Residual stress measurement method by combining thermal coupled simulation, ESPI and ANN

Abstract

The fusion thermal coupling simulation, ESPI and ANN residual stress measurement method belongs to the field of surface residual stress nondestructive testing. The present application combines thermal coupling finite element simulation and electronic speckle pattern interferometry experiment: first, the strain field under different residual stresses is simulated by thermal coupling simulation, and a large number of "residual stress-strain" data sets are constructed; second, the strain of the heating area is measured by using local heating combined with ESPI technology; finally, the strain data obtained by simulation and experiment are used as input, and the corresponding residual stress is used as output, the artificial neural network model is trained, and the nonlinear mapping relationship between residual stress and strain is established. In the measurement of residual stress, only the measurement point is heated locally and the strain is measured by ESPI and input into the trained model, and the residual stress of the point can be quickly inverted. The present application can realize the nondestructive and efficient detection of the surface residual stress of the complex structure.

Description

Technical Field

[0001] This invention belongs to the field of nondestructive testing of surface residual stress, and relates to a residual stress measurement method that integrates thermo-mechanical coupling simulation, ESPI and ANN. Background Technology

[0002] In the aerospace field, residual stress from machining can accelerate or inhibit damage to various components (such as fuselage frames, wing spars, and rocket fuel tanks) during service, thereby inducing or preventing fracture failure. Therefore, considering residual stress in the reliability design of components is crucial. Thus, achieving rapid, non-destructive prediction of residual stress in components is of great significance for optimizing and ensuring the service performance of aerospace equipment.

[0003] Currently, commonly used methods for residual stress measurement include neutron diffraction, X-ray diffraction, and the pinhole method. However, these methods are significantly limited in practical applications by the size and structure of the components: neutron diffraction and X-ray diffraction require sophisticated equipment, have limited detection areas, and are difficult to apply to large components; the pinhole method is a micro-destructive testing method and is not suitable for the non-destructive evaluation of in-service components. Therefore, there is an urgent need to develop a residual stress testing technology that is not limited by space and can achieve in-situ non-destructive testing, especially for measuring residual stress on the surface of large structures, which has urgent and important engineering value.

[0004] In recent years, some patented technologies have been developed for nondestructive testing, but certain limitations still exist.

[0005] Chinese invention patent CN117169461A discloses a "Method for Measuring Average Residual Stress in a Work-Hardened Layer on a Metal Surface." This method calculates the average residual stress inside the thin strip by measuring the geometric changes of the thin strip before wire cutting. However, the measurement accuracy of this method is affected by various factors such as instrument accuracy, the number of measuring points, and the deformation of the sample during the cutting process, resulting in a large measurement error for residual stress. Chinese invention patent CN118857530A discloses a "Method for Measuring Residual Stress and a Method for Predicting Part Deformation During Machining." This method introduces residual stress distribution into a simulation model, establishes a calculation model for analyzing the influence of residual stress, and conducts numerical simulation of the cutting process to simulate the release and redistribution of residual stress after machining. It then calculates the resulting workpiece deformation and ultimately predicts the machining deformation results. Chinese invention patent CN115144423B discloses a "Method and Device for Measuring Residual Stress Distribution Inside a Turbine Disk." This method combines electrolytic corrosion and X-ray diffraction (XRD) technology to measure the residual stress distribution at different depths on the surface and inside the turbine disk. However, for tenon and groove areas with complex geometries, the method is difficult to effectively measure the residual stress in these areas due to the irregular shape and poor accessibility of the working surface, thus limiting its application.

[0006] In summary, existing methods for measuring residual stress in machining are mainly limited by two factors: first, the residual stress inversion method makes it difficult to quantitatively establish the relationship between residual stress and measured strain, resulting in low accuracy of residual stress inversion; second, the geometric structure of the test sample is complex, and the reachability of the measurement probe is insufficient. Therefore, a residual stress measurement method integrating thermo-mechanical coupling simulation, ESPI, and ANN was designed. Summary of the Invention

[0007] To overcome the shortcomings of existing technologies, this invention provides a residual stress measurement method integrating thermo-coupling simulation, ESPI, and ANN, which is a method for inverting residual stress. This invention combines thermo-coupling finite element simulation with electronic speckle interferometry (ESPI) experiments: First, the strain field under different residual stresses is simulated through thermo-coupling simulation to construct a large number of "residual stress-strain" training samples; second, the strain in the heated area is measured using local heating combined with ESPI technology; finally, the strain data obtained from simulation and experiments are used as input, and the corresponding residual stress is used as output to train an artificial neural network (ANN) model, establishing a nonlinear mapping relationship between residual stress and strain. During residual stress measurement, the residual stress at the measurement point can be quickly inverted by simply heating the point locally and inputting the strain measured by ESPI into the trained ANN model, thus achieving non-destructive and efficient detection of residual stress on complex structural surfaces.

[0008] To achieve the above objectives, the technical solution adopted by the present invention is as follows: A residual stress measurement method integrating thermo-mechanical coupling simulation, ESPI and ANN is proposed and implemented based on a residual stress measurement system, including a heating system and a displacement measurement optical path system. The heating system is a heating laser used to heat the test sample.

