A method and system for evaluating the mechanical properties of a laminate based on strain data

Abstract

The present application relates to the technical field of mechanical property testing, and more particularly to a method and system for evaluating the mechanical properties of a laminate based on strain data. The method comprises the following steps: performing a biaxial tensile test on a laminate sample, simultaneously collecting load time history data and full-field speckle images, constructing a full-field strain matrix and discretizing it into nodes using a mapping algorithm, dividing the core pure shear zone and stress diffusion zone, obtaining the first principal strain through principal component analysis, calculating the load transfer efficiency attenuation spectrum and the full-field strain energy entropy density, respectively, combining the preset shape feature parameters to construct a dynamic correction model, and using the model to calculate the correction coefficient to obtain the true constitutive curve by inverting the nominal stress. The introduction of thermodynamic entropy characteristics solves the problem of stress field distortion in the nonlinear damage stage that is difficult to perceive by traditional methods, effectively improving the accuracy of biaxial mechanical property testing of laminates.

Description

Technical Field

[0001] This invention relates to the field of mechanical property testing technology, and in particular to a method and system for evaluating the mechanical properties of laminates based on strain data. Background Technology

[0002] Laminate structures are widely used in aerospace, new energy vehicles, and high-end equipment manufacturing due to their excellent specific strength, specific modulus, and designability. To accurately characterize the mechanical behavior of laminates under complex stress states, biaxial tensile testing is a crucial testing method. In biaxial tensile testing, a cruciform specimen is typically used to apply biaxial loads in the plane. However, this specific geometry, while achieving biaxial loading, also introduces testing errors, mainly manifested in geometric effects and material damage.

[0003] First, regarding geometric effects, the cruciform specimen must transfer the load through four loading arms. Due to the complex stress transfer path at the interface between the loading arms and the central region, part of the load is transferred through a shear hysteresis mechanism, resulting in the actual effective stress in the central test area being less than the nominal stress applied at the fixture end. This shear hysteresis effect makes the directly measured nominal stress-strain curve unable to accurately represent the true constitutive relationship of the laminate. Existing techniques typically use a static correction factor based on the linear elastic assumption to compensate for this error, assuming a fixed proportional relationship between the nominal stress and the actual stress.

[0004] Secondly, regarding material damage, laminates exhibit nonlinear damage evolution characteristics. In the initial loading stage, the material is in a linear elastic phase, and the static correction coefficient is still approximately applicable. However, as the load increases, microcracks in the matrix, fiber breakage, and interlaminar debonding gradually develop within the laminate. The accumulation of these damages leads to nonlinear degradation of stiffness, causing distortion and redistribution of the stress field within the specimen, altering the initial load transfer path. At this point, the static correction coefficient determined based on the initial undamaged geometry becomes ineffective. Existing correction methods struggle to quantify this dynamic change in the stress field caused by internal damage, resulting in increased testing errors in the nonlinear stage after yielding, affecting the accurate assessment of the laminate's ultimate bearing capacity.

[0005] Therefore, developing a method that can dynamically correct for geometric transfer efficiency and material damage state is a technical problem that needs to be solved in the field of laminate mechanical testing. Summary of the Invention

[0006] To address the problem of low testing accuracy caused by the inability to adapt to nonlinear damage and shear hysteresis effects in laminates, this invention provides a method and system for evaluating the mechanical properties of laminates based on strain data.

[0007] In a first aspect, the present invention provides a method for evaluating the mechanical properties of laminates based on strain data, employing the following technical solution:

[0008] A method for evaluating the mechanical properties of laminates based on strain data, comprising the following steps:

[0009] Biaxial tensile tests were conducted on laminate specimens, and load time history data and full-field speckle images were acquired simultaneously. A full-field strain matrix was constructed using a mapping algorithm, and the full-field strain matrix was discretized into several nodes. Based on the nodes, the region was divided into a core pure shear zone and a stress diffusion zone.

[0010] Principal component analysis is performed on each node to obtain the first principal strain along the first principal axis; based on the difference in the first principal strain gradient between the stress diffusion region and the core pure shear region, the load transfer efficiency attenuation spectrum characterizing the load transfer fidelity is calculated.

