Method and device for detecting performance of polyurethane material

Abstract

In the technical field of detection of the present invention, a performance detection method for polyurethane materials is disclosed, including clamping a material to be tested on a loading platform, simulating the force action of a target application scenario through collaborative control, using a vision unit to capture sequential images of the microscopic morphology of the material surface, synchronously collecting internal strain distribution data through a distributed strain sensing array, obtaining a dynamic stress response spectrum in combination with a piezoelectric mechanical sensor, and generating a multi-modal deformation feature data set; performing crack density and texture orientation analysis on the sequential images, extracting the crack propagation rate and the direction consistency coefficient, and marking the strain concentration area; calculating the apparent elastic modulus decay rate based on the dynamic stress response spectrum, establishing a material performance degradation model, generating a dynamic flexibility performance index through a weighted fusion algorithm, and grading and evaluating the material; upgrading the material detection from a single mechanical test to an intelligent evaluation system of multi-physical field coupling, and providing accurate and reliable detection support for various application scenarios.

Description

Technical Field

[0001] The present invention relates to the field of detection technologies, and in particular to a method and device for detecting the performance of polyurethane materials. Background Art

[0002] Due to its excellent flexibility, wear resistance and environmental adaptability, polyurethane materials are widely used in precision fields such as hinges of foldable electronic devices and dynamic seals of automobiles. With the development of flexible electronic devices towards ultra-thinness and high-frequency bending, and the increasing requirements for the dynamic durability of the sealing system in automotive lightweighting, traditional static mechanical property indicators (such as tensile strength and hardness) can no longer meet the performance evaluation needs of materials under complex alternating stresses. Especially under extreme working conditions such as thousands of times of bending per day for hinges and long-term vibration compression for seals, the correlation mechanism between micro-damage accumulation and macroscopic property degradation of materials has become the key bottleneck restricting product reliability.

[0003] In the prior art, the performance detection of polyurethane materials mostly adopts a static observation scheme combining a uniaxial tensile testing machine and a digital microscope. For example, the flexible performance is evaluated by measuring parameters such as elongation at break and elastic recovery rate. This type of method has a core defect of insufficient adaptability to dynamic scenarios: its uniaxial loading mode cannot simulate the multi-directional composite stress in real applications, resulting in a deviation of up to 30%-50% between the detection results and the actual working conditions; at the same time, static parameters cannot characterize the performance attenuation law of materials under cyclic loading, and it is difficult to predict the crack initiation position and the end point of life. This defect makes the traditional detection method have a serious design guidance blind area in scenarios such as material selection for the rotating shaft of foldable mobile phones and verification of flexible seals for new energy vehicle battery packs, and there is an urgent need to develop a multi-dimensional detection technology that can reproduce dynamic working conditions.

[0004] In view of this, it is necessary to improve the performance detection technology of polyurethane materials in the prior art to solve the technical problems of its single detection dimension and low accuracy. Summary of the Invention

[0005] The purpose of the present invention is to provide a method and device for detecting the performance of polyurethane materials to solve the above technical problems.

[0006] To achieve this purpose, the present invention adopts the following technical solutions:

[0007] A method for detecting the performance of polyurethane materials, comprising:

[0008] In a preset detection environment, a strip-shaped polyurethane material to be tested is clamped on a multi-axis deformation loading platform, and the force acting in the target application scenario is simulated through coordinated control, and the real-time temperature change during the deformation process is recorded;

[0009] Capture the sequential images of the microscopic topography of the material surface using a vision unit, synchronously collect the internal strain distribution data through a distributed strain sensing array, and combine with a piezoelectric mechanical sensor to obtain the dynamic stress response spectrum, generating a multi-modal deformation feature dataset;

[0010] Perform crack density and texture orientation analysis on the sequential images, extract the crack propagation rate and direction consistency coefficient, and at the same time perform three-dimensional deformation field reconstruction on the internal strain data, marking the coordinates of the strain concentration area and the change amount of the strain gradient;

[0011] Calculate the apparent elastic modulus decay rate based on the dynamic stress response spectrum, combine with the crack propagation rate and the change amount of the strain gradient, establish a material performance degradation model, and correlate with the cyclic life prediction equation of the target application scenario;

[0012] Generate a dynamic flexibility performance index according to the crack density, direction consistency coefficient, and apparent elastic modulus decay rate, and grade and evaluate the material based on a preset scenario fitness threshold.

[0013] Optionally, in a preset detection environment, clamp the strip-shaped polyurethane material to be tested on a multi-axis deformation loading platform, simulate the force action of the target application scenario through coordinated control, and record the real-time temperature change during the deformation process, specifically including:

[0014] Build a multi-axis deformation loading platform in a constant temperature environmental chamber, clamp both ends of the strip-shaped polyurethane material to be tested with pneumatic clamps, the surface of the pneumatic clamps is coated with a polytetrafluoroethylene anti-slip layer, and calibrate the spatial angle deviation between the material clamping axis and the platform reference plane through a laser alignment instrument;

[0015] Set the loading mode based on the working condition parameters of the target application scenario;

[0016] Drive the axial tensile unit through a servo motor, apply linear tensile stress at a preset tensile rate, and at the same time control the multi-directional bending module by a pneumatic driving device to generate dynamic bending deformation in the middle of the material;

[0017] Adopt a closed-loop feedback control system to synchronously adjust the phase relationship of the stretching, bending, and pressing actions, and trigger a protection mechanism according to the phase relationship. The protection mechanism includes: when the yield point of the material is detected, automatically reduce the tensile rate to 50% of the set value, and trigger the indenter retraction protection mechanism when the bending angle reaches the threshold;

[0018] Mount a temperature sensing array on the surface of the material, record the temperature fluctuation data during the deformation detection at a preset sampling frequency, and generate a temperature-deformation sequential correlation curve.

[0019] Optionally, the temperature sensing array consists of 16 thin film sensors, and the 16 thin film sensors cover the stretching area, bending area and pressing point in a serpentine layout; wherein, the thin film sensor is a PT100 thin film sensor.

[0020] Optionally, the method uses a vision unit to capture sequential images of the microscopic morphology of the material surface, synchronously collects internal strain distribution data through a distributed strain sensing array, combines a piezoelectric mechanical sensor to obtain a dynamic stress response spectrum, and generates a multi-modal deformation feature dataset, specifically including:

[0021] Deploy an imaging system on the material surface, use a polarized light source to irradiate the area to be measured at a preset incident angle, adjust the polarization direction to form a preset angle with the material stretching axis, and capture sequential images of the surface microscopic morphology;

[0022] Embed a distributed grating array along the interior of the material, and use a demodulator to monitor the strain values of each grating node in real time to generate a three-dimensional strain distribution dataset in the axial, radial and tangential directions;

[0023] Install piezoelectric mechanical sensors at the material clamping end and the bending area, collect the dynamic stress spectrum through an anti-aliasing filter, and synchronously record the time stamp.

