A method and system for detecting sunroof display films

Abstract

This invention relates to the field of optical thin film inspection technology, specifically disclosing a method and system for inspecting skylight display thin films. The method includes: applying a frequency-converting electric field under controllable humidity partial voltage modulation, simultaneously acquiring picosecond photoelastic imaging data to construct a dielectric relaxation-photoelastic coupling field, identifying interface debonding precursors and encoding them as a subcritical defect entropy state matrix; using the spatial coordinates of this matrix as the focusing point to excite a multi-physics field phase-locked coherent superposition signal, reconstructing a three-band energy dissipation topology network, identifying topological defect nodes and registering them with the entropy state matrix to generate a failure prediction map; fusing the coupling field and the topology network to establish a service limit extrapolation model based on the self-organized criticality of dissipative structures, generating a personalized adaptive compensation strategy, and encapsulating it into a quality adaptive digital twin that can be embedded in the thin film. This invention achieves a leap from post-event verification to pre-event prediction, providing a new paradigm for the reliability assessment of intelligent display dimming films.

Description

Technical Field

[0001] This invention relates to the field of optical thin film testing technology, and specifically to a testing method and system for a skylight display thin film. Background Technology

[0002] With the rapid development of intelligent vehicles and building curtain wall technologies, intelligent thin films that combine dimming and display functions have become key components for improving user experience and achieving energy conservation and emission reduction. Currently, mainstream dimming technologies mainly include PDLC (polymer dispersed liquid crystal), DLC (dye liquid crystal), SPD (suspended particle), and EC (electrochromic). Among them, PDLC technology achieves switching between transparent and hazy states by controlling the arrangement of liquid crystal molecules through an electric field, but suffers from problems such as high haze in the transparent state and limited heat insulation performance; DLC technology utilizes the host-guest effect to adjust light transmittance and has been applied to sunroofs in many car models; SPD technology adjusts luminous flux through the orientation of suspended particles, but has a high driving voltage and severe color distortion; EC technology can achieve color changes, but its response speed is slow. In recent years, the emergence of multi-color microcrystalline dielectric dimming technology, through independent research and development of three core material systems, has achieved three-dimensional adjustment of brightness, color, and opacity, and can integrate full-color display functions, representing the cutting-edge development direction of the industry.

[0003] However, existing methods for detecting such thin films have fundamental limitations. Current optical parameter testing only measures static transmittance and haze. While environmental reliability testing covers high and low temperatures, humidity, and xenon lamp aging, these are all post-test verification methods under single stress—that is, removing the film after a certain period of stress and measuring its performance changes or observing physical defects. These methods can only answer whether the film has failed, but cannot answer when it will fail or where the failure will begin. Although on-off cycle testing involves electrical actuation, it only focuses on changes in static parameters before and after the test, ignoring the real-time evolution of performance during dynamic processes.

[0004] More importantly, the aforementioned detection methods are all based on an implicit assumption: that thin film failure is uniform and can be captured by changes in macroscopic parameters. However, in practical applications, thin film failure under complex operating conditions often begins with subcritical debonding at the microcrystal-ligand interface. At this stage, the macroscopic optical properties of the thin film may not have changed significantly, while microscopic damage has already accumulated locally. Existing technologies are completely unable to capture this subcritical state—that is, the stage where the interface begins to debond but has not yet formed macroscopic defects—let alone reveal the cross-scale evolution path of the failure mechanism under multi-physics coupling. As dimming thin films upgrade from simple brightness adjustment to full-color dynamic display functions, the requirements for thin films to maintain color gamut accuracy and response consistency under dynamic operating conditions are increasing, and this fundamental limitation of existing detection methods has become a technical bottleneck restricting the development of the industry.

[0005] Therefore, how to predict subcritical failure, reveal the mechanism of multi-field coupled failure, and generate personalized compensation strategies have become technical problems that urgently need to be solved in this field. Summary of the Invention

[0006] The purpose of this invention is to provide a detection method and system for sunroof display films, so as to solve the technical problem that the subcritical failure behavior of smart dimming films under complex environments cannot be predicted in the prior art.

[0007] To solve the above-mentioned technical problems, the present invention specifically provides the following technical solution: A method for testing a sunroof display film includes the following steps: S1. Apply a frequency-converting electric field to the thin film under controllable wet partial voltage modulation, simultaneously acquire picosecond photoelastic imaging data, construct a dielectric relaxation-photoelastic coupling field, and identify interface debonding precursors based on the non-cooperative evolution region of birefringence and wave velocity, and encode them as a subcritical defect entropy state matrix. S2. Using the spatial coordinates in the entropy state matrix as the focal point, photoacoustic, thermoacoustic and ultrasonic guided wave signals are simultaneously excited to form a phase-locked coherent superposition detection mechanism. The amplitude and strain energy spectrum are extracted by phase-locked demodulation, the three-band energy dissipation topology network is reconstructed, the topology defect nodes are identified and registered with the entropy state matrix, and a cross-physical field coupling failure prediction map is generated. S3. Integrate the coupled field and the topology network to establish a service limit extrapolation model based on the self-organized criticality of dissipative structures, and generate an adaptive compensation strategy based on the entropy state matrix and the distribution of topological defect nodes. Encapsulate the service limit, compensation strategy and failure map into a quality adaptive digital twin.

[0008] As a preferred embodiment of the present invention, S1 specifically includes: S11. The thin film sample to be tested is encapsulated in the detection chamber, and a controllable moisture partial pressure modulated by a sine wave is introduced into the chamber. The frequency and amplitude of the sine wave simulate the interlayer water vapor permeation-desorption cycle caused by the breathing effect of day and night temperature difference under the edge condition of the skylight seal failure. S12. While the wet partial voltage is modulated, an alternating electric field with a frequency ranging from Hertz to kilohertz is applied to the thin film electrode in a logarithmic scan, covering the response frequency band from the microcrystalline medium interface polarization to the ligand dipole orientation. S13. Synchronously trigger a picosecond-level time-resolved pump-probe photoelastic imaging system to collect the spatial distribution of birefringence delay and ultrasonic surface wave velocity field distribution under each electric field frequency-wet partial voltage coupling state, and construct a dielectric relaxation-photoelastic coupling field. S14. Identify the interface debonding precursors in the coupled field, extract the spatial coordinates, the coupling phase difference between the current electric field frequency and the wet voltage modulation waveform, and the relaxation time distribution of the birefringence delay, and encode the three into a subcritical defect entropy state matrix.