[0011] Based on the strain energy density distribution weights of each node within the core pure shear zone, the full-field strain energy entropy density, which characterizes the orderliness of the stress field, is calculated.

[0012] Retrieve preset shape feature parameters, use the shape feature parameters to weight and couple the load transfer efficiency attenuation spectrum and the full-field strain energy entropy density to construct a dynamic correction model, and use the dynamic correction model to calculate the correction coefficient.

[0013] The nominal stress is corrected using the aforementioned correction factor to obtain the true constitutive curve, thus completing the mechanical property evaluation of the laminate.

[0014] This invention achieves standardized processing of experimental raw data by constructing a full-field strain matrix and dividing it into regions; by calculating the load transfer efficiency attenuation spectrum and the full-field strain energy entropy density, it achieves separate measurement and dynamic decoupling of geometric errors and material damage characteristics, solving the problem that traditional static correction cannot adapt to the nonlinear stage, thereby improving the accuracy of the evaluation.

[0015] Preferably, the step of constructing the full-field strain matrix through a mapping algorithm includes:

[0016] Displacement field calculations were performed on the acquired full-field speckle images to obtain unstructured scattered strain data.

[0017] A bilinear interpolation algorithm based on shape functions is used to map the scattered strain data onto a regular grid of integral points, forming a full-field strain matrix that includes the spatial location of nodes and the time dimension.

[0018] Preferably, the load transfer efficiency attenuation spectrum satisfies the following relationship:

[0019]

[0020] In the formula, Indicates time The load transfer efficiency attenuation spectrum; Represents an exponential function with the natural constant as its base; This represents the set of nodes in the stress diffusion region; This represents the set of nodes representing the core pure pruning region; Indicating the stress diffusion region in the first... Each node at time... The first principal strain value along the first principal axis; Indicates the first core pure shear region Each node at time... The first principal strain value along the first principal axis; Indicates time The average first principal strain value of all nodes within the core pure shear zone.

[0021] This invention uses the regional integral ratio to measure the strain gradient between the stress diffusion zone and the core zone, thereby characterizing the degree of energy dissipation during load transfer and dynamically reflecting the changes in transfer efficiency caused by the geometric deformation of the specimen.

[0022] Preferably, the total strain energy entropy density satisfies the following relationship:

[0023]

[0024] In the formula, Indicates time The full-field strain energy entropy density; This represents the total number of nodes within the core pure scissor region; Represents the natural logarithm function; Indicates time Within the core pure shearing zone, the first The strain energy density distribution weight of the nth node, whose value is the nth node. Each node at time... strain energy density and time The ratio of the total strain energy density in the core pure shear region.

[0025] This invention uses the full-field strain energy entropy density to describe the degree of order of the stress field in the core region. When the material is damaged, the stress field will change from uniform to non-uniform. This index can capture this change, thus providing a data basis for nonlinear damage correction.

[0026] Preferably, the correction coefficients calculated using the dynamic correction model satisfy the following relationship:

[0027]

[0028] In the formula, Indicates time The correction factor; Represents the geometric sensitivity constant; This represents the load transfer efficiency attenuation spectrum; This represents the Poisson sensitivity constant; This represents the total strain energy entropy density. The load normalization weight is represented by the geometric sensitivity constant and the Poisson sensitivity constant, which are the shape feature parameters.

[0029] This invention effectively decouples geometric transfer characteristics and material damage characteristics, thereby achieving accurate correction throughout the entire loading process.

[0030] Preferably, the method for obtaining the load normalization weights is as follows:

[0031] Calculate the deviation of the load value relative to the values ​​in the load sequence;

[0032] The load normalization weight is calculated using the Sigmoid function based on the deviation, so that the load normalization weight increases non-linearly with the increase of load.

[0033] Preferably, the step of correcting the nominal stress using the correction coefficient to obtain the true constitutive curve includes:

[0034] Calculate the product of the thickness of the laminate specimen and the effective loading width, and use the ratio of the load to the product as the nominal stress;

[0035] Multiply the nominal stress by the correction factor to obtain the true stress;

[0036] Obtain the corresponding strain data, establish the correspondence between the actual stress and the strain data, and generate the actual constitutive curve.