[0024] Optionally, after collecting the dynamic stress spectrum through the anti-aliasing filter and synchronously recording the time stamp, the following steps are further included:

[0025] Establish a multi-source data synchronization and fusion mechanism: when the deformation action is triggered on the loading platform, send a hardware synchronization pulse signal to the vision unit, fiber optic demodulator and piezoelectric sensor, so that the time deviation between the sequential image, strain data and stress spectrum is within the allowable range;

[0026] Preprocess the collected data: perform non-uniform illumination correction and motion blur compensation on the sequential image, use median filtering to eliminate the instantaneous noise of the three-dimensional strain distribution dataset, and perform baseline drift correction and dimension normalization on the dynamic stress spectrum;

[0027] Align the preprocessed surface image, three-dimensional strain distribution dataset and dynamic stress spectrum along the time axis, and construct a multi-modal deformation feature database including spatial coordinates, strain gradient, stress amplitude and texture feature vectors.

[0028] Optionally, perform crack density and texture orientation analysis on the sequential image, extract the crack propagation rate and direction consistency coefficient, and at the same time perform three-dimensional deformation field reconstruction on the internal strain data, and mark the coordinates and strain gradient change amounts of the strain concentration area, specifically including:

[0029] Perform dynamic noise reduction processing on time-series images, separate background textures and crack pixels, and compensate for image displacement artifacts caused by material deformation based on the optical flow method to generate a denoised and enhanced image set;

[0030] Use the U-Net network to perform crack semantic segmentation on the denoised and enhanced image set. Set the crack length threshold to 50μm, extract the crack skeleton lines of each frame and calculate the fractal dimension, and statistically calculate the total crack length per unit area to generate a crack density heat map;

[0031] Analyze the crack skeleton line direction through the histogram of oriented gradients, use weighted transformation to detect the main crack propagation direction, and calculate the anisotropy coefficient η = 1 - energy of the secondary dominant direction / energy of the main direction;

[0032] Perform spatial interpolation on the three-dimensional strain distribution data set, reconstruct the three-dimensional deformation field including axial, radial, and tangential strain components, calculate the strain gradient change amount, and identify the local areas where the strain gradient change rate exceeds 0.5% / mm² as strain concentration areas;

[0033] Map the coordinates of the strain concentration areas to the three-dimensional solid model of the material, calculate the strain energy density W = 0.5σ·ε of each hot spot area, generate a strain energy density cloud map and perform spatial registration with the crack density heat map;

[0034] Establish a crack-strain coupling analysis matrix, mark the coordinates of the dangerous areas that simultaneously satisfy the crack density ≥ 0.8mm / mm² and the strain energy density ≥ 15kJ / m³, and record their spatial distribution patterns and strain gradient change amounts.

[0035] Optionally, calculate the apparent elastic modulus decay rate based on the dynamic stress response spectrum, combine the crack propagation rate and the strain gradient change amount, establish a material property degradation model, and associate the cyclic life prediction equation of the target application scenario, specifically including:

[0036] Perform time-domain segmentation processing on the dynamic stress response spectrum, divide the stress relaxation stage, steady-state holding stage, and elastic rebound stage based on the deformation loading characteristics, calculate the stress decay slope and creep recovery rate of each stage, and generate a modulus decay feature vector;

[0037] Fit the apparent elastic modulus decay curve with a piecewise exponential function, use the peak stress ratio of adjacent deformation cycles as the decay factor, and solve the time-varying weight coefficient of the elastic modulus decay rate through the nonlinear least squares method;

[0038] Extract the spatial distribution heat map of the crack propagation rate, calculate the cosine value of the angle between the crack main direction and the strain gradient vector, and mark the three-dimensional coordinate point set where the strain gradient change amount exceeds the preset change threshold;

[0039] Construct a performance degradation differential equation that includes the modulus decay rate, crack-strain coupling factor, and local gradient mutation, and use the adaptive iteration method to solve the material damage accumulation rate;

[0040] Reconstruct the accelerated aging test parameters based on the typical load spectrum of the target scenario, substitute the damage accumulation rate into the Paris-Erdogan equation, and establish a cyclic life prediction model considering the multiaxial stress state;

[0041] By sampling and simulating the material performance degradation path, calculate the confidence interval of the life prediction value, and trigger the model parameter dynamic calibration mechanism when the interval width exceeds the preset threshold.

[0042] Optionally, according to the crack density, direction consistency coefficient, and apparent elastic modulus decay rate, generate a dynamic flexible performance index through a weighted fusion algorithm, and conduct a hierarchical evaluation of the material based on the preset scenario fitness threshold, specifically including:

[0043] Normalize the crack density, direction consistency coefficient, and apparent elastic modulus decay rate, convert the crack density into a relative damage degree in the range of 0-1, calibrate the matching index of the direction consistency coefficient according to the cosine similarity, and map the modulus decay rate to the preset degradation level;

[0044] Based on the failure mode library of the target application scenario, use the analytic hierarchy process to allocate weight coefficients;

[0045] Construct a dynamic flexible performance index calculation model: linearly weighted sum the normalized parameters and weight coefficients, and superimpose the nonlinear correction term of the strain energy density gradient to generate a comprehensive performance score;

[0046] Set the fitness threshold according to the historical test data and the limit value of the scenario working conditions;

[0047] Compare the comprehensive score with the preset threshold, and execute the hierarchical evaluation decision according to the comparison result;

[0048] Generate a hierarchical evaluation map according to the hierarchical evaluation decision, divide the material surface into several grid units, assign red-yellow-green three-color warning signs according to the performance scores within the grid units, and mark the spatial coordinates and weight rankings of the priority improvement areas.

[0049] Optionally, after the hierarchical evaluation of the material is conducted based on the preset scenario fitness threshold, it further includes:

[0050] Integrate the hierarchical evaluation results with the deformation field distribution and life prediction data to generate a visual inspection document that includes a three-dimensional damage evolution map, process optimization suggestions, and a scenario fitness report, and transmit it to the terminal interaction interface through an encrypted channel.

