and the Forman equation, the correlation between the energy release rate and the stress intensity range is confirmed: Where U is the cumulative released energy, N is the number of load cycles, is the energy release rate; c1 and m1 are the first material parameters, ΔK is the stress intensity range; R is the load ratio, K c is the fracture toughness; The energy release rate is confirmed by the first fitting function and linearly fitted with the corresponding stress intensity range to obtain the first material parameter.

5. A method for predicting fatigue cracks of a wheel disc structure based on acoustic emission signals and in-situ images according to claim 1 or 2, characterized in that: Generating the second fitting function specifically includes: Acquire the crack length according to the in-situ image to obtain a series of second associated data points consisting of the crack length and the number of cycles; The second associated data points are fitted and differentiated to obtain a crack length extension rate curve, that is, the second fitting function.

6. The method for predicting fatigue cracks of a wheel disc structure based on acoustic emission signals and in-situ images according to claim 5, characterized in that: Determining the second material parameter according to the second fitting function in combination with the Forman equation specifically includes: The Forman equation is: Where a is the crack length, N is the number of load cycles, is the crack length extension rate, c2 and m2 are the second material parameters; The crack length extension rate is confirmed according to the second fitting function, and a linear fitting is performed in combination with a corresponding stress intensity range to confirm the second material parameter.

7. A method for predicting fatigue cracks of a wheel disc structure based on acoustic emission signals and in-situ images according to claim 1 or 2, characterized in that: The predicted crack length specifically includes: Obtaining a prediction target, wherein the prediction target is a target number of cycles; Confirming a current roulette state, the current roulette state including a current crack length and a current number of cycles; predicting a crack length in a next cycle based on the current roulette state, and iterating until the target number of cycles is reached; The predicting of the crack length in the next cycle specifically includes: The crack growth rate of the next cycle is determined according to the first material parameter and the second material parameter: The crack length of the next cycle is obtained based on the crack growth rate of the next cycle and the current crack length.

8. The method for predicting fatigue cracks of a wheel disc structure based on acoustic emission signals and in-situ images according to claim 2 is characterized in that: The confirming of the remaining life specifically includes: setting a crack length threshold, and when the crack length is predicted to exceed the crack length threshold, outputting the current number of cycles to obtain the remaining life.

9. A method for predicting fatigue cracks in a wheel disc structure based on acoustic emission signals and in-situ images according to claim 1, characterized in that: The steps also include: clustering according to the acoustic emission signals, and associating the clustering results with multiple crack extension stages to obtain multiple damage state labels; and comparing the damage state labels with the actual damage to confirm the damage state of the actual damage.

10. A wheel disc structure fatigue crack prediction system based on acoustic emission signals and in-situ images, characterized in that: The prediction method according to any one of claims 1 to 9 is adopted, comprising: an acoustic emission monitoring subsystem, the acoustic emission monitoring subsystem is used to obtain the acoustic emission signal and the in-situ image of the crack of the target structure, comprising: an acoustic emission sensor, a piezoelectric chip, an image acquisition device and a controller; The acoustic emission sensor, the piezoelectric and the image acquisition device are electrically connected to the controller; The acoustic emission sensor is used to monitor the acoustic emission vibration signal; The piezoelectric chip is used to convert the acoustic emission vibration signal into an electrical signal and send it to the controller; The image acquisition device is used to acquire an in-situ image of the wheel structure and send it to the controller; The controller is used to predict the crack length of the wheel disc and confirm the remaining life according to the acoustic emission signal and the in-situ image.

Images