Ultrasonic fatigue experiment monitoring method based on digital twinning

Abstract

The invention belongs to the field of material performance testing of mechanical structure strength, and particularly relates to an ultrasonic fatigue experiment monitoring method based on digital twinning. According to the method, on the basis of an ultrasonic fatigue testing machine platform, an ultrasonic fatigue digital twin model of a tested object is established by collecting frequency, temperature, load and other test data, and real-time online monitoring of ultrasonic fatigue and life prediction are carried out in combination with a fatigue life prediction model. Compared with a traditional test method, the method has the advantages that ultrasonic detection data and a digital twinborn model technology are combined, the fatigue life of the material is predicted and evaluated, the test efficiency is improved, the long-life test cost is reduced, and the accuracy and reliability of test data are improved.

Description

Technical Field

[0001] The present invention relates to the fields of materials science and engineering and digital technology, and in particular to an ultrasonic fatigue test monitoring method based on digital twins. Background Art

[0002] The ultra-long life and safe service of mechanical structures are on the rise. As an important load-bearing form of the structure, fatigue and its life are gradually developing from low-cycle fatigue to high-cycle fatigue and ultra-high-cycle fatigue. The actual service life of key components such as turbine rotors and engine blades has reached 10 8 -10 10 Cycles and above pose new challenges to the ultra-high cycle fatigue strength of materials and structures. In this context, traditional fatigue performance research can no longer meet the needs of development. In order to save time and cost, the use of higher frequency fatigue testing machines to conduct material mechanical behavior tests is an important option. Among them, the working frequency of ultrasonic fatigue is 20kHz±500Hz, which is widely used in the ultra-high cycle fatigue behavior of materials. However, higher frequencies will affect the resonance conditions, and the performance requirements for specimens and equipment are also higher. In addition, the ultrasonic fatigue test itself takes a long time and requires high accuracy in life testing. It is difficult to use sensors to collect data at a frequency of 20kHz, which makes it difficult to achieve online monitoring of the fatigue process, even if lasers may be a This is a prediction method, but it provides very little data and it is difficult to judge the results. Digital twin technology, with its advantages of real-time monitoring, high-precision prediction and comprehensive evaluation, can monitor the fatigue process online and realize real-time display of the ultrasonic fatigue test process, which is of great value to improving the test accuracy of ultrasonic fatigue test processes at home and abroad. At the same time, traditional material fatigue life prediction methods have problems such as low prediction accuracy and high test cost. Therefore, the ultrasonic fatigue experiment monitoring method based on digital twin can not only realize the comprehensive monitoring of the specimen during the ultrasonic fatigue test, state visualization and convenient prediction of the fatigue life of the specimen, but also help promote the development of the material fatigue life prediction industry towards digitalization and intelligence, and has broad application prospects.

[0003] The ultrasonic fatigue test system is a resonant system composed of an ultrasonic fatigue test mechanical device and a test piece. First, the ultrasonic generator converts the voltage into an electrical signal, and the horn amplifies the vibration from the transducer, thereby converting it into a larger amplitude. The tool head can transmit the high-frequency vibration of the system to the sample test, and finally transmit it to the end of the test piece to complete the ultra-high frequency load loading of the test piece; Digital twin technology is a method that combines the actual system with the virtual model, and can predict the performance and behavior of the system through simulation; At present, the application of digital twin technology in the industrial field is becoming more and more extensive, but there is relatively little research on material fatigue life prediction;

[0004] As a key enabling technology for solving the interaction problem between digital models and physical entities and implementing the concept and goals of digital transformation, digital twin technology will play an important role in supporting the entire process of product development and facilitating the integrated innovation of scientific research, production and management. Ultrasonic fatigue testing is a new technology suitable for the ultra-high cycle fatigue performance of metal materials. Building a digital twin model with real-time characteristics for it will help expand its scope of application, seek real-time parameter monitoring, intelligent work, and practicality. Summary of the invention

[0005] Based on the existing technical problems, the present invention proposes an ultrasonic fatigue experiment monitoring method based on digital twin.