[0051] The present invention also provides a performance detection device for polyurethane materials, which is used for the performance detection method of polyurethane materials as described above. The performance detection device includes:

[0052] A constant temperature environmental chamber, in which a laser alignment instrument is arranged;

[0053] A multi-axis deformation loading platform, which is provided with a pneumatic fixture, an axial tension unit, a multi-directional bending module and a servo motor;

[0054] A multi-modal sensing module, including a vision unit, a temperature sensing array, a piezoelectric mechanical sensor and a distributed grating array;

[0055] A data processing unit, which is used to process the acquired temperature data, sequential images, strain distribution data and dynamic stress response spectra to generate a multi-modal deformation feature data set;

[0056] A decision-making system, which is used to carry a data processing model and conduct a hierarchical evaluation of the material.

[0057] Compared with the prior art, the present invention has the following beneficial effects: First, the composite stress action of the target scenario is simulated through the multi-axis deformation loading platform to trigger the evolution of microscopic damage of the material; the vision unit, strain sensing array and piezoelectric sensor are used to synchronously capture the surface crack propagation, internal strain distribution and dynamic mechanical response, and a multi-modal feature data set is constructed; the density and orientation of the surface cracks are analyzed, and combined with the three-dimensional strain field reconstruction, the damage hot spots of the material are identified; based on the dynamic stress response spectrum, the elastic modulus decay rate is calculated, a performance degradation model is established and the cycle life is predicted; by weighted fusion of the crack density, direction consistency and modulus decay parameters, a dynamic flexibility performance index is generated to realize the quantitative evaluation of the material scenario adaptability; this method upgrades the material detection from a single mechanical test to an intelligent evaluation system of multi-physical field coupling, solves the technical problem that the traditional method cannot characterize the mismatch between the dynamic flexibility performance and the real scenario, and provides accurate and reliable detection support for various application scenarios. Description of the Drawings

[0058] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained according to these drawings without creative efforts.

[0059] The structures, proportions, sizes, etc. shown in the drawings of this specification are only used to cooperate with the content disclosed in the specification for those familiar with this technology to understand and read, and are not used to limit the conditions for the implementation of the present invention. Therefore, they do not have substantial technical significance. Any modification of the structure, change in the proportional relationship, or adjustment of the size, without affecting the effects that the present invention can produce and the purposes that can be achieved, should still fall within the scope covered by the technical content disclosed in the present invention.

[0060] Figure 1 One of the schematic flowcharts of the performance detection method of the polyurethane material in the first embodiment;

[0061] Figure 2 Another schematic flowchart of the performance detection method of the polyurethane material in the first embodiment;

[0062] Figure 3 Schematic diagram of the performance detection device of the polyurethane material in the second embodiment. Detailed implementation manners

[0063] In order to make the invention objectives, features, and advantages of the present invention more obvious and understandable, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the embodiments described below are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the scope of protection of the present invention.

[0064] In the description of the present invention, it should be understood that the orientation or positional relationships indicated by the terms "upper", "lower", "top", "bottom", "inner", "outer", etc. are based on the orientation or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be construed as a limitation of the present invention. It should be noted that when a component is considered to be "connected" to another component, it can be directly connected to the other component or there may be intermediate components present.

[0065] The technical solutions of the present invention will be further described below in conjunction with the drawings and through specific implementation manners.

[0066] Embodiment 1:

[0067] Please refer to Figure 1 and Figure 2 As shown, the embodiment of the present invention provides a performance detection method for a polyurethane material, including:

[0068] S1. In a preset detection environment, clamp the strip-shaped polyurethane material to be tested on a multi-axis deformation loading platform, simulate the force action of the target application scenario through coordinated control, and record the real-time temperature change during the deformation process;

[0069] In a thermostatic and humidistatic chamber, apply composite stresses (tensile, bending, pressing) simulating real working conditions to the strip-shaped polyurethane material through a multi-axis deformation loading platform. The servo motor controls the axial tensile rate, the pneumatic bending module dynamically adjusts the bending radius, and the piezoelectric actuator array applies local periodic pressure.

[0070] S2. Use a vision unit to capture sequential images of the microscopic morphology of the material surface. Synchronously collect internal strain distribution data through a distributed strain sensing array, and combine with a piezoelectric mechanical sensor to obtain a dynamic stress response spectrum, generating a multi-modal deformation feature dataset;

[0071] The vision unit is preferably a polarization microscopy system to capture the surface crack initiation process at 120 fps, and the light intensity difference enhances the contrast of microcracks. The embedded fiber Bragg grating array (2 mm spacing) calculates three-dimensional strain (axial / radial / tangential) through wavelength shift, and the piezoelectric sensor collects dynamic stress signals, which are converted into voltage spectra by a charge amplifier. The hardware synchronous trigger module sends TTL pulses to align the time stamps of optical, mechanical, and thermal data.

[0072] S3. Perform crack density and texture orientation analysis on the sequential images, extract the crack propagation rate and direction consistency coefficient. At the same time, perform three-dimensional deformation field reconstruction on the internal strain data, and mark the coordinates and strain gradient change amounts of the strain concentration regions.

[0073] S4. Calculate the apparent elastic modulus decay rate based on the dynamic stress response spectrum, combine the crack propagation rate and the strain gradient change amount, establish a material performance degradation model, and correlate with the cycle life prediction equation of the target application scenario;

[0074] Wavelet packet energy feature: The propagation of microcracks inside the material will excite stress waves in a specific frequency band (2 - 4 kHz), and the energy accumulation is positively correlated with the damage degree.

[0075] Gradient-corrected Paris model: The traditional Paris law only considers the stress intensity factor ΔK. Introduce the strain gradient term α∇ε to reflect the influence of multi-axis loads and improve the prediction accuracy.

[0076] S5. Generate a dynamic flexibility performance index through a weighted fusion algorithm according to the crack density, direction consistency coefficient, and apparent elastic modulus decay rate, and perform a grading evaluation on the material according to the preset scenario adaptability threshold.

[0077] The working principle of the present invention is as follows: First, a multi-axis deformation loading platform is used to simulate the combined stress action of the target scenario, triggering the evolution of microscopic damage in the material. A vision unit, a strain sensing array, and a piezoelectric sensor are used to synchronously capture the surface crack propagation, internal strain distribution, and dynamic mechanical response, constructing a multi-modal feature dataset. The density and orientation of the surface cracks are analyzed, and combined with the three-dimensional strain field reconstruction, the material damage hotspots are identified. Based on the dynamic stress response spectrum, the elastic modulus decay rate is calculated, a performance degradation model is established and the cyclic life is predicted. By weighted fusion of the crack density, direction consistency, and modulus decay parameters, a dynamic flexibility performance index is generated to realize the quantitative evaluation of the material scene adaptability. This method upgrades the material detection from a single mechanical test to an intelligent evaluation system of multi-physical field coupling, solves the technical problem that the traditional method cannot characterize the mismatch between the dynamic flexibility performance and the real scene, and provides accurate and reliable detection support for various application scenarios. In this embodiment, specifically, step S1 specifically includes:

[0078] S11, Build a multi-axis deformation loading platform in a constant temperature environmental chamber, clamp both ends of the strip-shaped polyurethane material to be tested on a pneumatic fixture, the surface of the pneumatic fixture is coated with a polytetrafluoroethylene anti-slip layer, and the spatial angle deviation between the material clamping axis and the platform reference plane is calibrated by a laser alignment instrument.