[0009] The displacement measurement optical path system includes a first reflecting mirror, a second reflecting mirror, a first beam splitter, a first laser, a third reflecting mirror, a second beam splitter, a fourth reflecting mirror, a second laser, a CCD camera, a polarizing mirror, and a heating zone. The incident direction of the heating laser is defined as the Z-direction, the optical path directions of the first reflecting mirror, second reflecting mirror, and first beam splitter are defined as the X-direction, and the optical path directions of the third reflecting mirror, second beam splitter, and fourth reflecting mirror are defined as the Y-direction. The first reflecting mirror, second reflecting mirror, first beam splitter, and first laser form the X-direction displacement measurement optical path system, which is arranged in the XOZ plane. The third reflecting mirror, second beam splitter, fourth reflecting mirror, and second laser form the Y-direction displacement measurement optical path system, which is arranged in the YOZ plane. The front end of the heating laser perpendicularly illuminates the surface of the test sample. The plane of residual stress to be measured on the test sample lies in the XOY plane, and the circular area covered by the laser spot is the heating zone. The first beam splitter, the first laser, the second beam splitter, and the second laser are sequentially positioned behind the heating laser. The two sides of the beam splitter are provided with a first reflector and a second reflector along the X direction, and the two sides of the second beam splitter are provided with a third reflector and a fourth reflector along the Y direction. The CCD camera and the polarizing filter are positioned on the side of the sample to be tested.