[0006] The present invention proposes an ultrasonic fatigue test monitoring method based on digital twin, comprising the following steps:

[0007] S1, obtain the geometric structure parameters and material properties of the laboratory ultrasonic fatigue specimen;

[0008] S2, obtaining the initial working conditions and environmental parameters of the ultrasonic fatigue test;

[0009] S3, build a digital twin model based on the data obtained in S2;

[0010] S4, builds a complete digital twin model based on the digital twin sub-model in S3;

[0011] S5, during the ultrasonic fatigue test, real-time monitoring of the relevant data of the fatigue life of metal materials;

[0012] S6, using signal processing and statistical tools in MATLAB to identify and extract features related to metal fatigue life, and preprocessing the data obtained in S5, including data cleaning, outlier removal, and data smoothing operations;

[0013] S7, based on the data processed by S6, the digital twin model updates the parameters of the digital twin model in S4 in real time and makes predictions;

[0014] S8, real-time calculation and collection of the deviation between the predicted result in S7 and the actual result in S5;

[0015] S9, adjusting and correcting the digital twin model in S4 according to the deviation data calculated in S8, obtaining a new and more accurate digital twin model, and optimizing the digital twin model;

[0016] S10, use the trained digital twin model to predict the fatigue life of the new specimen. The prediction result is the fatigue life of the material under the corresponding load under certain environmental conditions, and the optimized design is carried out based on this.

[0017] Preferably, the performance parameters of the material to be tested are queried in S1, including its brand and mechanical performance parameters, the ultrasonic fatigue testing machine in S2 is composed of an ultrasonic fatigue testing mechanical device and a test piece to form a resonant system, and the working conditions and environmental parameters in S2 include the working frequency, temperature and load during the ultrasonic fatigue test.

[0018] Preferably, the digital twin model in S3 includes building a specimen frequency evolution model, a stress model, a strain model, a temperature field model, and a fatigue life prediction model.

[0019] Preferably, S3 comprises the following steps:

[0020] S31, build the specimen frequency evolution model based on the resonance frequency evolution law:

[0021] S32, build a stress model based on the maximum stress in the middle section of the specimen;

[0022] S33, build a strain model based on Hooke's law, the dynamic elastic modulus of the material and the stress model in S32;

[0023] S34, build a temperature field model based on thermal analysis;

[0024] S35, build a fatigue life prediction model based on infrared thermography.

[0025] Preferably, in step S31, during the ultrasonic fatigue process of the metal material, as the number of cycles increases, the resonant frequency of the sample continues to decrease, and the frequency image collected in real time is used as a reference to draw a relative frequency-cycle relationship diagram of the relative frequency f and the number of cycles N, and obtain the frequency change rate df / dN, thereby preliminarily understanding the ultra-high cycle fatigue damage of the sample.

[0026] Preferably, in step S32, the ultrasonic fatigue test selects ultrasonic fatigue specimens according to different test conditions, loading methods, and frequency effects to adapt to different test conditions. It is assumed that several commonly used ultrasonic fatigue specimens all use the axial center of the specimen as the coordinate origin. During the test, the stress amplitude is the largest in the middle of the specimen and is zero at both ends; while the displacement is zero in the middle of the specimen and is the largest at both ends. Therefore, a stress model is constructed based on the maximum stress amplitude at the middle section of the specimen:

[0027]

[0028] Where E is the elastic modulus of the material, and U0 is the maximum displacement of the specimen.

[0029] Preferably, in step S33, the stress and strain of the ultrasonic fatigue specimen belong to the elastic deformation range, and the conversion calculation of the corresponding stress amplitude and strain amplitude is performed by Hooke's law and the dynamic elastic modulus of the material to build a strain model:

[0030]

[0031] In the formula, σ max is the maximum stress in the middle section of the specimen, E d is the dynamic elastic modulus of the material, ε max is the maximum strain in the middle section of the specimen.

[0032] Preferably, in step S34, based on the aspect ratio of the ultrasonic fatigue specimen and the temperature distribution of the specimen along the length direction due to heat conduction, the surface temperature of the specimen can be simplified into a one-dimensional field, and the temperature is distributed in a gradient along the length direction of the specimen. The center of the specimen is taken as the origin, the direction close to the clamping end is the negative direction of the x-axis, and the direction of the free end is the positive direction of the x-axis. The temperature field model of the specimen is:

[0033] θ(x)=Α1x 2 +Α2x+Α3

[0034] Wherein, Α1<0, θ is the one-dimensional temperature value, and Α2 and Α3 are fitting parameters.