[0079] S12, Set the loading mode based on the working condition parameters of the target application scenario; for example: for the hinge scenario of a foldable device, input the parameter set of 500 ± 50 times of daily bending, a bending radius of 0.5 - 2 mm, and a stretching amplitude of 120% - 150%; for the automotive dynamic sealing scenario, input the parameter set of a vibration frequency of 20 - 100 Hz and a compression deformation rate of 15% - 30%.

[0080] S13, Drive the axial stretching unit through a servo motor, apply a linear tensile stress at a preset stretching rate, and at the same time, control the multi-directional bending module by a pneumatic driving device to generate a dynamic bending deformation in the middle section of the material;

[0081] The servo motor drives the ball screw to achieve axial stretching, and the speed is controlled in segments: 0.5 mm / s in the initial stage (elastic region), and the speed is reduced to 0.25 mm / s after the yield point (plastic region). The pneumatic bending module applies dynamic bending to the middle section of the material through a profiling die, and the bending radius is switched in real time through the curvature radius of the die. Synchronously trigger the piezoelectric actuator to apply a point pressure load of 0.1 - 5 N in the bending area to simulate the repeated pressing condition of a touch screen.

[0082] S14, Adopt a closed-loop feedback control system to synchronously adjust the phase relationship of the stretching, bending, and pressing actions, and trigger a protection mechanism according to the phase relationship. The protection mechanism includes: when the yield point of the material is detected, automatically reduce the stretching rate to 50% of the set value, and trigger the indenter retraction protection mechanism when the bending angle reaches the threshold;

[0083] The closed-loop feedback system controls through the linkage of three signals:

[0084] Yield point detection: Calculate the second derivative of the stress-strain curve in real time. When d²σ / dε² < -10 MPa, it is determined that the material enters the plastic stage, and a speed reduction command is triggered.

[0085] Bending protection: A high-precision rotary encoder monitors the bending angle. When it exceeds the preset threshold (such as 90°), the solenoid valve exhausts air emergently to reset the cylinder.

[0086] Phase synchronization: Synchronously control the phase difference of the stretching, bending, and pressing actions through the EtherCAT bus to be < 0.2 ms to prevent multi-directional stress interference.

[0087] S15. Mount a temperature sensing array on the material surface, record the temperature fluctuation data during deformation detection at a preset sampling frequency, and generate a temperature-deformation time-series correlation curve.

[0088] Mount a flexible temperature sensing array on the material surface, and adopt a serpentine wire layout (covering the stretching area (5 sensors at each end), the bending area (4 sensors), and the pressing point (2 sensors)). The PT100 thin film sensors are attached by polyimide tape, the sampling rate is 10 Hz, and the data is converted into temperature values through the RTD module. The temperature-deformation time-series curve shows that the temperature rise rate in the stretching stage is about 0.2 °C / s, and the local hot spot temperature difference in the bending area reaches 3 °C.

[0089] In this embodiment, the temperature sensing array is composed of 16 thin film sensors, and the 16 thin film sensors cover the stretching area, the bending area, and the pressing point with a serpentine wire layout; among them, the thin film sensors are PT100 thin film sensors.

[0090] It should be noted that the 16-point layout (10 points in the stretching area, 4 points in the bending area, and 2 points in the pressing point) realizes the full coverage of the deformation heat effect, and the spatial resolution reaches 5 mm; the serpentine wire layout avoids stress concentration caused by wire concentration and ensures uniform sensor spacing at the same time.

[0091] The time-series curve can identify the viscous thermal hysteresis effect (such as the temperature peak in the bending area lags the stress peak by 1 - 2 s); the thermal gradient distribution reflects the energy dissipation difference caused by internal friction heating of the material.

[0092] In this embodiment, specifically, step S2 specifically includes:

[0093] S21. Deploy an imaging system on the material surface, irradiate the area to be measured with a polarized light source at a preset incident angle, adjust the polarization direction to form a preset angle with the material stretching axis, and capture the time-series images of the surface microtopography.

[0094] Deploy a high-resolution polarization microscopy imaging system on the material surface. Use a 532nm polarization light source to irradiate the area to be measured at an incident angle of 30°. Adjust the polarization direction to form a 45° angle with the material's tensile axis to enhance the crack edge contrast.

[0095] S22, Embed a distributed grating array inside the material. Real-time monitor the strain values of each grating node through a demodulator to generate a three-dimensional strain distribution dataset in the axial, radial, and tangential directions;

[0096] Embed a distributed fiber Bragg grating array inside the material with a grating spacing of 2mm. Real-time monitor the strain values of each grating node through a wavelength demodulator (resolution 1pm, sampling rate 1kHz). Based on the linear relationship between the wavelength shift Δλ_B and the strain ε (Δλ_B = k·ε, k ≈ 1.2pm / με), calculate the axial, radial, and tangential strain components to generate a three-dimensional strain distribution dataset.

[0097] S23, Install piezoelectric mechanical sensors at the material clamping end and the bending area. Collect the dynamic stress spectrum through an anti-aliasing filter and synchronously record the timestamp; Set the sensor range to ±10N and the resolution to 0.01N.

[0098] The anti-aliasing filter filters out high-frequency noise (such as motor vibration interference). The dynamic stress spectrum is recorded at a sampling rate of 1kHz, and the timestamp is synchronized for multi-source data alignment.

[0099] S24, Establish a multi-source data synchronization and fusion mechanism: When the loading platform triggers a deformation action, send a hardware synchronization pulse signal to the vision unit, fiber optic demodulator, and piezoelectric sensor to make the time deviation between the sequential images, strain data, and stress spectrum within the allowable range;

[0100] Establish a hardware-level multi-source data synchronization and fusion mechanism: When the main controller of the loading platform triggers a deformation action, send a TTL synchronization pulse signal to the vision unit, fiber optic demodulator, and piezoelectric sensor. Each acquisition module aligns the data time axis based on the synchronization pulse to ensure that the time deviation between the sequential images, strain data, and stress spectrum ≤ 0.1ms. The synchronization error is corrected in real-time through the time-domain cross-correlation algorithm to eliminate the phase mismatch caused by transmission delay.