[0010] The residual stress measurement method includes the following steps: Step 1: Construct a parametric model for thermo-mechanical coupling simulation based on the three-dimensional finite element model of the sample to be tested, and obtain a dataset with strain as input and residual stress as output; First, a three-dimensional finite element model of the sample to be tested is established, and the mechanical parameters of the material are added. Second, residual stresses of different sizes are added to the three-dimensional finite element model. Then, a laser heat source and heat flux density Q are added to the central region of the surface of the three-dimensional finite element model through a heating laser to simulate local heating. The radius r of the heating zone is the radius of the heating spot. A parametric modeling framework for thermo-mechanical coupling simulation is built, and calculations are submitted in batches. Finally, the thermally induced strain on the sample surface under different combinations of residual stresses is extracted through post-processing. Specifically: Step 1.1: Establish a three-dimensional finite element model of the sample to be tested, and mesh it. Assign mechanical parameters of the material to the mesh, including: material density, yield strength, elastic modulus, specific heat capacity, coefficient of thermal expansion and thermal conductivity. Step 1.2: Add residual stress to the mesh established in Step 1.1 in the form of a predefined field. By changing the magnitude of the residual stress, different combinations of residual stress are formed. Set the center of the surface of the three-dimensional finite element model in Step 1 as the heating zone, and apply heat source parameters to the heating zone to realize the parametric modeling of the thermo-mechanical coupling simulation of the three-dimensional finite element model. Step 1.3: Based on the thermo-mechanical coupling simulation parameterized model established in Step 1.2, perform transient thermo-mechanical coupling analysis to obtain the variation law of residual stress on the sample surface during heating. After the calculation is completed, extract the strain in the heating zone. Step 1.4: Pair each set of preset residual stress with its corresponding strain one by one, using strain as input and residual stress as output, to construct an input-output dataset. The dataset consists of n pairs of data. Step 2: Spray white paint on the entire surface of the test sample to enable the CCD camera to capture the light intensity change before and after the deformation of the heated area. Coaxially clamp the Z-axis of the test sample and the Z-axis of the displacement measurement optical path system. Adjust the CCD camera so that it can capture the entire heated area on the test sample. Then adjust the CCD camera to move the heated area to the center of the field of view. Step 3: After the test sample is clamped, measure the interference fringes in the X and Y directions before heating. Then, irradiate the heated area with a heating laser. After heating, turn off the heating laser and measure the interference fringes in the X and Y directions after heating. Output the light intensity of each pixel in the interference fringes before and after heating. The specific steps are as follows: Step 3.1: Before heating after the test sample is clamped: The first laser emits a 532nm wavelength laser beam, which is split into two coherent beams by the first beam splitter. These beams are reflected by the first and second mirrors, respectively, onto the test sample. After diffuse reflection, the two coherent beams first enter the polarizing mirror and then the 532nm wavelength channel in the CCD camera. By changing the deflection angle of the polarizing mirror to 0°, 90°, 180°, and 270°, the CCD camera captures the interference fringes formed by the mutual interference in the X direction at the corresponding angles, and outputs the light intensity of the corresponding pixel under the corresponding interference pattern. The 472nm wavelength laser emitted by the second laser is split into two coherent beams by the second beam splitter. These beams are reflected onto the test sample by the first and second mirrors, respectively. After diffuse reflection, the two coherent beams first enter the polarizing mirror and then the 472nm wavelength channel in the CCD camera. By changing the deflection angle of the polarizing mirror (0°, 90°, 180°, and 270°), the CCD camera captures the interference fringes formed in the Y direction at the corresponding angles, and outputs the light intensity of the corresponding pixel under the corresponding interference pattern. Step 3.2: Heat the heating area using a heating laser for 10~30s; Step 3.3: After heating, repeat step 3.1 to obtain the interference fringe patterns in the X and Y directions after heating, as well as the light intensity of the corresponding pixel under the corresponding interference pattern; Step 4: Combining the light intensity data obtained from the CCD camera in Steps 3.1 and 3.3 at 0°, 90°, 180°, and 270°, the phase of each pixel is obtained using a four-step phase-shifting method. The phase difference between each pixel before and after heating is then calculated. The specific calculation steps are as follows: Step 4.1: Calculate the phase of each pixel before and after heating using the four-step phase-shifting method, as shown in the following formula: (1) in, This indicates that the phase of each pixel is obtained by a four-step phase-shifting method before and after heating. I 1 represents the light intensity corresponding to each pixel at 0°; I 2 represents the light intensity corresponding to each pixel at 90°; I 3 represents the light intensity corresponding to each pixel at 180°. I 3 represents the light intensity corresponding to each pixel at 270°; Step 4.2: Subtract the phase difference before and after heating to calculate the phase difference of each pixel before and after heating. The calculation formula is as follows: (2) Where, ΔΦ( x , y ) represents the phase difference between each pixel before and after heating, Φ 加热后 ( x , y ) represents the phase of each pixel obtained by the four-step phase-shifting method after heating, Φ 加热前 ( x , y () indicates the phase of each pixel obtained by the phase shifting method in the first four steps of heating; Step 5: Based on the phase difference of each pixel calculated in Step 4.2, establish the X-direction displacement. u and displacement in the Y direction v The relationship between phase difference and phase difference is calculated using the following formula: (3) in, u Indicates displacement in the X direction. v Indicates displacement in the Y direction. i This indicates the incident angle of the laser beam from the first or second laser. l Indicates the wavelength of the laser light from the first or second laser; Step 6: Combining formula (3), by... u and v Perform differentiation, calculate the difference step size in the X and Y directions, then calculate the phase difference change in the X and Y direction difference step size directions, and finally obtain the strain components in the X and Y directions. Step 6.1: For u and v The specific formula for differentiation is as follows: (4) in, e res,x This represents the strain caused by residual stress in the X direction. e res,y ΔΦ represents the strain caused by residual stress in the Y direction. R ( x , y ) represents the change in phase difference along the X-direction differential step size, ΔΦ G ( x , y ) represents the change in phase difference along the Y-direction differential step size, Δ x and Δ y This represents the difference step size in the X and Y directions. l R and l G This indicates the laser wavelengths corresponding to the first and second lasers. i R and i G This indicates the incident angle of the first and second lasers relative to the test sample; Step 6.2: Calculate the difference step size in the X and Y directions. The specific formula is as follows: (5) (6) in, n x This indicates the number of pixels during the X-direction differencing process. n y This indicates the number of pixels during the Y-direction differencing process. L This indicates the field of view of a CCD camera. W This indicates the field of view width of a CCD camera. N x This indicates the X-axis resolution of the CCD camera. N y This indicates the resolution in the Y direction of the CCD camera; Step 6.3: Calculate the phase difference change in the X and Y directions along the differential step size. The specific formula is as follows: (7) (8) Where, ΔΦ(Δ x , y ) represents the phase difference of the corresponding pixels after the X-direction differentiation, ΔΦ( x ,Δ y () represents the phase difference of the corresponding pixels after differential analysis in the Y direction; Step 6.4: Substitute formulas (5~8) into formula (4) to calculate the strain in the X and Y directions; Step 7: Use the XRD stress measurement method to measure the residual stress on the surface of the test sample. Based on the strain calculated in Step 6.4, pair each group of measured residual stress with its corresponding strain. Use strain as input and residual stress as output to construct an input-output dataset. The dataset contains a total of m pairs of data. Step 8: Based on the n pairs of data obtained through parametric modeling in Step 1 and the m pairs of data obtained through experimental measurement in Step 7, combine them into z pairs of data, where z = m + n, and establish a total dataset with strain as input and residual stress as output. Take 0.75 z pairs as the training set and 0.25 z pairs as the validation set. Step 9: Based on the dataset established in Step 8, a residual stress prediction model is established using an ANN neural network as the basic architecture. The residual stress prediction model is trained using the training set and validated using the validation set. Step 10: Select any test sample with unknown surface residual stress, measure the strain caused by local heating through steps 4-6, and input it into the residual stress prediction model trained in step 9 to measure the residual stress on the surface of the test sample.

[0011] Furthermore, in the X-direction displacement measurement optical path system: the laser emitted from the first laser passes through the first beam splitter and is divided into two coherent beams of equal energy. These beams are reflected by the first and second mirrors onto the entire surface of the test sample. The coherent light in the observation area of ​​the CCD camera on the test sample surface is diffusely reflected, first entering the polarizing mirror and then the target surface in the CCD camera to form an interference fringe pattern in the X direction.

[0012] Furthermore, in the optical path system for measuring displacement in the Y direction: the laser emitted from the second laser passes through a beam splitter and is divided into two coherent beams of equal energy. These beams are reflected by the third and fourth mirrors onto the entire surface of the test sample, respectively. The coherent light in the observation area of ​​the CCD camera on the surface of the test sample is diffusely reflected, first entering the polarizing mirror and then the target surface in the CCD camera to form an interference fringe pattern in the Y direction.

[0013] Furthermore, the XOZ plane where the X-direction displacement measurement optical path system is located is at a 90° angle to the YOZ plane where the Y-direction displacement measurement optical path system is located. The interference fringe patterns in the X-direction and Y-direction are perpendicular to each other.