[0035] Preferably, in step S35, the fatigue life prediction model of infrared thermal imaging is used to predict the fatigue life of the material based on the initial temperature rise slope R. The fatigue life prediction expression is as follows:

[0036] N f R=C

[0037] Where N f is the material fatigue life, R is the temperature rise slope, and C is the material constant.

[0038] Preferably, in step S4, the parameters of each digital twin sub-model are integrated based on the Simulink visualization tool platform in MATLAB, and fully debugged to build an integrated simulation platform of multiple physical fields to form a complete digital twin model;

[0039] The data related to the fatigue life of the material in S5 include ultrasonic fatigue test related data, including the end amplitude, temperature, test frequency, load amplitude, and strain of the specimen. The data related to the fatigue life of the specimen material in S5 is obtained by actual records of the control system of the ultrasonic fatigue testing machine;

[0040] The method for adjusting and updating the parameters in S9 adopts an extended Kalman filter algorithm.

[0041] The beneficial effects of the present invention are:

[0042] 1. By setting up a digital twin model of the research object with an ultrasonic fatigue testing machine as the experimental platform, a fatigue prediction model is established, and fatigue life prediction and optimization are performed based on the ultrasonic fatigue digital twin, which effectively reduces production and maintenance costs and improves product reliability and safety. Compared with traditional test methods, the present invention combines ultrasonic detection data with mathematical modeling and simulation technology to realize the prediction and evaluation of material fatigue life, reduce safety risks, and also provide a more accurate basis for structural design, optimization and improvement to improve performance and economic benefits. In addition, before conducting actual tests to understand the fatigue performance of materials, preliminary experiments can be carried out with the help of the system to lay the foundation for formal experiments, thereby reducing the consumption of manpower and material resources and improving experimental efficiency to achieve low-carbon and environmental protection.

[0043] 2. Improve efficiency: The effective use of ultrasonic fatigue digital twins to predict material fatigue life can calculate, simulate and predict the fatigue life of metals. Compared with traditional fatigue tests that consume a lot of time and resources, this patent develops an experimental monitoring method that saves test time and costs;

[0044] Improved accuracy: Ultrasonic fatigue testing itself takes a long time and requires high accuracy in fatigue life testing. This patent proposes a method for predicting material fatigue life using ultrasonic fatigue digital twins. It predicts material fatigue life based on the initial temperature rise slope R, and can provide more accurate fatigue life prediction results.

[0045] Process visualization: The ultrasonic fatigue digital twin method for predicting material fatigue life can monitor the fatigue process online and realize the real-time display of the ultrasonic fatigue test process, which is of great value in improving the test accuracy of the ultrasonic fatigue test process at home and abroad.

[0046] Optimization design: By simulating and predicting metal fatigue life, the ultrasonic fatigue digital twin method for predicting material fatigue life can be used to optimize the design and manufacturing process of metal materials to improve the durability and reliability of components.

[0047] 3. Through the application of digital twin technology, a quick and accurate fatigue prediction method is provided, without the need for a large number of experiments, reducing the waste of time and resources; by comparing and matching the physical properties of actual materials with digital models, this technology can quickly and accurately predict the fatigue life of materials, simulate different fatigue loading conditions and material properties, and thus help researchers understand the factors affecting fatigue life; in addition, digital twin technology improves the efficiency of structural design and optimization, helps researchers quickly evaluate the fatigue performance of materials, provides feasibility studies and references for structural safety, and reduces the cost of trial and error and iteration; the application of digital twin technology fully reflects the new concept of data-driven manufacturing development in the context of new generation information technology, industrial Internet technology and intelligent manufacturing concepts. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 This is a step diagram of an ultrasonic fatigue test monitoring method based on digital twin proposed by the present invention;

[0049] Figure 2 This is a flow chart of an ultrasonic fatigue test monitoring method based on digital twin proposed in the present invention;

[0050] Figure 3 Several commonly used ultrasonic fatigue specimens and their stress amplitude and displacement schematic diagrams of an ultrasonic fatigue test monitoring method based on digital twin proposed by the present invention;

[0051] Figure 4 This is a dog-bone specimen dimension diagram of an ultrasonic fatigue test monitoring method based on digital twin proposed in the present invention;

[0052] Figure 5 A stereoscopic diagram of a transducer model of an ultrasonic fatigue test monitoring method based on digital twin proposed in the present invention;