[0037] This invention eliminates errors caused by geometric and damage effects in nominal stress through inversion calculation, restores the true load-bearing capacity of the laminate, and constructs an accurate stress-strain relationship.

[0038] Preferably, the method for determining the geometric sensitivity constant and the Poisson sensitivity constant includes:

[0039] Construct a solid finite element model with the same geometric dimensions as the cross-shaped specimen to be tested, and an ideal infinite plate model containing only the core pure shear zone features;

[0040] A synchronized sequence of virtual loads is applied to the solid finite element model and the ideal infinite plate model;

[0041] Using the root mean square error of the stress response at the center point of the two models as the objective function, the geometric sensitivity constant and the Poisson sensitivity constant are obtained through iterative calculation using the nonlinear least squares method.

[0042] Preferably, the method for obtaining the strain energy density in the strain energy density distribution weight is as follows:

[0043] Obtain the time of each node within the core pure scissor region. The first principal strain along the first principal axis and the second principal strain along the second principal axis;

[0044] Based on the material constitutive stiffness matrix of the laminate, the first principal stress along the first principal axis and the second principal stress along the second principal axis are calculated using the first principal strain and the second principal strain.

[0045] The product of the first principal stress and the first principal strain, and half the sum of the products of the second principal stress and the second principal strain, are calculated as the strain energy density of the node.

[0046] Secondly, this invention provides a laminate mechanical property evaluation system based on strain data, employing the following technical solution:

[0047] A laminate mechanical property evaluation system based on strain data includes a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the aforementioned laminate mechanical property evaluation method based on strain data is implemented.

[0048] By adopting the above technical solution, a computer program is generated from the above-mentioned method for evaluating the mechanical properties of laminates based on strain data, and stored in a memory so that it can be loaded and executed by a processor. A terminal device can then be made based on the memory and processor for convenient use.

[0049] The present invention has the following technical effects:

[0050] This invention constructs a dynamic correction model that includes the load transfer efficiency attenuation spectrum and the full-field strain energy entropy density. Through mathematical decoupling, it distinguishes and describes the transfer error caused by the specimen geometry and the nonlinear error caused by internal material damage, thereby achieving accurate correction of the nominal stress throughout the entire loading process and significantly improving the accuracy of the biaxial mechanical property evaluation of laminates.

[0051] Furthermore, the present invention provides a full-field strain energy entropy density index, which utilizes the characteristic of the stress field evolving from ordered uniformity to disordered distortion to keenly capture the weakening effect of microscopic damage on macroscopic load-bearing capacity of materials, effectively solving the technical problem that traditional averaging methods cannot detect local stress concentration and stiffness degradation.

[0052] This invention introduces a load normalization weight based on the Sigmoid function to achieve adaptive switching of the correction logic, thereby focusing on correcting geometric transfer errors in the early stage of loading and on correcting damage nonlinear errors in the later stage of loading. This smooth transition mechanism conforms to the natural physical law of material damage accumulation, avoids the deviation caused by a single correction strategy at different loading stages, and ensures that the obtained true constitutive curve has extremely high confidence in both the elastic segment and the plastic or damaged segment. Attached Figure Description

[0053] Figure 1 This is a flowchart of a method for evaluating the mechanical properties of laminates based on strain data, provided in an embodiment of the present invention.

[0054] Figure 2 A comparison diagram of the biaxial tensile mechanical properties of laminates provided in the embodiments of the present invention. Detailed Implementation

[0055] This invention discloses a method for evaluating the mechanical properties of laminates based on strain data, referring to... Figure 1 This includes steps S1-S5:

[0056] S1: A biaxial tensile test was conducted on the laminated plate specimen. Load time history data and full-field speckle images were acquired simultaneously. A full-field strain matrix was constructed using a mapping algorithm, and the full-field strain matrix was discretized into several nodes. Based on the nodes, the region was divided into a core pure shear zone and a stress diffusion zone.