[0014] Compared with the prior art, the present invention has the following beneficial effects: (1) This invention uses artificial neural networks to train the sample data generated by finite element simulation to establish a mapping model between residual stress and thermal strain, which can reduce the complex theoretical derivation and parameter solution process and improve the efficiency of residual stress inversion.

[0015] (2) After the ANN residual stress prediction model established by this invention is trained, the residual stress result can be obtained quickly by only inputting the strain information measured by ESPI, which significantly improves the residual stress detection efficiency and is suitable for online detection and rapid evaluation.

[0016] (3) This invention uses ESPI measurement to obtain the full-field strain distribution in the X and Y directions, and combines it with a machine learning model to synchronously invert the residual stress in the two orthogonal directions, thereby realizing the quantitative measurement of the surface bidirectional residual stress and improving the integrity of the residual stress characterization. Attached Figure Description

[0017] Figure 1 The strain measurement system of the present invention; Figure 2 This invention relates to a method for measuring residual stress. Figure 3 The thermal coupling simulation is performed on a three-dimensional finite element model, where (a) is the three-dimensional finite element model and (b) is the heat flux density distribution diagram of the Gaussian heat source. Figure 4 The results of the thermo-mechanical coupling simulation of the three-dimensional finite element model are shown, where (a) is the simulated temperature distribution, (b) is the simulated strain, and (c) is the simulated stress. Figure 5 The network architecture for the residual stress prediction model; In the diagram: 1 heating laser, 2 first reflecting mirror, 3 second reflecting mirror, 4 first beam splitter, 5 first laser, 6 third reflecting mirror, 7 second beam splitter, 8 fourth reflecting mirror, 9 second laser, 10 CCD camera, 11 polarizing mirror, 12 heating zone. Detailed Implementation

[0018] The technical solutions in the design process of this invention will be clearly and completely described below in conjunction with the design concept of this invention.

[0019] This embodiment provides a residual stress measurement method that integrates thermo-mechanical coupling simulation, ESPI, and ANN, based on... Figure 1The strain measurement system shown includes a heating system and a displacement measurement optical path system; specifically: The heating system is a heating laser 1, which is used to heat the test sample.

[0020] The displacement measurement optical path system includes a first reflecting mirror 2, a second reflecting mirror 3, a first beam splitter 4, a first laser 5, a third reflecting mirror 6, a second beam splitter 7, a fourth reflecting mirror 8, a second laser 9, a CCD camera 10, a polarizing mirror 11, and a heating zone 12. The incident direction of the heating laser 1 is defined as the Z-direction, the optical path directions of the first reflecting mirror 2, the second reflecting mirror 3, and the first beam splitter 4 are defined as the X-direction, and the optical path directions of the third reflecting mirror 6, the second beam splitter 7, and the fourth reflecting mirror 8 are defined as the Y-direction. The first reflecting mirror 2, the second reflecting mirror 3, the first beam splitter 4, and the first laser 5 form the X-direction displacement measurement optical path system, which is arranged in the XOZ plane. The third reflecting mirror 6, the second beam splitter 7, the fourth reflecting mirror 8, and the second laser 9 form the Y-direction displacement measurement optical path system, which is arranged in the YOZ plane. The front end of the heating laser 1 vertically illuminates the surface of the test sample. The plane of the residual stress to be measured of the test sample is in the XOY plane, and the circular area covered by the laser spot is the heating area 12. The first beam splitter 4, the first laser 5, the second beam splitter 7, and the second laser 9 are sequentially arranged behind the heating laser 1. The beam splitter 4 has a first reflector 2 and a second reflector 3 arranged on both sides along the X direction, and the second beam splitter 7 has a third reflector 6 and a fourth reflector 8 arranged on both sides along the Y direction. The CCD camera 10 and the polarizing filter 11 are arranged on the side of the sample to be tested.

[0021] In the X-direction displacement measurement optical path system: the laser emitted from the first laser 5 passes through the first beam splitter prism 4 and is split into two coherent beams of equal energy. These beams are reflected by the first reflector 2 and the second reflector 3 onto the entire surface of the test sample. The coherent light in the observation area of ​​the CCD camera 10 on the surface of the test sample is diffusely reflected and first enters the polarizer 11 before entering the target surface in the CCD camera 10 to form an interference fringe pattern in the X direction.

[0022] In the optical path system for measuring displacement in the Y direction: the laser emitted from the second laser 9 is split into two coherent beams of equal energy by the beam splitter prism 7, and is reflected onto the entire surface of the test sample by the third reflector 6 and the fourth reflector 8, respectively. The coherent light in the observation area of ​​the CCD camera 10 on the surface of the test sample is diffusely reflected and first enters the polarizer 11 and then enters the target surface in the CCD camera 10 to form an interference fringe pattern in the Y direction.

[0023] Furthermore, the XOZ plane where the X-direction displacement measurement optical path system is located is at a 90° angle to the YOZ plane where the Y-direction displacement measurement optical path system is located. The interference fringe patterns in the X-direction and Y-direction are perpendicular to each other.