[0053] Figure 6 A stereoscopic diagram of a horn model of an ultrasonic fatigue test monitoring method based on digital twinning proposed in the present invention;

[0054] Figure 7 A stereoscopic diagram of a dog-bone specimen model of an ultrasonic fatigue test monitoring method based on digital twinning proposed in the present invention;

[0055] Figure 8 This is a three-dimensional diagram of the key structure assembly model of the ultrasonic fatigue test based on the digital twin ultrasonic fatigue test monitoring method proposed in the present invention. DETAILED DESCRIPTION

[0056] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0057] Reference Figure 1-Figure 8 , an ultrasonic fatigue test monitoring method based on digital twin. In order to make the purpose, technical scheme and advantages of the present invention clearer, the present invention is further described in detail below in conjunction with the accompanying drawings and the ultrasonic fatigue test of 25Cr2Ni2MoV steel welded joint.

[0058] like Figure 1 As shown, the digital twin modeling method for the laboratory ultrasonic fatigue test process proposed in the present invention includes the following steps:

[0059] S1, obtaining the geometric structure parameters and material properties of the laboratory ultrasonic fatigue specimen, wherein the geometric structure parameters can be obtained from the drawing file of the specimen; the material properties at least include querying the grade and mechanical property parameters of the material to be tested, which can be obtained through the material manual or laboratory test;

[0060] Furthermore, commonly used specimens include uniform cross-section cylindrical specimens, variable cross-section cylindrical specimens (variable cross-section cylindrical specimens are also called dog-bone specimens, which are divided into two types of cylindrical specimens: one without an intermediate uniform cross-section section, and the other with an intermediate uniform cross-section section), and plate specimens; among them, uniform cross-section specimens are the simplest type of specimens, but have a low displacement stress coefficient and are not suitable for ultrasonic fatigue testing of high-strength steel. They are generally only used for displacement calibration of ultrasonic fatigue testing machines; variable cross-section cylindrical specimens can shorten the specimen length and obtain a larger displacement stress coefficient, thereby increasing the heat dissipation area of ​​the specimen; when it is necessary to study the effect of surface treatment on the fatigue properties of the material, it is required to have an equal stress cross-section area in the middle of the specimen, and a variable cross-section circular arc specimen cannot meet the requirements, and a circular arc specimen with an intermediate uniform cross-section section is required; and when it is necessary to perform ultrasonic fatigue on a plate specimen, the ultrasonic fatigue testing machine system control software is equipped with a circular specimen that cannot meet the test requirements, and a plate-shaped ultrasonic fatigue specimen is required.

[0061] S2, ultrasonic fatigue testing machine is composed of ultrasonic fatigue testing mechanical device and test piece to form a resonant system. It simulates the fatigue damage of materials in actual use by applying high-frequency ultrasonic vibration load. When conducting ultrasonic fatigue test, it is necessary to determine the working parameters, mainly including ultrasonic generator power, frequency, amplitude, temperature and load mode. By adjusting these parameters, the fatigue performance of materials under different working conditions can be simulated, providing a theoretical basis for material fatigue life prediction and performance evaluation, and designing test plans to carry out research.

[0062] A common ultrasonic fatigue testing machine system mainly includes an ultrasonic generator, a piezoelectric transducer, an amplifier, a specimen and a cooling device; the ultrasonic generator acts as an exciter with an operating frequency of 20kHz±500Hz, which is generated through system resonance; the piezoelectric transducer converts the high-frequency electrical signal into a high-frequency resonance, which is transmitted to the ultrasonic fatigue specimen through a displacement amplifier to obtain the corresponding stress.

[0063] S3, building a digital twin model of the test bench according to the characteristics, initial working conditions and environmental parameters obtained in S2 and the physical action relationship in the ultrasonic fatigue test; the vibration behavior of the specimen in the ultrasonic fatigue test can be described by dynamic equations and vibration theory, so as to build a specimen frequency evolution model, stress model, strain model and temperature field model by taking the dog-bone specimen of the 25Cr2Ni2MoV steel welded joint as the test object, including the following steps:

[0064] S31, build the specimen frequency evolution model based on the resonance frequency evolution law;

[0065] S32, build a stress model based on the maximum stress in the middle section of the specimen;

[0066] S33, build a strain model based on Hooke's law, the dynamic elastic modulus of the material and the stress model in S32;

[0067] S34, build a temperature field model based on thermal analysis;

[0068] S35, build a fatigue life prediction model based on infrared thermography.