[0057] It should be noted that in order to invert the true mechanical properties of the laminate, high-precision full-field strain data needs to be obtained during the experiment.

[0058] It should be further explained that although traditional digital image correlation (DIC) devices can acquire full-field information, their output raw strain data is usually an unstructured scatter cloud based on camera pixel coordinates. This scatter-like strain data is spatially unevenly distributed, and if directly used for subsequent gradient calculations and energy integration, it will introduce significant computational errors due to data discreteness. Furthermore, this disordered data structure makes it impossible to directly define regular physical region boundaries. If conventional methods are used for region division, boundary effects are easily introduced, affecting the accuracy of stress transfer efficiency calculations.

[0059] Preferably, as an example, a biaxial tensile test is performed on the laminated plate specimen, and load time history data and full-field speckle images are acquired simultaneously. A full-field strain matrix is ​​constructed using a mapping algorithm and divided into a core pure shear zone and a stress diffusion zone, including:

[0060] First, a load is applied to the cross-shaped laminate specimen to be tested using a biaxial testing machine, and the load time history data is recorded simultaneously to obtain the load sequence. Then, a speckle image sequence of the specimen surface is acquired using a DIC device.

[0061] Next, displacement field calculations were performed on the acquired full-field speckle images to obtain unstructured scattered strain data.

[0062] Subsequently, a regular integral point grid covering the observation area of ​​the sample is constructed, with each grid point serving as a node. Using a bilinear interpolation algorithm, the scattered strain data are mapped onto this regular grid to form a time-evolving full-field strain matrix.

[0063] Finally, based on this, according to the geometric characteristics of the specimen, the full-field strain matrix is ​​divided into two physical functional regions: the set of nodes in the pre-defined rhomboid region located at the center of the specimen is defined as the core pure shear region; the set of nodes in the transition region extending from the edge of the rhomboid region to the loading arm is defined as the stress diffusion region.

[0064] Understandably, the full-field strain matrix characterizes the spatiotemporal distribution of the strain field on the specimen surface. When unstructured scattered strain data is mapped to a regular grid, the spatial continuity of the strain data is guaranteed. This verifies that the aforementioned mathematical methods meet the data regularity requirements for subsequent integration calculations, reducing computational errors caused by data discretization. Simultaneously, the clearly defined core pure shear region and stress diffusion region based on the regular grid provide a precise integration domain for the subsequent calculation of the load transfer efficiency attenuation spectrum and the full-field strain energy entropy density, ensuring the spatial accuracy of feature extraction.

[0065] S2: Perform principal component analysis on each node to obtain the first principal strain along the first principal axis; calculate the load transfer efficiency attenuation spectrum characterizing the load transfer fidelity based on the difference in the first principal strain gradient between the stress diffusion region and the core pure shear region.

[0066] It should be noted that in order to achieve geometric correction of the nominal stress, the efficiency characteristics of load transfer need to be extracted from the full-field strain data.

[0067] It should be further explained that the loading arm structure of the cruciform specimen leads to a shear hysteresis effect, which prevents the load applied at the clamping end from being fully transferred to the central region. This shear hysteresis characteristic results in a significant strain gradient between the stress diffusion zone and the core pure shear zone. If conventional static coefficients are used, it will lead to an overestimation or underestimation of the stress level, neglecting the dynamic influence of geometric structural effects during deformation.

[0068] Preferably, as an example, principal component analysis is performed on each node to obtain the first principal strain along the first principal axis; based on the difference in the first principal strain gradient between the stress diffusion region and the core pure shear region, the load transfer efficiency attenuation spectrum characterizing the load transfer fidelity is calculated, including:

[0069] First, taking the node data in the divided full-field strain matrix as the processing object, principal component analysis is performed on the strain tensor of each node. The principal component direction with the largest eigenvalue is extracted as the first principal axis, and the principal component vector with the smallest eigenvalue is obtained as the second principal axis. The projection values ​​of each node along the first principal axis and the second principal axis are obtained as the first principal strain of each node along the first principal axis and the second principal strain of each node along the second principal axis, respectively.