[0024] Figure 2 The residual stress measurement method of the present invention, based on the above-mentioned strain measurement system, includes the following steps: Step 1: First, establish a three-dimensional finite element model of the sample to be tested and add the mechanical parameters of the material. Second, add residual stresses of different sizes in the model. Then, add a laser heat source (gaussian heat source is used as an example in this invention) in the central area of ​​the model surface to realize the simulation of local heating. Based on the above settings, build a parametric modeling framework for thermo-mechanical coupling simulation and submit the calculation in batches. Finally, extract the thermal strain on the sample surface under different combinations of residual stresses through post-processing. Step 1.1: Establish a three-dimensional finite element model of the sample to be tested, such as... Figure 3 As shown in (a), the material is meshed and then assigned mechanical parameters to the mesh, including: material density, yield strength, elastic modulus, specific heat capacity, coefficient of thermal expansion and thermal conductivity. Step 1.2: Add residual stress to the mesh established in Step 1.1 in the form of a predefined field. By changing the magnitude of the residual stress, different combinations of residual stress are formed. Set the center of the surface of the 3D finite element model in Step 1 as the heating region, the radius r of the heating region as the radius of the heating spot, and apply Gaussian heat source parameter Q to the heating region. The heat flux density of the Gaussian heat source is as follows: Figure 3 (b) shows that the thermo-mechanical coupling simulation parameter model of the three-dimensional finite element model of the test sample is realized; Step 1.3: Submit the thermo-mechanical coupling simulation parameter model established in Step 1.2, perform transient thermo-mechanical coupling analysis to obtain the variation law of residual stress on the sample surface during heating, and after the calculation is completed, extract the strain of the heating area (i.e., area 12 defined in Step 1.2); Step 1.4: Pair each set of preset residual stress with its corresponding strain one by one, and organize them into a dataset of "input (strain) - output (residual stress)", with a total of n pairs of data in the dataset; Step 2: Spray white paint on the entire surface of the test sample, so that the CCD camera 10 can capture the light intensity change before and after the deformation of the heating area 12. The Z-axis of the test sample and the Z-axis of the displacement measurement optical path system are coaxially clamped. The CCD camera 10 is adjusted so that it can capture the complete heating area 12 on the test sample. The CCD camera 10 is then adjusted to move the heating area 12 to the center of the field of view. Step 3: After the test sample is clamped, measure the interference fringes in the X and Y directions before heating. Then, irradiate the heating area 12 with heating laser 1. After heating, turn off heating laser 1 and measure the interference fringes in the X and Y directions after heating. Output the light intensity of each pixel in the interference fringes before and after heating. The specific steps are as follows: Step 3.1: After the test sample is clamped and before heating, the 532nm wavelength laser emitted by the first laser 5 is split into two coherent beams by the beam splitter 4. These beams are reflected onto the test sample by the first reflecting mirror 2 and the second reflecting mirror 3, respectively. After diffuse reflection, the two coherent beams first enter the polarizer 11 and then the 532nm wavelength channel in the CCD camera 10. The deflection angle of the polarizer 11 is changed to 0°, 90°, 180°, and 270°. At the corresponding angles, the CCD camera 10 captures the interference fringe patterns formed by the mutual interference in the X direction and outputs the corresponding interference pattern. The light intensity of the corresponding pixel is as follows: The laser with a wavelength of 472nm emitted by the second laser 9 is split into two coherent beams by the second beam splitter 7. These beams are reflected onto the test sample by the first reflector 6 and the second reflector 8, respectively. The two coherent beams first enter the polarizer 11 after diffuse reflection and then enter the 472nm channel of the CCD camera 10. The deflection angle of the polarizer 11 is changed to 0°, 90°, 180° and 270°. The CCD camera 10 captures the interference fringe pattern formed by mutual interference in the Y direction at the corresponding angles and outputs the light intensity of the corresponding pixel under the corresponding interference pattern. Step 3.2: Use heating laser 1 to heat heating area 12 for 20 seconds at a heating temperature of 52℃; Step 3.3: After heating, repeat step 3.1 to obtain the interference fringe patterns in the X and Y directions after heating, as well as the light intensity of the corresponding pixel under the corresponding interference pattern; Step 4: Combining the light intensity data obtained from CCD camera 10 at 0°, 90°, 180°, and 270° in Steps 3.1 and 3.3, the phase corresponding to each pixel is obtained using a four-step phase-shifting method. The phase difference between each pixel before and after heating is then calculated. The specific calculation steps are as follows: Step 4.1: Calculate the phase of each pixel before and after heating using the four-step phase-shifting method, as shown in the following formula: (1) in, This indicates that the phase of each pixel is obtained by a four-step phase-shifting method before and after heating. I 1 represents the light intensity corresponding to each pixel at 0°; I 2 represents the light intensity corresponding to each pixel at 90°; I 3 represents the light intensity corresponding to each pixel at 180°.I 3 represents the light intensity corresponding to each pixel at 270°; Step 4.2: Subtract the phase difference before and after heating to calculate the phase difference of each pixel before and after heating. The calculation formula is as follows: (2) Where, ΔΦ( x , y ) represents the phase difference between each pixel before and after heating, Φ 加热后 ( x , y ) represents the phase of each pixel obtained by the four-step phase-shifting method after heating, Φ 加热前 ( x , y () indicates the phase of each pixel obtained by the phase shifting method in the first four steps of heating; Step 5: Based on the phase difference of each pixel calculated in Step 4.2, establish the X-direction displacement. u and displacement in the Y direction v The relationship between phase difference and phase difference is calculated using the following formula: (3) in, u Indicates displacement in the X direction. v Indicates displacement in the Y direction. i This indicates the incident angle of the laser from either the first laser 5 or the second laser 9. l This indicates the wavelength of the laser beam from either the first laser 5 or the second laser 9. Step 6: Combining formula (3), by... u and v Perform differentiation, calculate the difference step size in the X and Y directions, then calculate the phase difference change in the X and Y direction difference step size directions, and finally obtain the strain components in the X and Y directions. Step 6.1: For u and v The specific formula for differentiation is as follows: (4) in, e res,x This represents the strain caused by residual stress in the X direction. e res,y ΔΦ represents the strain caused by residual stress in the Y direction. R ( x , y ) represents the change in phase difference along the X-direction differential step size, ΔΦ G ( x , y ) represents the change in phase difference along the Y-direction differential step size, Δ x and Δy This represents the difference step size in the X and Y directions. l R and l G This indicates the laser wavelengths corresponding to the first laser 5 and the second laser 9. i R and i G This indicates the incident angle of the first laser 5 and the second laser 9 relative to the sample under test; Step 6.2: Calculate the difference step size in the X and Y directions. The specific formula is as follows: (5) (6) in, n x This indicates the number of pixels during the X-direction differencing process. n y This indicates the number of pixels during the Y-direction differencing process. L This indicates the field of view of a CCD camera. W This indicates the field of view width of a CCD camera. N x This indicates the X-axis resolution of the CCD camera. N y This indicates the resolution in the Y direction of the CCD camera; Step 6.3: Calculate the phase difference change in the X and Y directions along the differential step size. The specific formula is as follows: (7) (8) Where, ΔΦ(Δ x , y ) represents the phase difference of the corresponding pixels after the X-direction differentiation, ΔΦ( x ,Δ y () represents the phase difference of the corresponding pixels after differential analysis in the Y direction; Step 6.4: Substitute formulas (5~8) into formula (4) to calculate the strain in the X and Y directions; Step 7: Using the XRD stress measurement method, the residual stress in the X and Y directions is measured by rotating the test sample by 90°. Based on the strain calculated in Step 6.4, each set of measured residual stress is paired with its corresponding strain and organized into a dataset of "input (strain) - output (residual stress)". The dataset contains a total of m pairs of data. Step 8: Based on the n pairs of data obtained through parametric modeling in Step 1 and the m pairs of data obtained through experimental measurement in Step 7, they are combined into z pairs of data, where z = m + n. In this implementation case, 4000 pairs of data are obtained from simulation in Step 1 and 1000 pairs of data are obtained from experiment in Step 7, for a total of 5000 pairs of residual stress and strain mapping data. A total dataset with strain as input and residual stress as output is established, and 3750 pairs are taken as the training set and 1250 pairs are taken as the validation set. Step 9: Based on the dataset established in Step 8, a residual stress prediction model is established using an ANN neural network as the basic architecture. The residual stress prediction model is trained using the training set and validated using the validation set. Step 10: Randomly select a test sample with unknown surface residual stress. Measure the strain caused by local heating through steps 4-6, and input it into the residual stress prediction model trained in step 9 to measure the residual stress on the surface of the test sample. The residual stress in the X direction is -316 MPa, and the residual stress in the Y direction is -296 MPa.