[0069] Furthermore, in step S31, through the law of resonance frequency evolution, it can be inferred that in the ultrasonic fatigue process of metal materials, as the number of cycles increases, the resonance frequency of the sample continues to decrease; the frequency image collected in real time can be used as a reference to draw a relative frequency-cycle relationship diagram of the relative frequency f and the number of cycles N, and obtain the frequency change rate df / dN, based on which the ultra-high cycle fatigue damage of the sample can be preliminarily understood.

[0070] In step S32, since the stress in the middle section of the dog-bone-shaped specimen does not change much, the stress in this section can be approximated to be at the maximum stress level. Therefore, a stress model is constructed based on the maximum stress in the middle section of the specimen:

[0071]

[0072] in,

[0073]

[0074]

[0075] In the formula, σ max is the maximum stress in the middle section of the specimen, E d is the dynamic elastic modulus of the material, w is the angular frequency; c is the propagation speed of ultrasound in the material; R1, R2, L1 and L2 are the specimen dimensions, L3 is the resonant length, and U0 is the maximum displacement amplitude of the specimen;

[0076] Furthermore, in step S33, since the stress and strain of the ultrasonic fatigue specimen are very small and belong to the elastic deformation range, the corresponding stress amplitude and strain amplitude can be converted and calculated by Hooke's law and the dynamic elastic modulus of the material to build a strain model:

[0077]

[0078] In the formula, σ max is the maximum stress in the middle section of the specimen, E dis the dynamic elastic modulus of the material, ε max is the maximum strain in the middle section of the specimen;

[0079] Further, in step S34, based on the aspect ratio of the ultrasonic fatigue specimen and the temperature distribution of the specimen along the length direction due to heat conduction, the surface temperature of the specimen can be simplified into a one-dimensional field, and the temperature is distributed in a gradient along the length direction of the specimen; further, with the center of the specimen as the origin, the direction close to the clamping end as the negative direction of the x-axis, and the direction of the free end as the positive direction of the x-axis, the temperature field model of the specimen temperature is:

[0080] θ(x)=Α1x 2 +Α2x+Α3

[0081] Wherein, Α1<0, θ is the one-dimensional temperature value, and Α2 and Α3 are fitting parameters.

[0082] In step S35, a fatigue life prediction model is built based on infrared thermal imaging to predict the fatigue life of 25Cr2Ni2MoV steel welded joints; by fitting the initial temperature rise slope and the fatigue life under the corresponding loading stress, the value of the material constant C in the fatigue life prediction expression is obtained, and then the predicted fatigue life is obtained from the initial temperature rise slope.

[0083] S4, consider the coordination relationship and interface cooperation between the different digital twin sub-models obtained in S3. First, it is necessary to clarify the data transmission and interaction methods between the sub-models. The Simulink visualization tool platform in MATLAB can be used to establish a multi-physics field integrated simulation platform containing multiple sub-models, integrate the executable sub-models, and fully debug to ensure the coordination and matching between the interfaces of each sub-model. At the same time, consider the interaction and influence between different physical fields, gradually adjust the parameters and model settings, provide more comprehensive simulation results, and improve the accuracy and reliability of the integrated simulation platform.

[0084] S5, the ultrasonic fatigue machine applies fatigue load to the material to be tested, and at the same time, the ultrasonic fatigue test machine control system monitors and records the relevant data of the ultrasonic fatigue test, mainly including the test frequency, load amplitude, strain, and temperature; among them, the temperature of the specimen during the test is analyzed and processed by the IRSoft software provided by the infrared thermal imaging instrument to analyze the distribution and evolution of the temperature; the computer control software of the ultrasonic fatigue testing machine can record the test resonance frequency and test time, and the strain distribution can be measured by the micro dynamic strain gauge pasted on the surface of the specimen. They together constitute the computer measurement and control device of the test. After that, the fatigue-life curve of specimens of different materials can be drawn by processing the data.