[0009] As a preferred embodiment of the present invention, S14 specifically includes: S141. Traverse each spatial pixel in the dielectric relaxation-photoelastic coupling field, calculate the Pearson correlation coefficient between the birefringence delay time series and the ultrasonic surface wave velocity time series of the pixel under all electric field frequency-wet partial voltage coupling states, and mark the pixel with the correlation coefficient lower than the preset threshold as a candidate non-cooperative evolution region. S142. Perform morphological closing operation on the candidate non-cooperative evolution region, merge adjacent pixels to form a connected region, and remove isolated noise points with an area smaller than the minimum identifiable defect size to obtain the spatial coordinate set of the effective interface debonding precursor region. S143. For each effective interface debonding precursor region, extract the electric field frequency value and wet voltage phase value when the birefringence delay at its center point first deviates from the reference value by more than three times the standard deviation, and calculate the phase difference between the two and the zero point of the wet voltage modulation waveform as the coupling phase difference. S144. Perform exponential fitting on the birefringence delay time series at the center point of the region, calculate the time required for it to decay from the peak to 1 / e as the relaxation time, and statistically analyze the distribution histogram of relaxation time of all pixels in the region. S145. Encode the spatial coordinate set, coupled phase difference, and relaxation time distribution histogram into a three-dimensional data structure and store it as a subcritical defect entropy state matrix.

[0010] As a preferred embodiment of the present invention, S2 specifically includes: S21. The spatial coordinates in the subcritical defect entropy state matrix encoded in S1 are used as multi-physics field focusing points. At the points, the narrowband photoacoustic signal induced by Q-switched pulse laser, the dielectric thermoacoustic signal induced by electric field frequency modulation, and the broadband ultrasonic guided wave signal injected by the external piezoelectric transducer are synchronously excited, so that the three signals form a phase-locked coherent superposition detection mechanism in the film thickness direction. S22. Using the dielectric thermoacoustic signal as a phase reference, the narrowband photoacoustic signal and the broadband ultrasonic guided wave signal are phase-locked demodulated to extract the displacement amplitude spectrum and strain energy density spectrum of the phase-locked coherent superposition detection mechanism. S23. Based on the dispersion characteristics of the displacement amplitude spectrum and the attenuation characteristics of the strain energy density spectrum, the three-band energy dissipation topology network of the thin film in the orientation polarization band of the microcrystalline medium, the viscoelastic relaxation band of the ligand material, and the resonance band of the whole film structure is reconstructed. S24. Identify nodes in the topology network whose energy flow convergence exceeds the critical dissipation threshold as topology defect nodes, spatially register the nodes with the subcritical defect entropy state matrix, and generate a cross-physical field coupling failure prediction map.

[0011] As a preferred embodiment of the present invention, S23 specifically includes: S231. From the displacement amplitude spectrum extracted in S22, the dispersion curves of the three characteristic frequency bands corresponding to the orientation polarization frequency band of the microcrystalline medium, the viscoelastic relaxation frequency band of the ligand material, and the resonance frequency band of the whole film structure are separated, and the attenuation coefficient spectrum corresponding to the three characteristic frequency bands is separated from the strain energy density spectrum. S232. Using the microstructure units of the thin film sample as nodes and the energy transfer efficiency between adjacent units as edges, calculate the vibration mode participation coefficient of each node in different frequency bands based on the dispersion curve, calculate the energy dissipation weight of each edge based on the attenuation coefficient spectrum, and construct an initial three-band energy dissipation topology network. S233. Perform spectral clustering analysis on the initial three-band energy dissipation topology network to identify the main and side paths of energy dissipation in each frequency band, remove redundant edges with weights lower than a preset threshold, and obtain the optimized three-band energy dissipation topology network.

[0012] As a preferred embodiment of the present invention, S24 specifically includes: S241. Traverse all nodes in the three-band energy dissipation topology network and calculate the energy flow convergence degree of each node. The energy flow convergence degree is the ratio of the sum of edge weights flowing into the node to the sum of edge weights flowing out of the node. Mark nodes whose energy flow convergence degree exceeds a preset critical dissipation threshold as candidate topology defect nodes. S242. Map the spatial coordinates in the subcritical defect entropy state matrix described in S1 to the node coordinate system of the three-band energy dissipation topology network, calculate the Euclidean distance between each candidate topology defect node and the spatial coordinates of the nearest subcritical defect entropy state matrix, and confirm the candidate nodes whose distance is less than the preset spatial association threshold as valid topology defect nodes. S243. For each valid topological defect node, trace its energy input path and dissipation path in the three-band energy dissipation topology network, and mark all nodes and edges on the path as potential failure propagation channels. S244. The effective topological defect nodes, potential failure propagation channels and their corresponding energy flow convergence and spatial coordinate registration information are superimposed onto the spatial structure map of the thin film sample to generate a cross-physical field coupled failure prediction map.

[0013] As a preferred embodiment of the present invention, S3 specifically includes: S31. By integrating the dielectric relaxation-optical elastic coupling field constructed in S1 with the three-band energy dissipation topology network reconstructed in S2, a service limit extrapolation model is established based on the self-organization criticality of the dissipation structure. This model uses the fluctuation amplitude of the non-equilibrium energy dissipation intensity index within the wet partial voltage modulation cycle as a measure of the degree of deviation from the equilibrium state, and uses the connectivity evolution of the three-band energy dissipation topology network as a measure of the change in the orderliness of the dissipation structure. S32. By tracking the critical point of the transition from the near-equilibrium linear region to the dissipative structure bifurcation region, the service limit number of the thin film under the cyclic breathing-light-electric drive coupling condition is extrapolated. S33. Based on the distribution density of interface debonding precursors and the spatial aggregation degree of topological defect nodes in the subcritical defect entropy state matrix, a personalized adaptive compensation strategy for the thin film is generated. This strategy includes frequency-phase pre-modulation parameters of the driving electric field to avoid dielectric loss peaks, local humidity barrier layer thickness gradient design parameters to suppress edge debonding propagation, and microcrystalline medium orientation axis pre-tilt angle distribution parameters to counteract moisture expansion anisotropy. S34. The service limit cycle number, the parameter set of the adaptive compensation strategy, and the sparse feature vector of the cross-physical field coupling failure prediction spectrum are encapsulated into a quality adaptive digital twin that can be embedded in a thin-film flexible integrated circuit.

[0014] As a preferred embodiment of the present invention, S32 specifically includes: S321. From the service limit extrapolation model established in S31, extract the time series data of the non-equilibrium energy dissipation intensity index within the wet partial voltage modulation cycle, perform sliding window variance analysis on the time series data of the non-equilibrium energy dissipation intensity index, calculate the fluctuation amplitude of the non-equilibrium energy dissipation intensity index within each window, and generate a time series sequence of the fluctuation amplitude of the non-equilibrium energy dissipation intensity index. S322. Simultaneously, from the service limit extrapolation model, extract the connectivity evolution data of the three-band energy dissipation topology network, where connectivity is the reciprocal of the average shortest path length between effective nodes in the network, and generate a connectivity time series. S323. Perform a joint analysis on the time series of fluctuation amplitude of the non-equilibrium energy dissipation intensity index and the time series of connectivity, calculate the cross-correlation function of the two on the time axis, and mark the time point when the cross-correlation function value first exceeds the preset critical correlation threshold as the critical point of the dissipation structure transitioning from the near-equilibrium linear region to the bifurcation region. S324. Based on the number of cycles corresponding to the critical point, and combined with the conversion relationship between the preset accelerated aging factor and the measured working conditions, extrapolate the service limit number of cycles of the film under the cyclic breathing-light irradiation-electric drive coupling condition.