[0025] The above embodiments are merely illustrative of the implementation methods of the present invention, but should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the protection scope of the present invention.

Claims

1. A residual stress measurement method integrating thermo-mechanical coupling simulation, ESPI, and ANN, characterized in that, The residual stress measurement method is based on a heated laser (1) and a displacement measurement optical path system. The displacement measurement optical path system includes a first reflecting mirror (2), a second reflecting mirror (3), a first beam splitter (4), a first laser (5), a third reflecting mirror (6), a second beam splitter (7), a fourth reflecting mirror (8), a second laser (9), a CCD camera (10), and a polarizing mirror (11). The method includes the following steps: Step 1: Construct a parametric model for thermo-mechanical coupling simulation based on the three-dimensional finite element model of the sample to be tested, and obtain a dataset with strain as input and residual stress as output; First, a three-dimensional finite element model of the sample to be tested is established, and material mechanical parameters are added; second, residual stresses of different sizes are added to the three-dimensional finite element model; then, local heating is simulated in the central region of the surface of the three-dimensional finite element model by heating laser (1), and a parameterized model of thermo-mechanical coupling simulation is constructed; finally, the strain of the sample surface under different residual stress combinations is extracted to obtain a dataset with strain as input and residual stress as output, totaling n pairs of data; Step 2: Spray white paint on the entire surface of the test sample, so that the CCD camera (10) can capture the light intensity change before and after the deformation of the heating area (12). The Z-axis of the test sample and the Z-axis of the displacement measurement optical path system are coaxially clamped. The heating area (12) on the complete test sample is captured by the CCD camera (10), and the heating area (12) is moved to the center of the field of view. Step 3: First, measure the interference fringe patterns in the X and Y directions before heating; then, irradiate the heating area (12) with the heating laser (1) and measure the interference fringe patterns in the X and Y directions after heating; finally, output the light intensity of each pixel in the interference fringe patterns before and after heating. Step 4: Based on the light intensity obtained in Step 3, obtain the phase corresponding to each pixel point through the four-step phase shift method, and calculate the phase difference of each pixel point before and after heating; Step 5: Establish the X-direction displacement based on the phase difference of each pixel. u and displacement in the Y direction v The relationship with phase difference; Step 6: By analyzing... u and v Perform differentiation, calculate the difference step size in the X and Y directions, and then calculate the phase difference change in the X and Y direction difference step size directions to obtain the strain components in the X and Y directions. Step 7: Use the XRD stress measurement method to measure the residual stress on the surface of the test sample. Based on the strain components in Step 6, pair each set of measured residual stress with its corresponding strain. Use strain as input and residual stress as output to construct an input-output dataset, for a total of m pairs of data. Step 8: Based on the n pairs of data obtained through parametric modeling in Step 1 and the m pairs of data obtained through experimental measurement in Step 7, combine them into z pairs of data, where z = m + n. Establish a total dataset with strain as input and residual stress as output, and divide it into training set and validation set. Step 9: Using an ANN neural network as the basic architecture, establish a residual stress prediction model, train the residual stress prediction model through the training set, and validate the residual stress prediction model through the validation set. Step 10: Select any test sample with unknown surface residual stress, measure the strain caused by local heating through steps 4 to 6, and input it into the residual stress prediction model trained in step 9 to obtain the residual stress on the surface of the test sample.