[0070] Then, the load transfer efficiency attenuation spectrum is calculated, specifically satisfying the following relationship:

[0071]

[0072] In the formula, Indicates time The load transfer efficiency attenuation spectrum; Represents an exponential function with the natural constant as its base; This represents the set of nodes in the stress diffusion region; This represents the set of nodes representing the core pure pruning region; Indicating the stress diffusion region in the first... Each node at time... The first principal strain value along the first principal axis; Indicates the first core pure shear region Each node at time... The first principal strain value along the first principal axis; Indicates time The average first principal strain value of all nodes within the core pure shear zone.

[0073] It is understandable that in the relational expression The term characterizes the degree of dissipation of strain energy along the transmission path. Logical deduction from the relational formula shows that when the shear hysteresis effect is significant, it will lead to… The downward trend changes, thus accurately describing the dissipation of strain energy along the transmission path. This directly verifies that the above processing can accurately describe the load transmission loss caused by the geometric structure, making up for the inability of the static correction coefficient to capture the dynamic transmission efficiency changes.

[0074] S3: Based on the strain energy density distribution weights of each node within the core pure shear zone, calculate the full-field strain energy entropy density, which characterizes the orderliness of the stress field.

[0075] It should be noted that, in order to correct the nonlinear damage stage of the material, damage evolution characteristics need to be extracted from the core pure shear region data.

[0076] It should be further explained that after yielding, laminates will develop microcracks, debonding, and other damage. This damage will cause a redistribution of internal stress, disrupting the original uniform stress field. This damage characteristic causes the strain distribution in the core pure shear region to change from uniform to non-uniform. If conventional averaging methods are used, it may mask the stress concentration caused by local damage and fail to accurately reflect the material's load-bearing state.

[0077] Preferably, as an example, the full-field strain energy entropy density, characterizing the stress field order, is calculated based on the strain energy density distribution weights of each node within the core pure shear zone, including:

[0078] First, using the material constitutive stiffness matrix of the laminate. According to Hooke's Law The calculations at time [time] are obtained for each node. First principal stress along the first principal axis and the second principal stress along the second principal axis .

[0079] Then, the strain energy density at this node is calculated, specifically satisfying the following relationship:

[0080]

[0081] In the formula, , The k-th node at time [time] The first principal strain and the first principal stress along the first principal axis; The k-th node at time [time] The second principal strain and the second principal stress along the second principal axis, For the k-th node at time... The strain energy density.

[0082] Next, the strain energy density distribution weights are calculated based on the strain energy density of all nodes, specifically satisfying the following relationship:

[0083]

[0084] Finally, the overall strain energy entropy density is calculated based on the strain energy density distribution weights of each node, specifically satisfying the following relationship:

[0085]

[0086] In the formula, Indicates time The full-field strain energy entropy density; This represents the total number of nodes in the core pure scissor region. For a moment No. The strain energy density distribution weight of each node represents the proportion of that single point's contribution to the overall energy distribution.

[0087] Understandably, the entropy value in the relation characterizes the purity or order of the stress field. Logical deduction from the relation reveals that when a material enters the nonlinear damage stage, the stress distribution becomes non-uniform, thus affecting the calculated relation. The significant decrease effectively reflects the physical phenomenon of stress field distortion caused by the accumulation of internal damage in the material.

[0088] This directly verifies that thermodynamic entropy analysis accurately responds to macroscopic stress field distortion caused by microscopic material damage, effectively eliminating the potential danger of traditional methods failing to detect the effects of nonlinear damage.

[0089] S4: Retrieve preset shape feature parameters, use the shape feature parameters to weighted couple the load transfer efficiency attenuation spectrum and the full-field strain energy entropy density to construct a dynamic correction model, and use the dynamic correction model to calculate the correction coefficient.

[0090] It should be noted that, in order to achieve the final stress correction, the efficiency characteristics of load transfer and the damage evolution characteristics need to be decoupled and fused.

[0091] It should be further noted that the correction coefficient is affected not only by real-time changing experimental data but also by the inherent geometric properties of the specimen. This coupling characteristic necessitates the introduction of specific shape characteristic parameters to describe the interaction between the efficiency characteristics of load transfer and the damage evolution characteristics.