[0015] As a preferred embodiment of the present invention, S33 specifically includes: S331. Read the distribution density data of interface debonding precursors in the subcritical defect entropy state matrix described in S1, and identify spatial regions with distribution density exceeding a preset threshold as high-risk areas; at the same time, read the spatial aggregation data of topological defect nodes described in S2, identify spatial regions with aggregation degree exceeding a preset threshold as energy dissipation anomaly areas, and take the spatial intersection of the two as the priority compensation area. S332. For the priority compensation region, extract the electric field frequency-wet voltage coupling phase difference recorded in S1, calculate the center value and dispersion of its statistical distribution, and generate the frequency-phase pre-modulation parameters of the driving electric field. The parameters enable the driving electric field to actively avoid the frequency-phase coupling interval corresponding to the dielectric loss peak when working. S333. Based on the spatial distribution characteristics of the priority compensation region at the edge of the film, and combined with the energy flow convergence direction of the region in the three-band energy dissipation topology network reconstructed in S2, the thickness gradient design parameters of the local humidity barrier layer are generated. The parameters make the barrier layer gradually thicken in the direction of edge debonding and expansion path to suppress expansion. S334. Based on the current distribution state of the orientation axis of the microcrystalline medium in the priority compensation region, and combined with the anisotropic evolution characteristics of the birefringence retardation in S1, a pre-tilt angle distribution parameter of the orientation axis of the microcrystalline medium is generated. The parameter causes the orientation axis to form a compensatory deflection in the main direction of wet expansion to counteract the anisotropy of wet expansion.

[0016] A detection system for a sunroof display film, used to implement a detection method for a sunroof display film, comprising: The multi-field coupling excitation and subcritical entropy state encoding module is used to apply a frequency-converting electric field to the thin film under controllable wet partial voltage modulation, and simultaneously acquire picosecond photoelastic imaging data to construct a dielectric relaxation-photoelastic coupling field, identify interface debonding precursors and encode them as a subcritical defect entropy state matrix. The multi-physics phase-locked superposition detection and energy dissipation topology reconstruction module is connected to the multi-field coupled excitation and subcritical entropy state encoding module. It is used to excite multi-physics phase-locked coherent superposition signals with the spatial coordinates in the subcritical defect entropy state matrix as the focusing points, extract the amplitude spectrum and strain energy spectrum through phase-locked demodulation, reconstruct the three-band energy dissipation topology network, identify topology defect nodes and register them with the entropy state matrix to generate a cross-physics coupling failure prediction map. The dissipative structure modeling and adaptive digital twin module is connected to the multi-field coupled excitation and subcritical entropy state encoding module and the multi-physics phase-locked superposition detection and energy dissipation topology reconstruction module. It is used to fuse the coupled fields and topology network to establish a service limit extrapolation model based on the self-organized criticality of dissipative structures, generate personalized adaptive compensation strategies, and encapsulate the service limit, compensation strategy and failure prediction map into a quality adaptive digital twin.

[0017] Compared with the prior art, the present invention has the following advantages: 1. This invention achieves dynamic capture of the precursors to debonding at the microcrystal-ligand interface by constructing a dielectric relaxation-photoelastic coupling field. The spatial coordinates, coupling phase difference, and relaxation time distribution of the precursors to interface debonding are encoded into a subcritical defect entropy state matrix, enabling the detection system to see the initiation of microscopic interface damage before macroscopic failure occurs, thus filling the technical gap in the detection of subcritical states of intelligent dimming films.

[0018] 2. This invention reconstructs a three-band energy dissipation topology network using a multi-physics phase-locked coherent superposition technique involving acoustic, optical, and electrical fields. It identifies topological defect nodes where the energy flow convergence exceeds the critical dissipation threshold and spatially registers them with the subcritical defect entropy state matrix. This establishes a complete causal chain visualization map from the interface debonding initiation to energy dissipation anomalies and then to the macroscopic failure channel, revealing the cross-scale migration path of energy flow during the failure process.

[0019] 3. This invention creates a service limit extrapolation model based on the self-organization criticality of dissipative structures. By tracking the critical transition point from the near-equilibrium linear region to the dissipative structure bifurcation region, it realizes a paradigm shift from empirical statistics to physical mechanism-driven lifetime prediction. At the same time, it generates personalized adaptive compensation strategies based on defect distribution, realizing an intelligent leap from a unified standard to a one-membrane-one-policy approach. Finally, it is packaged into a quality adaptive digital twin that can be embedded in thin films, enabling the thin films to have self-sensing, self-diagnostic, and self-optimizing full life cycle intelligent management capabilities. Attached Figure Description

[0020] To more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings in the following description are merely exemplary, and those skilled in the art can derive other embodiments based on the provided drawings without creative effort.

[0021] Figure 1 This is a flowchart illustrating the method described in Embodiment 1 of the present invention.

[0022] Figure 2 This is a framework diagram of the system described in Embodiment 2 of the present invention. Detailed Implementation

[0023] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0024] The concepts involved in this application will first be described with reference to the accompanying drawings. It should be noted that the following descriptions of various concepts are only for the purpose of making the content of this application easier to understand and do not constitute a limitation on the scope of protection of this application; furthermore, the embodiments and features in the embodiments of this application can be combined with each other unless otherwise specified. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0025] Example 1 like Figure 1 As shown, the present invention provides a method for detecting a sunroof display film, comprising the following steps: S1. A frequency-converted electric field is applied to the thin film under test under controllable wet partial voltage modulation, and picosecond photoelastic imaging data is acquired simultaneously to construct a dielectric relaxation-photoelastic coupling field. Based on the non-cooperative evolution region of birefringence and wave velocity, interface debonding precursors are identified and encoded as a subcritical defect entropy state matrix; specifically including: S11. Establishment and packaging of a controllable moisture partial voltage modulation environment, specifically: The thin film sample to be tested is placed in a detection chamber with an airtight structure. The detection chamber is equipped with a humidity generation module and a dew point monitoring module. The humidity generation module splits the dry carrier gas into two paths. One path is saturated and humidified by a constant temperature deionized water bath, while the other path remains dry. The two gases are dynamically mixed by a proportional valve and then introduced into the chamber to achieve continuous adjustment of the partial pressure of moisture. The dew point monitoring module is based on the principle of a cold mirror dew point meter and provides real-time feedback on the actual water vapor partial pressure value in the chamber.

[0026] The control unit drives the proportional valve to change its opening degree with a sine wave, so that the moisture partial pressure in the cavity is modulated according to a preset sine wave pattern. The frequency of the sine wave is set to the reciprocal of the 24-hour cycle to simulate the breathing effect of the sealed cavity caused by the temperature difference between day and night in the actual service of a car sunroof. The amplitude of the sine wave is set from the water vapor partial pressure corresponding to 10% of the membrane design sealing level to the water vapor partial pressure corresponding to 90% to cover the extreme state of interlayer water vapor permeation-desorption cycle under the edge of seal failure.