2. The residual stress measurement method integrating thermo-mechanical coupling simulation, ESPI, and ANN as described in claim 1, is characterized in that, The incident direction of the heating laser (1) is defined as the Z direction, the optical path direction of the first reflector (2), the second reflector (3), and the first beam splitter (4) is defined as the X direction, and the optical path direction of the third reflector (6), the second beam splitter (7), and the fourth reflector (8) is defined as the Y direction. The first reflector (2), the second reflector (3), the first beam splitter (4), and the first laser (5) form an X-direction displacement measurement optical path system, which is arranged in the XOZ plane. The third reflector (6), the second beam splitter (7), the fourth reflector (8), and the second laser (9) form a Y-direction displacement measurement optical path system, which is arranged in the YOZ plane. The front end of the heating laser (1) vertically illuminates the surface of the test sample. The plane of residual stress to be measured on the test sample is in the XOY plane, and the circular area covered by the light spot is the heating area (12). The first beam splitter (4), the first laser (5), the second beam splitter (7), and the second laser (9) are arranged in sequence behind the heating laser (1). The first beam splitter (4) has a first reflector (2) and a second reflector (3) on both sides along the X direction. The second beam splitter (7) has a third reflector (6) and a fourth reflector (8) on both sides along the Y direction. The CCD camera (10) and the polarizer (11) are arranged on the side of the test sample.

3. The residual stress measurement method integrating thermo-mechanical coupling simulation, ESPI, and ANN as described in claim 2, is characterized in that, In the X-direction displacement measurement optical path system: the laser emitted from the first laser (5) is split into two coherent beams of equal energy by the first beam splitter (4), and reflected by the first mirror (2) and the second mirror (3) to the entire surface of the sample to be tested. The coherent light in the observation area of ​​the CCD camera (10) on the sample to be tested is diffusely reflected and first enters the polarizer (11) and then enters the target surface of the CCD camera (10) to form an interference fringe pattern in the X direction. In the Y-direction displacement measurement optical path system: the laser emitted from the second laser (9) is split into two coherent beams of equal energy by the second beam splitter (7), and reflected by the third mirror (6) and the fourth mirror (8) to the entire surface of the sample to be tested. The coherent light in the observation area of ​​the CCD camera (10) on the sample to be tested is diffusely reflected and first enters the polarizer (11) and then enters the target surface of the CCD camera (10) to form an interference fringe pattern in the Y direction.

4. The residual stress measurement method integrating thermo-mechanical coupling simulation, ESPI, and ANN as described in claim 2, is characterized in that, The XOZ plane where the X-direction displacement measurement optical path system is located is distributed at 90° with the YOZ plane where the Y-direction displacement measurement optical path system is located; the interference fringe pattern in the X direction is perpendicular to the interference fringe pattern in the Y direction.

5. The residual stress measurement method integrating thermo-mechanical coupling simulation, ESPI, and ANN as described in claim 4, is characterized in that... Specifically, step 1 is as follows: Step 1.1: Establish a three-dimensional finite element model of the sample to be tested, and mesh it. Assign mechanical parameters of the material to the mesh, including: material density, yield strength, elastic modulus, specific heat capacity, coefficient of thermal expansion and thermal conductivity. Step 1.2: Add residual stress in the form of a predefined field on the mesh established in Step 1.

1. By changing the magnitude of the residual stress, different combinations of residual stress are formed. Set the center of the surface of the three-dimensional finite element model in Step 1 as the heating zone (12) and apply heat source parameters to the heating zone (12) to construct the thermo-mechanical coupling simulation parameterized model of the three-dimensional finite element model. Step 1.3: Perform transient thermo-mechanical coupling analysis based on the thermo-mechanical coupling simulation parameterized model to obtain the variation law of residual stress on the sample surface during the heating process. After the calculation is completed, extract the strain of the heating zone (12). Step 1.4: Pair each set of preset residual stresses with their corresponding strains one by one, using strain as input and residual stress as output, to construct an input-output dataset. The dataset consists of n pairs of data.