[0101] S25, Preprocess the collected data: Perform non-uniform illumination correction and motion blur compensation on the sequential images, use median filtering to eliminate the instantaneous noise in the three-dimensional strain distribution dataset, and perform baseline drift correction and dimension normalization on the dynamic stress spectrum;

[0102] Apply median filtering to eliminate instantaneous noise and retain effective strain gradient information; correct baseline drift through polynomial fitting and normalize it to stress values according to the sensor sensitivity coefficient. The preprocessed data is stored in the cache area (DDR4, bandwidth 3200MHz) for subsequent analysis and call.

[0103] S26. Align the preprocessed surface images, three-dimensional strain distribution datasets, and dynamic stress spectra along the time axis to construct a multi-modal deformation feature database containing spatial coordinates, strain gradients, stress amplitudes, and texture feature vectors.

[0104] Align the preprocessed surface images, three-dimensional strain distribution datasets, and dynamic stress spectra along the time axis to construct a multi-modal deformation feature database. The database adopts a hierarchical storage structure:

[0105] Spatial coordinates: Record the crack position and the coordinates of the strain concentration area (XYZ);

[0106] Strain gradient: Store the axial, radial, and tangential strain gradient values;

[0107] Stress amplitude: Record the peak stress (σ_max) and valley stress (σ_min) of the dynamic stress spectrum;

[0108] Texture features: Extract parameters such as the crack direction angle (θ) and fractal dimension (D_f).

[0109] In this embodiment, specifically, step S3 specifically includes:

[0110] S31. Perform dynamic noise reduction processing on the time-series images, separate the background texture and crack pixels, and compensate for the image displacement artifacts caused by material deformation based on the optical flow method to generate a denoised and enhanced image set;

[0111] Perform dynamic noise reduction processing on the time-series images. Use an adaptive Gaussian mixture model to separate the background texture (such as material surface grain boundaries and filler particles) and crack pixels. Calculate the pixel displacement field between adjacent frames based on the optical flow method to compensate for the image displacement artifacts caused by material deformation. The denoised images are enhanced in contrast through histogram equalization to generate a denoised and enhanced image set. During the noise reduction process, the separation threshold of the background texture is dynamically adjusted according to the local gray variance to avoid losing crack details due to over-smoothing.

[0112] S32. Use the U-Net network to perform crack semantic segmentation on the denoised and enhanced image set, set the crack length threshold to 50μm, extract the crack skeleton lines of each frame and calculate the fractal dimension, and statistically calculate the total crack length per unit area to generate a crack density heat map;

[0113] Use the U-Net network to perform crack semantic segmentation on the denoised and enhanced image set. The network adds an attention mechanism module to improve the recognition rate of microcracks. The segmentation result is processed morphologically to extract the crack skeleton line. The box counting method is used to calculate the fractal dimension \(D_f\) to quantify the complexity of the crack morphology. The total crack length per unit area is statistically calculated to generate a crack density heat map, improving the spatial resolution.

[0114] S33. Analyze the crack skeleton line orientation through the histogram of oriented gradients, use the weighted transform to detect the main crack propagation direction, and calculate the anisotropy coefficient \(\eta = 1 - (\text{energy of the secondary dominant direction} / \text{energy of the main direction})\);

[0115] Analyze the crack skeleton line orientation through the histogram of oriented gradients, extract the local gradient directions (0° - 180°, with an interval of 10°) and statistically calculate the energy distribution. Use the weighted Hough transform to detect the main crack propagation direction, giving higher weights to long crack segments to avoid interference from short cracks. Calculate the anisotropy coefficient \(\eta = 1 - (\text{energy of the secondary dominant direction} / \text{energy of the main direction})\). If \(\eta < 0.3\), it is determined as a uniform damage mode; if \(\eta \geq 0.6\), it is a directional expansion mode. The angle \(\theta\) between the main direction and the material tensile axis is used to evaluate the correlation between the crack propagation path and the load direction.

[0116] S34. Perform spatial interpolation on the three-dimensional strain distribution dataset, reconstruct the three-dimensional deformation field containing axial, radial, and tangential strain components, calculate the strain gradient change amount, and identify the local areas where the strain gradient change rate exceeds 0.5% / mm² as strain concentration areas;

[0117] Specifically: Perform Kriging spatial interpolation on the three-dimensional strain distribution dataset, fit the spatial correlation based on the semi-variogram \(\gamma(h)=C0 + C1(1 - e^{(-h / a)})\), and reconstruct the three-dimensional deformation field containing axial, radial, and tangential strain components. Calculate the strain gradient change amount \(\nabla\varepsilon=\sqrt{(\frac{\partial\varepsilon_x}{\partial x})^2+(\frac{\partial\varepsilon_y}{\partial y})^2+(\frac{\partial\varepsilon_z}{\partial z})^2}\), and identify the local areas where the gradient change rate exceeds 0.5% / mm² as strain concentration areas. Mark the three-dimensional coordinates \((x, y, z)\) and gradient values of the strain concentration areas to generate a strain gradient heat map for subsequent energy density calculation and damage risk assessment.

[0118] S35. Map the coordinates of the strain concentration areas to the three-dimensional solid model of the material, calculate the strain energy density \(W = 0.5\sigma\cdot\varepsilon\) of each hot spot area, generate a strain energy density cloud map, and perform spatial registration with the crack density heat map;

[0119] Combine the strain \(\varepsilon\) and stress \(\sigma\) in the strain concentration areas to calculate the strain energy density and quantify the degree of local energy accumulation;

[0120] Three-dimensional entity mapping: Map the strain energy density values to the three-dimensional entity model of the material (such as a CAD model or a point cloud model) to generate an energy density contour map;

[0121] Then, align the coordinates of the energy density contour map and the crack density thermogram to locate the overlapping areas of high energy density and high crack density.

[0122] The simple strain gradient marker (S34) only reflects geometric deformation, while the energy density calculation (S35) reveals the internal energy distribution of the material, which is more directly related to the damage mechanism; Spatial registration establishes the physical connection between cracks (surface damage) and strain energy (internal energy), breaking through the limitations of traditional single data analysis.

[0123] S36: Establish a crack-strain coupling analysis matrix, mark the coordinates of the dangerous areas that simultaneously meet the conditions of crack density ≥ 0.8 mm / mm² and strain energy density ≥ 15 kJ / m³, and record their spatial distribution patterns and the change amount of strain gradient.