[0027] The thin film sample to be tested is fixed in the middle of the cavity by edge clamping. Its main plane is perpendicular to the air outlet direction of the humidity generation module, ensuring that the change in moisture partial pressure can be uniformly applied to both sides of the thin film, and then diffuse and penetrate into the microcrystalline medium layer and ligand material layer inside the thin film.

[0028] S12. Application and frequency band coverage of the frequency logarithmic scanning alternating electric field, specifically: While the wet partial voltage sinusoidal modulation continues, an alternating electric field is applied to the transparent conductive electrode of the thin film sample under test through a function generator and a power amplifier. The voltage amplitude of this alternating electric field remains constant and equal to the effective value of the rated operating voltage of the thin film; the frequency is scanned logarithmically from the lower limit of the Hertz level to the upper limit of the kilohertz level, where the lower limit of the Hertz level covers the space charge polarization response characteristic frequency at the interface between the microcrystalline medium and the wide-temperature matrix material, and the upper limit of the kilohertz level covers the response characteristic frequency of the dipole orientation polarization of the molecular chain segments of the ligand material.

[0029] The scanning employs a logarithmic stepping method, with the dwell time at each frequency point being no less than fifty times the corresponding period to ensure that the dielectric response reaches a quasi-steady state. The trigger signal for frequency scanning is synchronized with the zero-crossing point of the wet voltage modulation waveform, ensuring that the acquisition of each electric field frequency-wet voltage coupling state is performed under the same wet voltage phase reference, facilitating subsequent phase difference calculation and data alignment.

[0030] S13. Synchronous triggering and data acquisition of the pump-probe photoelastic imaging system, specifically: Under each electric field frequency-wet partial voltage coupling state, a picosecond-time-resolved pump-probe photoelastic imaging system is synchronously triggered. The pump source of this system is a femtosecond pulsed laser output from a Ti:Sapphire laser, tuned by an optical parametric amplifier to the near-infrared wavelength corresponding to the dichroic absorption peak of the microcrystalline medium. The pulse width is compressed to below ten picoseconds, and the pulse energy is adjusted to below the threshold that only induces transient thermoelastic deformation without causing a thermally induced phase transition in the microcrystalline medium. The pump beam is shaped into a line source by a cylindrical lens and incident on the surface of the thin film sample under test at a Brewster angle, exciting transient thermal stress pulses in the thickness direction of the film.

[0031] The probe light source is a continuous-wave semiconductor laser collinear with the pump light, and the wavelength selection is in the visible light band, which is sensitive to changes in the orientation of the microcrystalline medium and is not absorbed. After the probe light passes through the thin film, its polarization state changes due to the birefringence effect of the thin film. The orthogonal polarization components are separated into a balanced photodetector by a Wollaston prism, and the differential signal reflects the transient change in the birefringence delay.

[0032] The measurement of ultrasonic surface wave velocity is achieved through an additional laser Doppler vibrometric optical path. This path scans the thin film surface in a direction perpendicular to the pump line light source, acquiring the time-series signal of the surface displacement. The surface wave velocity is calculated using the arrival time difference of the displacement signal. The data acquisition of the spatial distribution of birefringence retardation and the ultrasonic surface wave velocity field distribution is repeated under each electric field frequency-wet partial voltage coupling state until the entire scanning frequency band is covered. All data are stored according to an index structure of electric field frequency, wet partial voltage phase, and spatial coordinates, constructing a complete dielectric relaxation-photoelastic coupling field.

[0033] S14. Identification of Precursors to Interface Debonding and Encoding of Subcritical Defect Entropy State Matrix: Based on the constructed dielectric relaxation-photoelastic coupling field, the identification and encoding of precursors to interface debonding are performed, specifically including the following sub-steps: S141. Preliminary marking of non-co-evolutionary regions, specifically: For each spatial pixel in the dielectric relaxation-photoelastic coupling field, the time series sequences of birefringence delay and ultrasonic surface velocity under all electric field frequencies and wet partial pressure phases are extracted. Detrending processing is performed on both sequences: a linear least squares method is used to fit the linear trend of the sequence with respect to the index, and the fitted trend term is subtracted from the original sequence to obtain the residual sequence. Normalization is then performed: the arithmetic mean and standard deviation of the residual sequence are calculated, and each value is subtracted from the mean and divided by the standard deviation to obtain a standardized sequence with a mean of zero and a variance of one.

[0034] Calculate the Pearson correlation coefficient between two standardized sequences: calculate the arithmetic mean of the product of corresponding elements to obtain the correlation coefficient, which ranges from -1 to +1. In regions with homogeneous materials and intact interfaces, the photoelastic deformation induced by dielectric relaxation and the change in acoustic velocity induced by thermoelastic effects are physically correlated, exhibiting a high positive correlation coefficient. In regions where there are precursors to debonding at the microcrystal-ligand interface, the interruption of stress transmission at the interface leads to decoupling between the birefringence retardation and acoustic velocity, resulting in a significantly reduced correlation coefficient.

[0035] A preset threshold is set to distinguish between cooperative and non-cooperative evolution regions: A reference thin film sample without interface defects, verified by destructive cutting, is selected. Its dielectric relaxation-photoelastic coupling field is constructed according to steps S11 to S13. The Pearson correlation coefficient of all pixels is calculated, and its distribution, mean, and standard deviation are calculated. The threshold is set as the mean minus three times the standard deviation. All pixels of the sample under test are traversed, and pixels with correlation coefficients lower than the threshold are marked as candidate non-cooperative evolution regions, generating a binary mask image with the same spatial coordinate dimension.

[0036] S142. Spatial location of the effective defect area, specifically: Morphological image processing is performed on the binary mask image generated by S141 to optimize the spatial representation of defect regions. First, a dilation operation is performed: a circular structuring element with a radius five times the spatial resolution is convolved on the binary mask, expanding candidate non-co-evolutionary regions outwards, connecting broken regions caused by local noise, and filling small holes. Next, an erosion operation is performed: a circular structuring element of the same radius is deconvolved on the dilated image, shrinking the region boundaries to approximately the original contours and eliminating edge burrs introduced by the dilation process. The combination of dilation and erosion constitutes a morphological closing operation, merging adjacent candidate pixels to form connected components while maintaining the basic invariance of the overall shape of the region.

[0037] To eliminate isolated spurious defects caused by noise, a minimum identifiable defect size is set: depending on the thin film application scenario and reliability requirements, this is typically set to one-thousandth of the total thin film area, or equivalently, a circular region with a diameter not less than three percent of the short side length of the thin film. The area of ​​each connected region after closing operations is calculated, and connected regions with areas smaller than this size are eliminated; these eliminated connected regions are identified as isolated noise points. The retained connected regions are the effective interface debonding precursor regions. The coordinates of the circumscribed rectangle boundary or convex hull polygon vertices of each region are recorded, and the x and y coordinates of all pixels within the boundary are extracted to form a spatial coordinate set, stored in a list data structure. Each element corresponds to an effective interface debonding precursor region, and its content is an ordered set of all pixel coordinate pairs contained in that region.