6. The residual stress measurement method integrating thermo-mechanical coupling simulation, ESPI, and ANN as described in claim 5, is characterized in that, Step 3 specifically includes: Step 3.1: Measure the interference fringes in the X and Y directions before heating; The laser with a wavelength of 532nm emitted by the first laser (5) is split into two coherent beams by the first beam splitter (4). The beams are reflected onto the test sample by the first reflector (2) and the second reflector (3), respectively. The two coherent beams enter the polarizer (11) through diffuse reflection and then enter the 532nm channel of the CCD camera (10). The deflection angle of the polarizer (11) is changed. The deflection angles are 0°, 90°, 180° and 270°, respectively. The CCD camera (10) captures the interference fringe pattern formed by mutual interference in the X direction at the corresponding angles, and outputs the light intensity of the corresponding pixel point under the corresponding interference pattern. The laser with a wavelength of 472nm emitted by the second laser (9) is split into two coherent beams by the second beam splitter (7). The beams are reflected onto the test sample by the first mirror (2) and the second mirror (3), respectively. The two coherent beams enter the polarizer (11) after diffuse reflection and then enter the 472nm channel of the CCD camera (10). The deflection angle of the polarizer (11) is changed. The deflection angles are 0°, 90°, 180° and 270°, respectively. The CCD camera (10) captures the interference fringe pattern formed by mutual interference in the Y direction at the corresponding angles, and outputs the light intensity of the corresponding pixel point under the corresponding interference pattern. Step 3.2: Heat the heating area (12) with a heating laser (1) for 10~30s; Step 3.3: After heating, repeat step 3.1 to obtain the interference fringe patterns in the X and Y directions after heating, as well as the light intensity of the corresponding pixel under the corresponding interference pattern.

7. The residual stress measurement method integrating thermo-mechanical coupling simulation, ESPI, and ANN as described in claim 6, is characterized in that, Step 4 specifically includes: Step 4.1: Calculate the phase of each pixel before and after heating using the four-step phase-shifting method, as shown in the following formula: in, This indicates that the phase of each pixel is obtained by a four-step phase-shifting method before and after heating. I 1 represents the light intensity corresponding to each pixel at 0°; I 2 represents the light intensity corresponding to each pixel at 90°; I 3 represents the light intensity corresponding to each pixel at 180°. I 3 represents the light intensity corresponding to each pixel at 270°; Step 4.2: Subtract the phase difference before and after heating to calculate the phase difference of each pixel before and after heating. The calculation formula is as follows: Where, ΔΦ( x , y ) represents the phase difference between each pixel before and after heating, Φ 加热后 ( x , y ) represents the phase of each pixel obtained by the four-step phase-shifting method after heating, Φ 加热前 ( x , y ) represents the phase of each pixel obtained by the phase shifting method in the first four steps before heating.

8. The residual stress measurement method integrating thermo-mechanical coupling simulation, ESPI, and ANN as described in claim 7, is characterized in that, In step 5, the displacement in the X direction u and displacement in the Y direction v The relationship between phase difference and phase difference is calculated using the following formula: in, u Indicates displacement in the X direction. v Indicates displacement in the Y direction. θ This indicates the incident angle of the laser from the first laser (5) or the second laser (9). λ This indicates the wavelength of the laser from the first laser (5) or the second laser (9).

9. The residual stress measurement method integrating thermo-mechanical coupling simulation, ESPI, and ANN as described in claim 8, is characterized in that, Specifically, step 6 includes: Step 6.1: For u and v The specific formula for differentiation is as follows: in, ε res,x This represents the strain caused by residual stress in the X direction. ε res,y ΔΦ represents the strain caused by residual stress in the Y direction. R ( x , y ) represents the change in phase difference along the X-direction differential step size, ΔΦ G ( x , y ) represents the change in phase difference along the Y-direction differential step size, Δ x and Δ y This represents the difference step size in the X and Y directions. λ R and λ G This indicates the laser wavelengths corresponding to the first laser (5) and the second laser (9). θ R and θ G This indicates the incident angle of the first laser (5) and the second laser (9) relative to the sample under test; Step 6.2: Calculate the difference step size in the X and Y directions. The specific formula is as follows: in, n x This indicates the number of pixels during the X-direction differencing process. n y This indicates the number of pixels during the Y-direction differencing process. L This indicates the field of view of a CCD camera. W This indicates the field of view width of a CCD camera. N x This indicates the X-axis resolution of the CCD camera. N y This indicates the resolution in the Y direction of the CCD camera; Step 6.3: Calculate the phase difference change in the X and Y directions along the differential step size. The specific formula is as follows: Where, ΔΦ(Δ x , y ) represents the phase difference of the corresponding pixels after the X-direction differentiation, ΔΦ( x ,Δ y () represents the phase difference of the corresponding pixels after differential analysis in the Y direction; Step 6.4: Substitute the formulas from Step 6.2 and Step 6.3 into the formula from Step 6.1 to calculate the strain in the X and Y directions.

Images