[0124] Use crack density (≥ 0.8 mm / mm²) and strain energy density (≥ 15 kJ / m³) as combined criteria to screen high-risk areas; Mark the coordinates of the areas that simultaneously meet the double thresholds, and record their spatial distribution patterns (such as banded distribution along the bending axis); Track the gradient change trend of the dangerous areas (such as the increasing rate of strain energy density over time) to predict the damage propagation direction.

[0125] A single index (only crack or only strain) may miss potential risks. For example:

[0126] Regions with high crack density but low energy: May be surface scratches, without the risk of deep damage;

[0127] Regions with high energy density but no cracks: May be potential damage initiation points;

[0128] The combined use of double thresholds can accurately identify the truly dangerous areas (i.e., the areas where high energy drives crack propagation), avoiding misjudgment.

[0129] In this embodiment, specifically, step S4 specifically includes:

[0130] S41: Perform time-domain segmentation processing on the dynamic stress response spectrum, divide the stress relaxation stage, steady-state holding stage, and elastic rebound stage based on the deformation loading characteristics, calculate the stress decay slope and creep recovery rate of each stage, and generate a modulus decay feature vector;

[0131] Perform time-domain segmentation processing on the dynamic stress response spectrum and divide it into three stages based on the deformation loading characteristics:

[0132] Stress relaxation stage: From the peak stress σ_max down to 90%σ_max, calculate the attenuation slope k1 = (σ_max - σ_90%) / Δt1;

[0133] Steady-state holding stage: The stress fluctuates within the range of ±5%σ_max, calculate the creep recovery rate η = (σ_end - σ_min) / σ_max;

[0134] Elastic rebound stage: The stress recovers from 90%σ_max to the initial value, calculate the rebound slope k2 = (σ_90% - σ_0) / Δt2.

[0135] Generate the modulus decay eigenvector V = [k1, η, k2] to characterize the viscoelastic behavior of the material.

[0136] S42, Fit the apparent elastic modulus decay curve using a piecewise exponential function, take the ratio of the peak stresses of adjacent deformation cycles as the decay factor, and solve the time-varying weight coefficient of the elastic modulus decay rate through the nonlinear least squares method;

[0137] Fit the apparent elastic modulus decay curve using a piecewise exponential function:

[0138] Function form: E(t) = E0·[A·exp(-α1t) + (1 - A)·exp(-α2t)], where A is the weight coefficient, and α1, α2 are the decay rates;

[0139] Decay factor: Take the ratio of the peak stresses of adjacent cycles r = σ_max,i / σ_max,i+1 as the decay factor, and trigger the decay rate adjustment when r < 0.95;

[0140] Parameter solution: Fit α1, α2, A through the Levenberg-Marquardt algorithm (maximum iteration 200 times, tolerance 0.1%) to generate the time-varying weight coefficient matrix.

[0141] This model can accurately describe the nonlinear modulus decay behavior of materials under cyclic loading.

[0142] S43, Extract the spatial distribution heat map of the crack propagation rate, calculate the cosine value of the angle between the main crack direction and the strain gradient vector, and mark the three-dimensional coordinate point set where the strain gradient change exceeds the preset change threshold; The steps reveal the correlation between the crack propagation path and the strain field distribution.

[0143] Rate calculation: Based on the results of the time-series image segmentation, calculate the crack length increment da / dt per unit time;

[0144] Direction analysis: Extract the main crack direction θ_c through the Histogram of Oriented Gradients (HOG), and calculate the cosine value of the angle between it and the strain gradient vector θ_s, cos(θ_c - θ_s);

[0145] Threshold marker: When the change rate of strain gradient Δ(∇ε) / Δt > 0.2% / mm² / s, mark the corresponding three-dimensional coordinate point set (x, y, z).

[0146] S44. Construct a performance degradation differential equation that includes the modulus decay rate, crack-strain coupling factor, and local gradient mutation, and use the adaptive iteration method to solve the material damage accumulation rate; comprehensively consider the multi-field coupling effects of modulus decay, crack propagation, and strain gradient mutation.

[0147] Equation form: dD / dt = k1·ΔE + k2·(da / dN) + k3·∇ε, where D is the damage degree, and k1, k2, and k3 are weight coefficients.

[0148] Coupling factor: The crack-strain coupling factor k2 = cos(θ_c - θ_s)·(da / dN), which reflects the dependence of crack propagation on local strain.

[0149] Iterative solution: Use the adaptive step Runge-Kutta method (initial step size 1s, error limit 0.001%) to solve the damage accumulation rate dD / dt.

[0150] S45. Reconstruct the accelerated aging test parameters based on the typical load spectrum of the target scenario, substitute the damage accumulation rate into the Paris-Erdogan equation, and establish a cyclic life prediction model considering the multiaxial stress state.

[0151] S46. Calculate the confidence interval of the life prediction value by sampling and simulating the material performance degradation path. When the interval width exceeds the preset threshold, trigger the model parameter dynamic calibration mechanism. This step ensures the reliability of the life prediction result and avoids misjudgment caused by parameter uncertainty.

[0152] Parameter distribution: Set the statistical distributions (normal distribution, coefficient of variation 10%) of the crack propagation rate C, n, and the strain gradient correction coefficient α.

[0153] Life prediction: Generate 10,000 groups of random parameter combinations, and calculate the mean μ and standard deviation σ of the life distribution.

[0154] Confidence interval: When the interval width (2σ / μ) > 20%, trigger the model parameter dynamic calibration mechanism and refit C, n, and α.

[0155] In this embodiment, specifically, step S5 specifically includes:

[0156] S51. Normalize the crack density, direction consistency coefficient, and apparent elastic modulus decay rate. Convert the crack density to the relative damage degree in the range of 0-1. Calibrate the matching index of the direction consistency coefficient according to the cosine similarity, and map the modulus decay rate to the preset deterioration level.

[0157] S52. Based on the failure mode library of the target application scenario, use the analytic hierarchy process to allocate weight coefficients. For example, in the foldable scenario, the weight of the crack density is 0.5, the direction consistency is 0.3, and the modulus decay is 0.2; in the automotive sealing scenario, the weight of the modulus decay is 0.6, the crack density is 0.25, and the direction consistency is 0.15.

[0158] S53. Build a dynamic flexible performance index calculation model: perform a linear weighted sum of the normalized parameters and the weight coefficients, and superimpose the non-linear correction term of the strain energy density gradient to generate a comprehensive performance score.

[0159] Linear weighted sum: DFI_base = w1·RDI + w2·DCI + w3·(1 - ΔE / E0), where w1, w2, and w3 are weight coefficients.

[0160] Non-linear correction term: Introduce the exponential correction term of the strain energy density gradient ∇W, DFI = DFI_base + λ·exp(∇W), and the correction coefficient λ = 0.05 to suppress the misjudgment caused by local energy concentration.