[0038] S143. Extraction of the coupling phase difference, specifically: For each effective interface debonding precursor region determined in S142, its geometric center point or centroid coordinates are calculated as a geometric feature point. A local extraction window is set with this geometric feature point as the center, and the window size covers the outer rectangle of the region or extends to 120% of the maximum size of the region. From the four-dimensional data structure of the dielectric relaxation-photoelastic coupling field, the time-series curve of the birefringence delay at this geometric feature point as a function of the wet partial pressure phase index is extracted.

[0039] Using the birefringence retardation of the thin film in a dry state with the electric field frequency at the lower limit of the Hertzian level as the reference value, the absolute deviation of the birefringence retardation at each sampling point in the time series curve from the reference value is calculated. Sampling points where the birefringence retardation first deviates from the reference value by more than three standard deviations are identified. This three standard deviations are determined based on the measurement noise statistics of the birefringence retardation in the dry state. The electric field frequency value and the wet partial voltage phase value corresponding to this sampling point are recorded. The wet partial voltage phase value is calculated using the wet partial voltage phase index of this sampling point and the period parameter of the wet partial voltage modulation waveform.

[0040] Calculate the coupling phase difference: Extract the zero-crossing time closest to the timestamp of the sampling point from the timing data of the wet voltage modulation waveform. Calculate the time difference between the wet voltage phase value at that sampling point and the zero-crossing time. Divide this time difference by the period of the wet voltage modulation waveform and multiply by 360 degrees to obtain the coupling phase difference expressed in degrees. This coupling phase difference reflects the hysteresis characteristic of the interface debonding precursor to wet voltage modulation; a positive value indicates a hysteresis response, and a negative value indicates a leading response.

[0041] S144. Calculation of relaxation time distribution, specifically: For each effective interface debonding precursor region identified in S143, the birefringence retardation time series is extracted to analyze the dynamic characteristics of dielectric relaxation. The horizontal axis of this time series represents time, mapped to the actual time axis by both the electric field frequency index and the wet partial voltage phase index, while the vertical axis represents the birefringence retardation value. A single-exponential or double-exponential decay function is fitted to this time series to extract characteristic relaxation times.

[0042] The mathematical form of a single exponential decay function is: the birefringence retardation equals the steady-state value plus the amplitude coefficient multiplied by the exponential decay term, with the time variable having a negative exponent divided by the relaxation time. A double exponential decay function adds a second amplitude coefficient and a second relaxation time term to the single exponential function, used to characterize the superposition of fast and slow relaxation processes.

[0043] A nonlinear least squares method was used for fitting optimization: initial values ​​for the fitting parameters were set, the steady-state value was the arithmetic mean of the values ​​at the end of the time series, the amplitude coefficient was the difference between the peak value and the steady-state value of the time series, and the relaxation time was one-tenth of the wet partial voltage modulation period. The objective function was constructed as the sum of squared residuals between the fitted values ​​and the actual observed values. The Levenberg-Marquardt algorithm was used for iterative optimization until the relative change in the sum of squared residuals was less than a preset convergence threshold or the maximum number of iterations was reached. The relaxation time was extracted from the optimized fitting parameters and defined as the time required for the birefringence delay to decay from the peak amplitude to 36.8%.

[0044] Repeat the above fitting process for all pixels within the effective interface debonding precursor region to obtain the relaxation time value for each pixel. Statistically analyze the distribution characteristics of the relaxation times of all pixels to generate a histogram data structure: set the number of bins to the square root of the total number of pixels, or adaptively determine it using the Sturgess formula; calculate the minimum and maximum values ​​of all relaxation times, uniformly divide the interval into bins of sub-intervals, and count the number of pixels falling within each sub-interval as the frequency. Record the bin boundary array, frequency array, and statistical features, including arithmetic mean, standard deviation, skewness, and kurtosis.

[0045] S145. Encoding and storage of three-dimensional data structures, specifically: The spatial coordinate set, coupling phase difference, and relaxation time distribution histogram obtained in the above steps are encoded into a three-dimensional data structure. The first dimension is the defect region index, with each index corresponding to a valid interface debonding precursor region. The second dimension is the attribute category, including the spatial coordinate substructure, the coupling phase difference scalar value, and the relaxation time distribution substructure. The third dimension varies according to the substructure type: the spatial coordinate substructure contains a list of pixel coordinates, the coupling phase difference is a single floating-point value, and the relaxation time distribution substructure contains a histogram bins array and a frequency array. This three-dimensional data structure is stored in binary format or a standardized data exchange format and named the subcritical defect entropy state matrix, serving as the input data for the subsequent S2 step.

[0046] S2. Using the spatial coordinates in the entropy state matrix as the focusing point, photoacoustic, thermoacoustic, and ultrasonic guided wave signals are simultaneously excited to form a phase-locked coherent superposition detection mechanism. Amplitude and strain energy spectra are extracted via phase-locked demodulation, and a three-band energy dissipation topology network is reconstructed. Topological defect nodes are identified and registered with the entropy state matrix to generate a cross-physics field coupled failure prediction map. Specifically, this includes: S21. The physical conditions for phase-locked coherent superposition detection and the signal excitation control, specifically: The spatial coordinates of the effective interface debonding precursor region are extracted from the subcritical defect entropy state matrix encoded by S14, and its geometric center coordinates are calculated as the multiphysics field focusing point. The thin film sample to be tested is positioned in the detection area using a precision displacement stage, ensuring that the point is precisely located at the triple intersection of the optical axis center of the Q-switched pulse laser, the geometric center of the electric field modulation electrode, and the focal zone of the external piezoelectric transducer acoustic beam, with positioning accuracy controlled at the micrometer level.

[0047] A phase-locked coherent superposition detection field is constructed at this site. The dielectric thermoacoustic signal induced by electric field frequency modulation is used as the reference phase source, and its phase zero point is locked by a phase-locked loop circuit. The narrowband photoacoustic signal induced by Q-switched pulse laser is controlled by an adjustable optical delay line to ensure that when the photoacoustic pressure wave reaches the focusing site, its wavefront phase maintains a preset phase difference with the dielectric thermoacoustic signal. This phase difference is usually set to 0° or 90° to achieve in-phase or orthogonal superposition. The broadband ultrasonic guided wave signal injected by the external piezoelectric transducer is adjusted by an electronically controlled delay trigger to ensure that its phase at the focusing site maintains a preset coherent relationship with the above two signals.

[0048] Although the three signals have different physical mechanisms and frequency bands, by precisely controlling the phase relationship of each signal, they can achieve controllable coherent superposition at the focal point. This superposition forms a demodulated composite response signal in the time domain, rather than a spatially stable standing wave distribution. Specifically, to achieve phase locking of the three different frequency signals, a sideband modulation coherent technique is used, with the modulation frequency f of the dielectric thermoacoustic signal... m As a reference, the output of the Q-switched pulsed laser is amplitude modulated using an electro-optic modulator to generate a frequency of f. m The low-frequency sideband envelope allows the narrowband photoacoustic signal to carry phase information at the same frequency as the dielectric thermoacoustic signal, with a modulation frequency f. m The frequency is set to be one-tenth lower than the repetition frequency of the Q-switched pulse laser to ensure that at least ten cycles of amplitude modulation are completed within the pulse interval, forming a stable low-frequency sideband envelope. Simultaneously, amplitude modulation is applied to the external piezoelectric transducer via an electronically controlled power amplifier, at the carrier frequency f of the broadband ultrasonic guided wave. u The frequency generated on both sides is f m The sideband components are extracted. Sideband components with the same frequency as dielectric thermoacoustic signals are extracted from the photoacoustic and ultrasonic guided wave signals using bandpass filters, ensuring that all three signals have a frequency of f at the focal point. m The phase-locked components form an equivalent detection field that can be coherently superimposed. Synchronous demodulation is performed using a lock-in amplifier with the dielectric thermoacoustic signal as a reference. The contribution components of each individual signal and their coupling terms are separated from the composite response signal, and the amplitude and phase information reflecting the microstructure state of the material are extracted.