[0085] Among them, the fatigue process of material samples is accompanied by heat dissipation. The heat dissipation cannot be obtained directly, but can be obtained indirectly from the temperature field detection of the material samples. Currently, non-contact measurement methods have been widely used. Infrared thermal imagers can accurately record the temperature evolution of the specimen surface during the entire fatigue process. It can directly observe the target at a certain distance and monitor and record the evolution of the specimen surface temperature during the fatigue process in real time.

[0086] S6, use signal processing and statistical tools in MATLAB to identify and extract features related to metal fatigue life, and preprocess the data obtained in S5, including data cleaning, removal of outliers, and data smoothing operations.

[0087] In MATLAB, the signal processing toolbox and the statistics toolbox can be used to identify and extract features related to metal fatigue life. For example, the signal processing toolbox can include filtering, spectrum analysis, waveform processing and other functions, which can be used to process signal information in experimental data, such as denoising, smoothing, feature extraction, etc. The statistics toolbox can implement statistical analysis, regression analysis, hypothesis testing functions, which can be used to perform statistical analysis and feature extraction on experimental data, such as calculating mean, variance, and correlation coefficient statistics. Data preprocessing tools can include data cleaning, outlier processing, and data smoothing functions, which can be used to preprocess raw data to improve data quality and reliability. When performing data preprocessing, the Min-Max normalization method can be used to normalize the data. Min-Max normalization is a linear transformation method that scales data to a specified range and is often used to map data to the [0,1] interval. This helps to eliminate dimensional differences between different features and improve the comparability of data and the stability of the model. By using various tools and methods in MATLAB to carefully process and extract features from experimental data, the features related to metal fatigue life can be analyzed more comprehensively, providing a scientific basis for further research and application.

[0088] S7, based on the data processed by S6, the digital twin model updates the digital twin model parameters in S4 in real time and makes predictions.

[0089] S8, real-time calculation and collection of the deviation between the predicted result in S7 and the actual result in S5;

[0090] S9, adjusting and correcting the digital twin model in S4 according to the deviation data calculated in S8 to obtain a new and more accurate digital twin model to achieve optimization of the digital twin model;

[0091] If the prediction error is large, the numerical simulation method can be combined to introduce the ultrasonic fatigue correction coefficient, and the test data can be corrected to optimize the model selection and other steps to further improve the prediction accuracy, so that the prediction results output by the model are consistent with the actual monitoring data to achieve the optimization of the digital twin model. The parameter adjustment and update method can adopt appropriate nonlinear system modeling methods and state estimation algorithms, such as extended Kalman filtering (EKF) or particle filtering, to more accurately predict fatigue life.

[0092] The trained digital twin model is used to predict the fatigue life of new specimens. The prediction result is the fatigue life of the material under the corresponding load under certain environmental conditions. The design can be optimized based on this, thereby improving the service life of the metal material in the actual working environment.

[0093] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. An ultrasonic fatigue test monitoring method based on digital twin, characterized by: The following steps are involved: S1, obtain the geometric structure parameters and material properties of the laboratory ultrasonic fatigue specimen; S2, obtaining the initial working conditions and environmental parameters of the ultrasonic fatigue test; S3, build a digital twin model based on the data obtained in S2; S4, builds a complete digital twin model based on the digital twin sub-model in S3; S5, during the ultrasonic fatigue test, real-time monitoring of the relevant data of the fatigue life of metal materials; S6, using signal processing and statistical tools in MATLAB to identify and extract features related to metal fatigue life, and preprocessing the data obtained in S5, including data cleaning, outlier removal, and data smoothing operations; S7, based on the data processed by S6, the digital twin model updates the parameters of the digital twin model in S4 in real time and makes predictions; S8, real-time calculation and collection of the deviation between the predicted result in S7 and the actual result in S5; S9, adjusting and correcting the digital twin model in S4 according to the deviation data calculated in S8, obtaining a new and more accurate digital twin model, and optimizing the digital twin model; S10, use the trained digital twin model to predict the fatigue life of the new specimen. The prediction result is the fatigue life of the material under the corresponding load under certain environmental conditions, and the optimized design is carried out based on this.

2. According to the ultrasonic fatigue test monitoring method based on digital twinning according to claim 1, it is characterized in that: The performance parameters of the material to be tested are queried in S1, including its brand and mechanical performance parameters. The ultrasonic fatigue testing machine in S2 is composed of an ultrasonic fatigue testing mechanical device and a test piece to form a resonant system. The working conditions and environmental parameters in S2 include the working frequency, temperature and load during the ultrasonic fatigue test.