[0161] Comprehensive performance score: The DFI value ranges from 0 to 1. DFI ≥ 0.85 indicates excellent performance, and DFI < 0.6 indicates a high failure risk.

[0162] This model combines linear and non-linear terms to take into account both global performance and local damage characteristics.

[0163] S54. Set the fitness threshold according to the historical test data and the scenario working condition limit value. For example, the threshold for the foldable scenario is set to 0.85 (corresponding to no fracture after 5000 cycles), and the threshold for the automotive sealing scenario is set to 0.75 (corresponding to the vibration fatigue life of 10^7 times).

[0164] S55. Compare the comprehensive score with the preset threshold, and perform a hierarchical evaluation decision according to the comparison result.

[0165] For example, when the comprehensive score ≥ the threshold, it is marked as "scenario adaptation", when the score is in the range of the threshold - 20%, it is marked as "critical monitoring", and when the score < the threshold - 20%, it is marked as "failure risk".

[0166] S56. Generate a hierarchical evaluation map according to the hierarchical evaluation decision. Divide the material surface into several grid units, assign red-yellow-green three-color warning signs according to the performance scores in the grid units, and mark the spatial coordinates and weight rankings of the priority improvement areas.

[0167] In this embodiment, specifically, after step S5, it further includes:

[0168] S6. Integrate the hierarchical evaluation results with the deformation field distribution and life prediction data to generate a visual inspection document including a three-dimensional damage evolution map, process optimization suggestions, and a scenario adaptation report, and transmit it to the terminal interaction interface through an encrypted channel.

[0169] Embodiment 2:

[0170] Combined with Figure 3 As shown in the figure, the present invention also provides a performance detection device for polyurethane materials, which is used for the performance detection method of polyurethane materials as in Embodiment 1. The performance detection device includes:

[0171] A constant temperature environmental chamber 10, in which a laser alignment instrument is provided.

[0172] A multi-axis deformation loading platform 20, which is provided with a pneumatic fixture 21, an axial tensile unit 22, a multi-directional bending module 23, and a servo motor 24.

[0173] A multi-modal sensing module 30, including a vision unit, a temperature sensing array, a piezoelectric mechanical sensor, and a distributed grating array.

[0174] A host module 40, which is provided with a data processing unit for processing the acquired temperature data, sequential images, strain distribution data, and dynamic stress response spectra to generate a multi-modal deformation feature data set.

[0175] A decision-making system for carrying a data processing model and performing hierarchical evaluation on the material.

[0176] As mentioned above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A method for detecting the performance of a polyurethane material, characterized in that, Including: In a preset detection environment, a strip-shaped polyurethane material to be tested is clamped on a multi-axis deformation loading platform, and the force acting in the target application scenario is simulated through coordinated control, and the real-time temperature change during the deformation process is recorded; A vision unit is used to capture the sequential images of the microscopic morphology of the material surface. At the same time, the internal strain distribution data is collected through a distributed strain sensing array, and the dynamic stress response spectrum is obtained by combining a piezoelectric mechanical sensor to generate a multi-modal deformation feature data set; Perform crack density and texture orientation analysis on the sequential images, extract the crack propagation rate and the direction consistency coefficient. At the same time, perform three-dimensional deformation field reconstruction on the internal strain data, and mark the coordinates of the strain concentration area and the strain gradient change amount; Calculate the apparent elastic modulus decay rate based on the dynamic stress response spectrum, combine the crack propagation rate and the strain gradient change amount to establish a material performance degradation model, and correlate the cyclic life prediction equation of the target application scenario; Generate a dynamic flexibility performance index through a weighted fusion algorithm according to the crack density, direction consistency coefficient and apparent elastic modulus decay rate, and grade and evaluate the material according to the preset scenario fitness threshold.

2. The performance detection method of the polyurethane material according to claim 1, characterized in that, The step of, in a preset detection environment, clamping a strip-shaped polyurethane material to be tested on a multi-axis deformation loading platform, simulating the force acting in the target application scenario through coordinated control, and recording the real-time temperature change during the deformation process specifically includes: Build a multi-axis deformation loading platform in a constant temperature incubator, clamp both ends of the strip-shaped polyurethane material to be tested with pneumatic clamps. The surface of the pneumatic clamps is coated with a polytetrafluoroethylene anti-slip layer, and the spatial angle deviation between the material clamping axis and the platform reference plane is calibrated by a laser alignment instrument; Set the loading mode based on the working conditions of the target application scenario; Drive the axial tension unit through a servo motor to apply a linear tensile stress at a preset tensile rate. At the same time, control the multi-directional bending module by a pneumatic drive device to generate a dynamic bending deformation in the middle section of the material; Use a closed-loop feedback control system to synchronously adjust the phase relationship of the stretching, bending and pressing actions, and trigger a protection mechanism according to the phase relationship. The protection mechanism includes: when the yield point of the material is detected, automatically reduce the tensile rate to 50% of the set value, and trigger the indenter retraction protection mechanism when the bending angle reaches the threshold; Mount a temperature sensing array on the material surface, record the temperature fluctuation data during deformation detection at a preset sampling frequency, and generate a temperature-deformation time series correlation curve.

3. The performance detection method of the polyurethane material according to claim 2, characterized in that, The temperature sensing array is composed of 16 thin film sensors, and the 16 thin film sensors cover the stretching area, bending area and pressing point in a serpentine wiring layout; among them, the thin film sensor is a PT100 thin film sensor.

4. The performance detection method of the polyurethane material according to claim 1, characterized in that, The step of using a vision unit to capture the sequential images of the microscopic morphology of the material surface, synchronously collecting the internal strain distribution data through a distributed strain sensing array, combining a piezoelectric mechanical sensor to obtain the dynamic stress response spectrum, and generating a multi-modal deformation feature data set specifically includes: Deploy an imaging system on the material surface, use a polarized light source to irradiate the area to be tested at a preset incident angle, adjust the polarization direction to form a preset angle with the material stretching axis, and capture the sequential images of the surface microscopic morphology; Embed a distributed grating array inside the material, and use a demodulator to monitor the strain values of each grating node in real time to generate a three-dimensional strain distribution dataset in the axial, radial, and tangential directions; Install piezoelectric mechanical sensors at the material clamping end and the bending area, collect the dynamic stress spectrum through an anti-aliasing filter, and synchronously record the time stamps.