[0092] Preferably, as an example, preset shape feature parameters are retrieved, and the load transfer efficiency attenuation spectrum and the full-field strain energy entropy density are weighted and coupled using the shape feature parameters to construct a dynamic correction model. The correction coefficients are calculated using the dynamic correction model, including:

[0093] First, determine the shape feature parameters, including:

[0094] A solid finite element model with the same geometric dimensions as the cruciform specimen to be tested, and an ideal infinite plate model containing only the core pure shear zone features, are constructed. A synchronized sequence of virtual loads is applied to both models, and the geometric sensitivity constant and Poisson sensitivity constant are calculated iteratively using the nonlinear least squares method with the objective function of minimizing the root mean square error of the stress response at the center point of both models. These geometric sensitivity constants and Poisson sensitivity constants are collectively referred to as shape characteristic parameters.

[0095] Next, the load normalization weights are calculated, including:

[0096]

[0097] In the formula, For a moment The load normalization weights, For a moment The load value, To represent the median of the load sequence, This is the shape adjustment coefficient, which controls the steepness of the Sigmoid curve; The higher the value, the more severe the transition from the elastic stage to the damage stage of the material. The smaller the value, the smoother the damage evolution; e is the natural constant.

[0098] It is understandable that when the load When smaller, Approaching When the load When it increases, Approaching This sigmoid-based weighting mechanism perfectly replicates the physical laws of material damage evolution, avoiding errors caused by prematurely introducing damage correction during the elastic stage.

[0099] Finally, the correction factor is calculated, specifically satisfying the following relationship:

[0100]

[0101] In the formula, Indicates time Correction factor; and These are the preset geometric sensitivity constant and Poisson sensitivity constant, respectively; For a moment The load normalization weights, It is a dynamically corrected model.

[0102] Understandably, when the load is low, Approaching 0, thus calculating It approaches 1. At this point... At this point, the correction factor is mainly determined by the transmission efficiency. The determination is consistent with the physical phenomenon that the linear elastic stage is mainly affected by shear hysteresis.

[0103] High load The stress density approaches 1, and due to damage accumulation, the disorder of the stress field increases, as does the entropy density. A significant decrease, at which point the denominator... The smaller term leads to a decrease in the overall denominator, thus making it more accurately reflect the decline in load-bearing capacity after damage.

[0104] The above process achieves a smooth transition from simple geometric correction to geometric-plus-damage coupling correction, ensuring the accuracy of the correction coefficients throughout the entire loading process.

[0105] S5: Use the correction coefficient to correct the nominal stress, obtain the true constitutive curve, and complete the mechanical property evaluation of the laminate.

[0106] It should be noted that, in order to obtain the true mechanical properties of the laminate, the calculated correction factor needs to be corrected for the nominal stress measured.

[0107] It should be further explained that the nominal stress calculated from the load data directly measured by the force sensor actually includes a mixture of the unique geometric structural effects and material damage effects of the cruciform specimen. If the nominal stress is used directly to construct the constitutive curve without inversion processing, it will lead to a significant deviation in the material strength design parameters.

[0108] Preferably, as an example, the nominal stress is corrected using the correction factor to obtain the true constitutive curve, and the mechanical properties of the laminate are evaluated, including:

[0109] First, at any time Based on the collected load values, the thickness of the laminate specimen, and the effective loading width, the nominal stress is calculated, specifically satisfying the following relationship:

[0110]

[0111] In the formula, For a moment nominal stress, For a moment The load value, The thickness of the laminate sample is given. This is the effective width of the test area at the center of the sample.

[0112] Next, the nominal stress is corrected in real time using a correction factor to obtain the true stress, which specifically satisfies the following relationship:

[0113]

[0114] In the formula, For a moment The actual stress borne by the core shear zone of the laminate. For a moment The dynamic correction coefficient.

[0115] Then, retrieve the time. Core pure cut area The average first principal strain value Build data pairs Iterate through all sampling moments of the entire loading process. Several data pairs are obtained, where T represents the number of samples during the loading process. All data are used to fit the final true stress-strain constitutive curve.