[0049] Maintaining the phase stability of the superimposed signal is achieved through real-time feedback control. The envelope stability of the composite response signal is monitored. When the envelope fluctuation exceeds the preset threshold, the delay of the optical delay line and the electronically controlled delay trigger is automatically adjusted to compensate for the phase drift caused by environmental disturbances, ensuring the continuous stability of the phase-locked state and guaranteeing the reliability of the detection signal.

[0050] S22. Extraction based on coherent superposition of amplitude spectrum and strain energy density spectrum, specifically: S221. Using the dielectric thermoacoustic signal excited by S21 as the phase reference for phase-locked loop (LLL) detection, a miniature piezoelectric sensor is arranged on the thin film surface or electrode edge. This sensor is directly attached to the thin film or transmits acoustic energy through an acoustic coupling agent, receiving the pressure wave output of the dielectric thermoacoustic signal. The sensor output is amplified by a low-noise preamplifier and then input into the reference channel of the digital LLL amplifier. This dielectric thermoacoustic signal has a phase relationship that is strictly locked to the electric field modulation frequency, which can effectively track the physical process of electric field-thermal-acoustic coupling and provide a stable phase reference for subsequent demodulation.

[0051] S222. Narrowband photoacoustic signals are received via a broadband photoacoustic detector or a laser interferometric displacement sensor arranged on the thin film surface. This detector is sensitive to nanoscale surface displacements caused by photoacoustic pressure waves. Broadband ultrasonic guided wave signals are received via a piezoelectric receiving array or a laser Doppler vibrometer arranged on the other side of the thin film, providing multi-channel parallel acquisition capability. After pre-amplification and bandpass filtering, the two signals, along with the dielectric thermoacoustic reference signal, are input into the signal channel of a digital lock-in amplifier, forming a three-channel synchronous acquisition architecture that ensures time alignment accuracy better than the nanosecond level.

[0052] Due to the significant frequency differences among the three types of signals, a frequency-band independent phase-locked-loop demodulation architecture is adopted: the dielectric thermoacoustic signal is directly input into the first channel of the digital lock-in amplifier, and baseband demodulation is performed using itself as a reference; the narrowband photoacoustic signal, after being filtered by a bandpass filter to remove the high-frequency carrier, has its low-frequency modulation sideband extracted and input into the second independent channel, and demodulation is performed using the same-frequency locked signal generated by the dielectric thermoacoustic reference signal through a frequency synthesizer as a reference; the broadband ultrasonic guided wave signal has its sub-bands shifted to the baseband through a digital downconverter, divided into several frequency bands, and then input into a multi-channel lock-in amplifier. Each channel is demodulated in parallel using the corresponding frequency reference signal generated by the frequency synthesizer from the dielectric thermoacoustic reference signal, ensuring that each frequency band signal completes phase-sensitive detection in an independent channel.

[0053] S223. After demodulation by independent phase-locked loop (PLL) in frequency bands as described in S222, the signals of each channel undergo vector synthesis and spectral separation processing in a digital LPL amplifier. Using the dielectric thermoacoustic signal of the reference channel as a reference, in-phase and quadrature reference waveforms are generated. The frequencies of the two reference waveforms are consistent with the fundamental frequency of the dielectric thermoacoustic signal, and the phase difference is strictly maintained at 90 degrees. The composite response signal received by the signal channel is multiplied by the in-phase and quadrature reference waveforms respectively. After low-pass filtering, the in-phase and quadrature components are obtained. Vector synthesis is then used to obtain the amplitude and phase of the coherent portion of the composite signal with the reference signal. A Fast Fourier Transform is further performed on the composite signal, and a digital bandpass filter separates the spectral contributions of narrowband photoacoustics, dielectric thermoacoustics, and broadband ultrasonic guided waves.

[0054] S224. Extract the frequency domain characteristic spectrum reflecting the mechanical state of the material's micro-regions from the phase-locked demodulation results. Define the displacement amplitude spectrum as the distribution curve of particle displacement amplitude as a function of frequency, obtained through coherent detection within the ultrasonic guided wave band. Identify local minimum points of amplitude as equivalent node locations and local maximum points as equivalent anti-node locations. The strain energy density spectrum is calculated by multiplying the square of the particle velocity amplitude at the equivalent anti-node by the product of the material density and the local elastic modulus, characterizing the degree of energy accumulation in the micro-region. After calibration and verification, the above spectral data are used as input data for the reconstruction of the three-band energy dissipation topology network.

[0055] S23. Reconstruction of the three-band energy dissipation topology: Based on the displacement amplitude spectrum and strain energy density spectrum extracted in S22, the three-band energy dissipation topology of the thin film in three characteristic frequency bands is reconstructed; the specific implementation includes the following sub-steps: S231. Three-band feature extraction and dispersion-attenuation analysis, specifically: From the displacement amplitude spectrum extracted from S22, equivalent dispersion curves were separated within three characteristic frequency bands: the orientation polarization frequency band of the microcrystalline medium, the viscoelastic relaxation frequency band of the ligand material, and the resonance frequency band of the whole-film structure. The equivalent dispersion curves were extracted as follows: peak picking was performed on the displacement amplitude spectrum within each characteristic frequency band, the frequency-wavenumber coordinates corresponding to the amplitude maxima were identified, and the functional relationship between frequency and wavenumber was fitted through phase velocity calculation and group velocity derivation to obtain the equivalent phase velocity dispersion curve and the equivalent group velocity dispersion curve within that frequency band. The phase velocity was calculated by dividing the angular frequency by the wavenumber, and the group velocity was calculated by the derivative of the angular frequency with respect to the wavenumber.

[0056] From the strain energy density spectrum extracted from S22, the equivalent attenuation coefficient spectra corresponding to the three characteristic frequency bands are separated. The equivalent attenuation coefficient is calculated by fitting the exponential attenuation characteristics of the strain energy density spectrum along the frequency axis, or by calculating based on the amplitude attenuation slope of the time-domain signal, to obtain the equivalent attenuation coefficient characterizing the energy dissipation rate during propagation, thus forming a distribution spectrum of the attenuation coefficient as a function of frequency.