5. The performance detection method of the polyurethane material according to claim 4, characterized in that After collecting the dynamic stress spectrum through the anti-aliasing filter and synchronously recording the time stamps, the following steps are further included: Establish a multi-source data synchronization and fusion mechanism: when the loading platform triggers a deformation action, send a hardware synchronization pulse signal to the vision unit, fiber optic demodulator, and piezoelectric sensor to make the time deviation between the sequential images, strain data, and stress spectrum within the allowable range; Preprocess the collected data: perform non-uniform illumination correction and motion blur compensation on the sequential images, use median filtering to eliminate the instantaneous noise in the three-dimensional strain distribution dataset, and perform baseline drift correction and dimension normalization on the dynamic stress spectrum; Align the preprocessed surface images, three-dimensional strain distribution dataset, and dynamic stress spectrum along the time axis to construct a multi-modal deformation feature database containing spatial coordinates, strain gradients, stress amplitudes, and texture feature vectors.

6. The performance detection method of the polyurethane material according to claim 5, characterized in that, Performing crack density and texture orientation analysis on the sequential images, extracting the crack propagation rate and direction consistency coefficient, and at the same time performing three-dimensional deformation field reconstruction on the internal strain data, marking the coordinates and strain gradient change amounts of the strain concentration regions, specifically including: Perform dynamic noise reduction processing on the sequential images, separate the background texture and crack pixels, and compensate for the image displacement artifacts caused by material deformation based on the optical flow method to generate a denoised and enhanced image set; Use the U-Net network to perform crack semantic segmentation on the denoised and enhanced image set, set the crack length threshold to 50μm, extract the crack skeleton lines of each frame and calculate the fractal dimension, and statistically calculate the total crack length per unit area to generate a crack density heat map; Analyze the crack skeleton line direction through the histogram of oriented gradients, use weighted transformation to detect the main crack propagation direction, and calculate the direction consistency coefficient η = 1 - energy of the secondary dominant direction / energy of the main direction; Perform spatial interpolation on the three-dimensional strain distribution dataset, reconstruct a three-dimensional deformation field containing axial, radial, and tangential strain components, calculate the strain gradient change amount, and identify the local regions where the strain gradient change rate exceeds 0.5% / mm² as strain concentration regions; Map the coordinates of the strain concentration regions to the three-dimensional solid model of the material, calculate the strain energy density W = 0.5σ·ε of each hot spot region, generate a strain energy density cloud map and perform spatial registration with the crack density heat map; where, ε is the strain in the strain concentration region, and σ is the stress in the strain concentration region; Establish a crack-strain coupling analysis matrix, mark the coordinates of the dangerous regions that simultaneously satisfy the crack density ≥ 0.8mm / mm² and the strain energy density ≥ 15kJ / m³, and record their spatial distribution patterns and strain gradient change amounts.

7. The performance detection method of the polyurethane material according to claim 1, characterized in that, Calculating the apparent elastic modulus decay rate based on the dynamic stress response spectrum, combining the crack propagation rate and the strain gradient change amount, establishing a material performance degradation model, and correlating with the cycle life prediction equation of the target application scenario, specifically including: Perform time-domain segmentation processing on the dynamic stress response spectrum, divide the stress relaxation stage, steady-state holding stage, and elastic rebound stage based on the deformation loading characteristics, calculate the stress attenuation slope and creep recovery rate in each stage, and generate a modulus attenuation feature vector; Use a piecewise exponential function to fit the apparent elastic modulus attenuation curve, take the peak stress ratio of adjacent deformation cycles as the attenuation factor, and solve the time-varying weight coefficient of the apparent elastic modulus attenuation rate by the nonlinear least squares method; Extract the spatial distribution heat map of the crack propagation rate, calculate the cosine value of the angle between the crack main direction and the strain gradient vector, and mark the three-dimensional coordinate point set where the strain gradient change amount exceeds the preset change threshold; Construct a performance degradation differential equation including the apparent elastic modulus attenuation rate, crack-strain coupling factor, and local gradient mutation, and use the adaptive iteration method to solve the material damage accumulation rate; Reconstruct the accelerated aging test parameters based on the typical load spectrum of the target scenario, substitute the damage accumulation rate into the Paris-Erdogan equation, and establish a cyclic life prediction model considering the multiaxial stress state; Simulate the material performance degradation path by sampling, calculate the confidence interval of the life prediction value, and trigger the model parameter dynamic calibration mechanism when the interval width exceeds the preset threshold; 8. The performance detection method of the polyurethane material according to claim 1, characterized in that Generate a dynamic flexibility performance index through a weighted fusion algorithm according to the crack density, direction consistency coefficient, and apparent elastic modulus attenuation rate, and perform a hierarchical evaluation of the material based on the preset scenario fitness threshold, specifically including: Normalize the crack density, direction consistency coefficient, and apparent elastic modulus attenuation rate, convert the crack density into a relative damage degree in the 0-1 interval, calibrate the matching index of the direction consistency coefficient according to the cosine similarity, and map the apparent elastic modulus attenuation rate to the preset degradation level; Based on the failure mode library of the target application scenario, use the analytic hierarchy process to assign weight coefficients; Construct a dynamic flexibility performance index calculation model: perform linear weighted summation of the normalized parameters and weight coefficients, and superimpose the nonlinear correction term of the strain energy density gradient to generate a comprehensive performance score; Set the fitness threshold according to the historical test data and the scenario working condition limit value; Compare the comprehensive score with the preset threshold, and execute the hierarchical evaluation decision according to the comparison result; Generate a hierarchical evaluation map according to the hierarchical evaluation decision, divide the material surface into several grid units, assign red-yellow-green three-color warning signs according to the performance scores within the grid units, and mark the spatial coordinates and weight rankings of the priority improvement areas; 9. The performance detection method of the polyurethane material according to claim 1, wherein After performing the hierarchical evaluation of the material according to the preset scenario fitness threshold, it further includes: Integrate the hierarchical evaluation results with the deformation field distribution and life prediction data to generate a visual inspection document including a three-dimensional damage evolution map, process optimization suggestions, and a scenario fitness report, and transmit it to the terminal interaction interface through an encrypted channel; 10. A performance detection device for a polyurethane material, characterized in that, A performance detection method for the polyurethane material according to any one of claims 1 to 9, wherein the performance detection device includes: A constant temperature environmental chamber, in which a laser alignment instrument is provided; A multiaxial deformation loading platform, which is provided with a pneumatic fixture, an axial tension unit, a multi-directional bending module, and a servo motor; The multimodal sensing module includes a vision unit, a temperature sensing array, a piezoelectric mechanical sensor, and a distributed grating array; The data processing unit is used to process the acquired temperature data, sequential images, strain distribution data, and dynamic stress response spectra to generate a multimodal deformation feature dataset; The decision-making system is used to carry a data processing model and conduct a hierarchical evaluation of the material.

Images