[0116] Understandably, in the elastic stage, the main correction is made for stress transfer losses caused by shear hysteresis, ensuring that the calculated true stress reflects the actual force on the material; in the damage stage, as damage intensifies, the correction factor... It will dynamically adjust according to changes in entropy density, reducing the portion of the nominal stress curve that is artificially inflated due to stiffness degradation.

[0117] The modified constitutive curve has a slope in the elastic segment that is more consistent with the theoretical value, and the hardening or softening behavior after yielding truly reflects the damage evolution of the material, eliminating the interference of the sample shape on the evaluation of material performance.

[0118] Finally, based on the generated true constitutive curves, the key mechanical performance indicators of the laminate are automatically calculated and output, specifically:

[0119] The true elastic modulus is obtained by processing the elastic linear segment data of the true constitutive curve using a linear regression algorithm; the true ultimate strength is obtained by processing the stress data of the entire constitutive curve using an extreme value search algorithm; and the failure strain is obtained by processing the end data of the true constitutive curve using a fracture feature recognition algorithm.

[0120] Figure 2 This is a comparison chart of the biaxial tensile mechanical properties of laminates. The horizontal axis represents strain, and the vertical axis represents stress. The solid black line represents the ideal material constitutive curve preset in the finite element digital twin model, without geometric defects in the cross-shaped specimen, serving as the standard reference value for this verification. The dashed red line represents the nominal stress curve calculated directly from the testing machine load using traditional methods. The red filled area is the region between the dashed red line and the solid black line, visually demonstrating the systematic errors present in existing technologies when testing cross-shaped specimens. The dotted blue line represents the true stress curve obtained after correction using this invention.

[0121] The image shows that the red dashed line is significantly higher than the black baseline, indicating that the test result is too high. The blue dotted line successfully presses down and conforms to the black baseline, proving that the invention effectively eliminates systematic errors. Furthermore, in the nonlinear stage where the strain is greater than 1.0%, the error of the red curve amplifies divergently, while the blue curve of the invention closely follows the black baseline in this stage. This demonstrates that by introducing the full-field strain energy entropy density, the algorithm can accurately correct complex nonlinear errors caused by material damage.

[0122] This invention also discloses a laminate mechanical property evaluation system based on strain data, including a processor and a memory. The memory stores computer program instructions, which, when executed by the processor, implement a laminate mechanical property evaluation method based on strain data according to this invention.

[0123] The system also includes other components well known to those skilled in the art, such as communication buses and communication interfaces, the settings and functions of which are known in the art and will not be described in detail here.

[0124] In this invention, the aforementioned memory can be any tangible medium containing or storing a program that can be used or combined with an instruction execution system, apparatus, or device. For example, a computer-readable storage medium can be any suitable magnetic or magneto-optical storage medium, such as resistive random access memory (DRAM), dynamic random access memory (DRAM), static random access memory (SRAM), enhanced dynamic random access memory (DRAM), high-bandwidth memory, hybrid memory cube, etc., or any other medium that can be used to store desired information and can be accessed by an application, module, or both. Any such computer storage medium can be part of a device or accessible to or connected to a device.

Claims

1. A method for evaluating the mechanical properties of laminates based on strain data, characterized in that, Including the following steps: Biaxial tensile tests were conducted on laminate specimens, and load time history data and full-field speckle images were acquired simultaneously. The full-field strain matrix was constructed by a mapping algorithm and discretized into several nodes. Based on the nodes, the region was divided into a core pure shear zone and a stress diffusion zone. Principal component analysis is performed on each node to obtain the first principal strain along the first principal axis; Based on the difference in the first principal strain gradient between the stress diffusion zone and the core pure shear zone, the load transfer efficiency attenuation spectrum, which characterizes the load transfer fidelity, is calculated. Based on the strain energy density distribution weights of each node within the core pure shear region, the full-field strain energy entropy density, characterizing the orderliness of the stress field, is calculated, satisfying: , For a moment The full-field strain energy entropy density, This represents the total number of nodes within the core pruning region. It is the natural logarithm function; Indicates time The first core pure cut zone The strain energy density distribution weight of the nth node, whose value is the nth node. Each node at time... strain energy density and time The ratio of total strain energy density in the core pure shear region; Retrieve preset shape feature parameters, and use these parameters to weightedly couple the load transfer efficiency attenuation spectrum and the full-field strain energy entropy density to construct a dynamic correction model. Calculate the correction coefficients using this model, satisfying the following conditions: , For a moment Correction factor, The geometric sensitivity constant is... The load transfer efficiency attenuation spectrum. Let be the Poisson sensitivity constant. The load normalization weights are used; the geometric sensitivity constant and Poisson sensitivity constant are the shape characteristic parameters. The nominal stress is corrected using a correction factor to obtain the true constitutive curve, thus completing the mechanical property evaluation of the laminate.