[0057] S232. Construction of the initial topology network, specifically: The basic node set for constructing a topological network is based on the microstructural units of the thin film sample. The microstructural units are divided as follows: the thin film is discretized into grid cells along the in-plane direction, with the grid size determined according to the detection spatial resolution, typically 10-100 micrometers; along the thickness direction, it is divided into a microcrystalline medium layer, a ligand material layer, and an interface transition layer based on the material's layered structure. The grid cells within each layer constitute an independent subset of nodes. Node attributes include: spatial coordinates, the identifier of the material layer to which the node belongs, and the vibrational mode participation coefficient calculated based on the equivalent dispersion curve extracted from S231. This coefficient reflects the energy proportion of the node in the vibration of a specific frequency band.

[0058] The vibration mode participation factor is calculated as follows: Modal analysis is performed on the equivalent dispersion curve extracted by S231 to identify the vibration mode corresponding to each dispersion branch, including the orientation polarization mode of the microcrystalline medium, the shear mode of the ligand material and the bending mode of the whole film; the ratio of the displacement amplitude of each node to the total energy in each mode is calculated as the participation factor of the node in that mode.

[0059] The connection relationships of the topological network are constructed using the energy transfer efficiency between adjacent units as edges. An edge is defined as follows: within the same material layer, it connects four adjacent mesh units in the same plane; between adjacent material layers, it connects mesh units that overlap in the thickness direction. Edge attributes include: equivalent energy transfer efficiency and energy dissipation weight. The equivalent energy transfer efficiency is calculated based on the equivalent dispersion curve extracted from S231 and the material calibration database, and is the product of the group velocity between adjacent nodes and the equivalent transmission coefficient. The equivalent transmission coefficient is the effective efficiency coefficient of energy transfer between adjacent microstructure units, obtained by pre-calibrating standard calibration samples and comprehensively considering the intrinsic material loss and interface scattering effects. The energy dissipation weight is calculated based on the equivalent attenuation coefficient spectrum extracted from S231, and is the product of the arithmetic mean of the equivalent attenuation coefficients between adjacent nodes and the distance between them.

[0060] By integrating the node set and edge set, an initial three-band energy dissipation topology network is constructed. This network is a three-layer composite structure, with each layer corresponding to a characteristic frequency band. The nodes and edges within a layer have the same definition, and the attribute values ​​within a layer are calculated independently based on the equivalent dispersion curve and equivalent attenuation coefficient spectrum of that frequency band.

[0061] S233. Spectral clustering optimization of topological networks, specifically: Spectral clustering analysis was performed on the initial three-band energy dissipation topology network constructed by S232 to identify the main and side paths of energy dissipation. The specific implementation of spectral clustering is as follows: The Laplacian matrix of the network is constructed, where the diagonal elements represent the degree of the nodes and the off-diagonal elements represent the negative edge weights; the eigenvalues ​​and eigenvectors of the Laplacian matrix are calculated, and the eigenvectors corresponding to the k smallest non-zero eigenvalues ​​are selected as the low-dimensional embedding representations of the nodes; k-means clustering is performed on the embedding representations to divide the network into k energy dissipation communities.

[0062] The main and secondary energy dissipation paths within each frequency band are identified. The main path is identified as follows: within each community, the shortest path between node pairs is calculated, and the shortest path connecting energy input nodes and energy output nodes in the network is selected as the candidate main path. The candidate main paths are ranked by importance based on the product or weighted sum of the weights of all edges on the path, and the path with the highest ranking is selected as the main path. The secondary paths are other valid paths within the community besides the main paths.

[0063] Redundant edges with weights below a preset threshold are removed to optimize the network structure. The preset threshold is set by calculating the statistical distribution of all edge weights and setting the threshold as the mean minus twice the standard deviation. All edges are traversed, and edges with weights below the preset threshold are marked as redundant and removed from the network, resulting in an optimized three-band energy dissipation topology network. This optimized network retains the main energy dissipation channels while reducing computational complexity and noise interference.

[0064] S24. Identification of topological defect nodes and generation of cross-physics field coupling failure prediction maps: Based on the three-band energy dissipation topology network optimized in S233, topological defect nodes are identified and cross-physics field coupling failure prediction maps are generated; the specific implementation includes the following sub-steps: S241. Calculation and marking of energy flow convergence degree of topological defect nodes, specifically: Traverse all nodes in the three-band energy dissipation topology network and calculate the energy flow convergence degree of each node. The energy flow convergence degree is defined as the ratio of the sum of the weights of all edges flowing into the node to the sum of the weights of all edges flowing out of the node; if the ratio is greater than one, the node is an energy convergence node where energy accumulates; if it is less than one, the node is an energy dissipation node where energy dissipates; if it is equal to one, the node is in energy balance.

[0065] A preset critical dissipation threshold is set to distinguish between normal nodes and defective nodes. This threshold is set based on the energy flow convergence distribution of the same batch of non-destructive reference samples, calculating the mean and standard deviation, and setting the threshold as the mean plus three times the standard deviation, or a critical value determined according to the failure physics model. Nodes with energy flow convergence exceeding the preset critical dissipation threshold are marked as candidate topological defective nodes. These nodes represent the locations of abnormal energy convergence, corresponding to potential stress concentrations or interface debonding propagation regions.

[0066] S242. Spatial registration and confirmation of valid topological defect nodes, specifically: The spatial coordinates in the S1-encoded subcritical defect entropy state matrix are mapped to the node coordinate system of the three-band energy dissipation topology network. The specific implementation of this coordinate mapping is as follows: extract the spatial coordinate set of each effective interface debonding precursor region from the subcritical defect entropy state matrix and calculate its geometric center coordinates; in the node set of the topology network, search for the node with the smallest Euclidean distance to this geometric center coordinate, and establish the correspondence between the subcritical defect and the topology node.

[0067] Calculate the Euclidean distance between each candidate topological defect node and the spatial coordinates of the entropy state matrix of its nearest subcritical defect. Set a preset spatial correlation threshold, which is determined based on the required spatial resolution and defect localization accuracy, typically two to five times the spatial resolution. Candidate topological defect nodes with a distance less than the preset spatial correlation threshold are identified as valid topological defect nodes. These nodes simultaneously satisfy the conditions of abnormal energy flow convergence and spatial proximity to subcritical defects, characterizing high-failure-risk areas.

[0068] S243. Tracking and marking failure propagation paths, specifically: For each valid topological defect node, trace its energy input and dissipation paths in the three-band energy dissipation topology network. The energy input path is traced as follows: starting from the node, search backwards for the incident edge with the largest edge weight, recursively tracing back to the energy input source node, and recording all nodes and edges along the path. The energy dissipation path is traced as follows: starting from the node, search forwards for the outgoing edge with the largest edge weight, recursively tracing back to the energy dissipation sink node, and recording all nodes and edges along the path. Merge the energy input and dissipation paths, removing duplicate nodes and edges to form a complete potential failure propagation channel.

[0069] Mark all nodes on the channel as high-risk nodes and all edges as high-risk edges, and record the quantified risk level values ​​of each node and edge.