3. According to the ultrasonic fatigue test monitoring method based on digital twinning according to claim 1, it is characterized in that: The digital twin model in S3 includes building a specimen frequency evolution model, a stress model, a strain model, a temperature field model, and a fatigue life prediction model.

4. According to the ultrasonic fatigue test monitoring method based on digital twinning according to claim 1, it is characterized in that: The S3 comprises the following steps: S31, build the specimen frequency evolution model based on the resonance frequency evolution law: S32, build a stress model based on the maximum stress in the middle section of the specimen; S33, build a strain model based on Hooke's law, the dynamic elastic modulus of the material and the stress model in S32; S34, build a temperature field model based on thermal analysis; S35, build a fatigue life prediction model based on infrared thermography.

5. The ultrasonic fatigue test monitoring method based on digital twin according to claim 1 is characterized in that: In the step S31, during the ultrasonic fatigue process of the metal material, as the number of cycles increases, the resonant frequency of the sample continues to decrease, and the frequency image collected in real time is used as a reference to draw a relative frequency-cycle relationship diagram of the relative frequency f and the number of cycles N, and obtain the frequency change rate df / dN, based on which a preliminary understanding of the ultra-high cycle fatigue damage of the sample is obtained.

6. The ultrasonic fatigue test monitoring method based on digital twin according to claim 1 is characterized in that: In step S32, the ultrasonic fatigue test selects ultrasonic fatigue specimens according to different test conditions, loading methods, and frequency effects to adapt to different test conditions. It is assumed that several commonly used ultrasonic fatigue specimens all use the axial center of the specimen as the coordinate origin. During the test, the stress amplitude is the largest in the middle of the specimen and is zero at both ends; while the displacement is zero in the middle of the specimen and is the largest at both ends. Therefore, a stress model is built based on the maximum stress amplitude at the middle section of the specimen: Where E is the elastic modulus of the material, and U0 is the maximum displacement of the specimen.

7. The ultrasonic fatigue test monitoring method based on digital twin according to claim 1 is characterized in that: In step S33, the stress and strain of the ultrasonic fatigue specimen belong to the elastic deformation range, and the corresponding stress amplitude and strain amplitude are converted and calculated by Hooke's law and the dynamic elastic modulus of the material to build a strain model: In the formula, σ max is the maximum stress in the middle section of the specimen, E d is the dynamic elastic modulus of the material, ε max is the maximum strain in the middle section of the specimen.

8. The ultrasonic fatigue test monitoring method based on digital twin according to claim 1 is characterized in that: In step S34, based on the aspect ratio of the ultrasonic fatigue specimen and the temperature distribution of the specimen along the length direction due to heat conduction, the surface temperature of the specimen can be simplified into a one-dimensional field, and the temperature is distributed in a gradient along the length direction of the specimen. The center of the specimen is taken as the origin, the direction close to the clamping end is the negative direction of the x-axis, and the direction of the free end is the positive direction of the x-axis. The temperature field model of the specimen is: θ(x)=A1x 2 +A2x+A3 Wherein, Α1<0, θ is the one-dimensional temperature value, and Α2 and Α3 are fitting parameters.

9. The ultrasonic fatigue test monitoring method based on digital twin according to claim 1 is characterized in that: In step S35, the fatigue life prediction model of infrared thermal imaging is used to predict the fatigue life of the material based on the initial temperature rise slope R. The fatigue life prediction expression is as follows: N f R=C Where N f is the material fatigue life, R is the temperature rise slope, and C is the material constant.

10. The ultrasonic fatigue test monitoring method based on digital twin according to claim 1 is characterized in that: In the step S4, the parameters of each digital twin sub-model are integrated based on the Simulink visualization tool platform in MATLAB, and fully debugged to build an integrated simulation platform of multiple physical fields to form a complete digital twin model; The data related to the fatigue life of the material in S5 include ultrasonic fatigue test related data, including the end amplitude, temperature, test frequency, load amplitude, and strain of the specimen. The data related to the fatigue life of the specimen material in S5 is obtained by actual records of the control system of the ultrasonic fatigue testing machine; The method for adjusting and updating the parameters in S9 adopts an extended Kalman filter algorithm.

Images