2. The method for evaluating the mechanical properties of laminates based on strain data according to claim 1, characterized in that, The construction of the full-field strain matrix through a mapping algorithm includes: Displacement field calculations were performed on the acquired full-field speckle images to obtain unstructured scattered strain data. A bilinear interpolation algorithm based on shape functions is used to map the scattered strain data onto a regular grid of integral points, forming a full-field strain matrix that includes the spatial location of nodes and the time dimension.

3. The method for evaluating the mechanical properties of laminates based on strain data according to claim 1, characterized in that, The load transfer efficiency attenuation spectrum satisfies the following relationship: In the formula, Indicates time The load transfer efficiency attenuation spectrum; Represents an exponential function with the natural constant as its base; This represents the set of nodes in the stress diffusion region; This represents the set of nodes representing the core pure pruning region; Indicating the stress diffusion region in the first... Each node at time... The first principal strain value along the first principal axis; Indicates the first core pure shear region Each node at time... The first principal strain value along the first principal axis; Indicates time The average first principal strain value of all nodes within the core pure shear zone.

4. The method for evaluating the mechanical properties of laminates based on strain data according to claim 1, characterized in that, The method for obtaining the load normalization weights is as follows: Calculate the deviation of the load value relative to the values ​​in the load sequence; The load normalization weight is calculated using the Sigmoid function based on the deviation, so that the load normalization weight increases non-linearly with the increase of load.

5. The method for evaluating the mechanical properties of laminates based on strain data according to claim 1, characterized in that, The step of correcting the nominal stress using the correction coefficient to obtain the true constitutive curve includes: Calculate the product of the thickness of the laminate specimen and the effective loading width, and use the ratio of the load to the product as the nominal stress; Multiply the nominal stress by the correction factor to obtain the true stress; Obtain the corresponding strain data, establish the correspondence between the actual stress and the strain data, and generate the actual constitutive curve.

6. The method for evaluating the mechanical properties of laminates based on strain data according to claim 1, characterized in that, The method for determining the geometric sensitivity constant and the Poisson sensitivity constant includes: Construct a solid finite element model with the same geometric dimensions as the cross-shaped specimen to be tested, and an ideal infinite plate model containing only the core pure shear zone features; A synchronized sequence of virtual loads is applied to the solid finite element model and the ideal infinite plate model; Using the root mean square error of the stress response at the center point of the two models as the objective function, the geometric sensitivity constant and the Poisson sensitivity constant are obtained through iterative calculation using the nonlinear least squares method.

7. The method for evaluating the mechanical properties of laminates based on strain data according to claim 1, characterized in that, The method for obtaining the strain energy density in the strain energy density distribution weight is as follows: Obtain the time of each node within the core pure scissor region. The first principal strain along the first principal axis and the second principal strain along the second principal axis; Based on the material constitutive stiffness matrix of the laminate, the first principal stress along the first principal axis and the second principal stress along the second principal axis are calculated using the first principal strain and the second principal strain. The product of the first principal stress and the first principal strain, and half the sum of the products of the second principal stress and the second principal strain, are calculated as the strain energy density of the node.

8. A system for evaluating the mechanical properties of laminates based on strain data, characterized in that, include: A processor and a memory, the memory storing computer program instructions that, when executed by the processor, implement a method for evaluating the mechanical properties of laminates based on strain data according to any one of claims 1-7.

Images