[0070] S244. Generation of cross-physics coupling failure prediction maps, specifically: The effective topological defect nodes, potential failure propagation channels, and their attribute information obtained from S241 to S243 are superimposed onto the spatial structure map of the thin film sample to generate a cross-physics field coupled failure prediction map. Specifically, a three-dimensional visualization space is established using the in-plane spatial coordinates of the thin film as the horizontal and vertical axes, and the material layer thickness direction as the depth axis. In this space, the energy flow convergence degree values ​​of effective topological defect nodes are encoded with different colors or transparency, and the edge weight values ​​of potential failure propagation channels are encoded with different line types or thicknesses. The boundaries of the subcritical defect regions identified in S1 are superimposed and displayed, and the confidence levels of different physics field detection results are distinguished by dashed lines or semi-transparent filling.

[0071] The data structure of this cross-physics coupled failure prediction map includes: spatial grid definition, node attribute list, edge attribute list, failure risk level labeling, and cross-reference index with the S1 subcritical defect entropy state matrix. The map is stored in both a visual image format and a structured data format, serving as input data for the S3 step, used for service limit extrapolation and adaptive compensation strategy generation.

[0072] S3. Integrating coupled fields and topological networks, a service limit extrapolation model based on the self-organized criticality of dissipative structures is established. An adaptive compensation strategy is generated based on the entropy state matrix and the distribution of topological defect nodes. The service limit, compensation strategy, and failure map are encapsulated into a quality adaptive digital twin; specifically including: S31. The establishment and dual-metric fusion of the self-organization criticality criterion for dissipative structures, specifically: By integrating the dielectric relaxation-photoelastic coupling field constructed in S1 with the three-band energy dissipation topology network reconstructed in S2, a service limit extrapolation model is established based on the self-organization criticality of the dissipation structure. This model is based on the dissipation structure theory in non-equilibrium thermodynamics, treating the aging process of the thin film under cyclic conditions as a non-equilibrium evolution process of an open system exchanging matter and energy with the external environment.

[0073] S311. Construction and calculation of the non-equilibrium energy dissipation intensity index, the specific implementation method is as follows: from the specific spatial coordinates identified by the subcritical defect entropy state matrix encoded in S14, extract the birefringence delay time sequence and the ultrasonic surface wave velocity time sequence of the corresponding pixel point in the dielectric relaxation-photoelastic coupling field constructed in S13.

[0074] The electric field-driven deviation is defined as the ratio of the rate of change of birefringence retardation to the reference sample baseline rate of change. This ratio characterizes the degree to which the polarization response of the material deviates from the equilibrium state under the action of the electric field. The mechanical dissipation deviation is defined as the ratio of the rate of change of ultrasonic surface wave velocity to the reference sample baseline wave velocity rate of change. This ratio characterizes the degree to which the elastic response of the material deviates from the equilibrium state under the action of mechanical stress.

[0075] The non-equilibrium energy dissipation intensity index is defined as the weighted geometric mean of the electric field-driven deviation and the mechanical dissipation deviation. The weighting coefficient is determined based on the energy proportions of the electric field and mechanical stress in the coupled field, and the weighting coefficient w is... e = E e / (E e + E m ), where E e E is the electric field energy density. m Elastic strain energy density; weighting coefficient w for mechanical dissipation deviation. m =E m / (E e + E m ), satisfying w e + w m = 1. This dimensionless non-equilibrium energy dissipation intensity index characterizes the relative degree to which the system deviates from the thermodynamic equilibrium state. When the index value exceeds a preset threshold, the system is determined to have entered a non-equilibrium dissipation state. The preset threshold ranges from 1.5 to 3.0, with a classical value of 2.0. Statistical analysis is performed on this index within each wet partial pressure modulation cycle, and its standard deviation is calculated as the fluctuation amplitude. This fluctuation amplitude serves as a quantitative measure of the degree to which the system deviates from the equilibrium state during that cycle.

[0076] S312. The connectivity evolution of the three-band energy dissipation topology network is used as a measure of the change in the orderliness of the dissipation structure. Specifically, the following implementation method is used: Extract the network adjacency matrix for each detection time point from the optimized three-band energy dissipation topology network (S233); calculate the shortest path length matrix between all valid nodes in the network; and take the reciprocal of the arithmetic mean of the non-infinite elements of this matrix as the global connectivity; or use the reciprocal of the characteristic path length, network efficiency, or algebraic connectivity as alternative indicators. Track the evolution trend of this connectivity with detection time or number of cycles as a quantitative indicator of the change in the orderliness of the dissipation structure.

[0077] The above two metrics are integrated into the core input of the service limit extrapolation model: a two-dimensional state space is constructed with the fluctuation amplitude of the non-equilibrium energy dissipation intensity index as the horizontal axis and connectivity as the vertical axis; in this state space, the evolution trajectory of the thin film moves from the high connectivity-low fluctuation region near the equilibrium region to the low connectivity-high fluctuation region far from the equilibrium region, and the curvature change or slope change of the trajectory indicates the phase transition or bifurcation of the dissipation structure.

[0078] S32. Tracking the critical point of dissipative structure transition and extrapolating the service limit cycle number: Based on the service limit extrapolation model established in S31, the service limit cycle number of the thin film under the cyclic breathing-illumination-electric drive coupling condition is extrapolated by tracking the critical point of the dissipative structure transition from the near-equilibrium linear region to the dissipative structure bifurcation region; the specific implementation includes the following sub-steps: S321. Time-series analysis of the fluctuation amplitude of the non-equilibrium energy dissipation intensity index, specifically: From the service limit extrapolation model established by S31, time-series data of the non-equilibrium energy dissipation intensity index within the wet partial voltage modulation cycle are extracted. This time-series data is a discrete sequence, with each data point corresponding to the calculated value of the non-equilibrium energy dissipation intensity index fluctuation amplitude for one wet partial voltage modulation cycle. A sliding window variance analysis is performed on this time-series data: the window width is set to ten to fifty wet partial voltage modulation cycles, and the window slides with a step size of one cycle. The variance or standard deviation of the non-equilibrium energy dissipation intensity index data within each window is calculated as the characterization value of the non-equilibrium energy dissipation intensity index fluctuation amplitude at the center time of that window. A time-series sequence of the non-equilibrium energy dissipation intensity index fluctuation amplitude is generated. The sampling interval of this sequence is equal to the sliding step size, and the sequence length is equal to the length of the original non-equilibrium energy dissipation intensity index time-series data minus the window width plus one.

[0079] S322. Temporal extraction of connectivity evolution, specifically: Simultaneously, connectivity evolution data of the three-band energy dissipation topology network is extracted from the service limit extrapolation model. This data extraction method is as follows: if the three-band energy dissipation topology network reconstruction of S2 is repeatedly performed at multiple detection time points, the connectivity calculation values ​​at each time point are directly extracted to form a time-series sequence; if only a single reconstruction is performed, based on the network structure of the single reconstruction, combined with the time-series evolution data of the dielectric relaxation-optical elastic coupling field of S1, the extrapolated values ​​of network parameters over time are calculated using linear response theory or perturbation approximation, thereby estimating the time-series evolution of connectivity. Connectivity is defined using the global connectivity described in S31, which is the reciprocal of the average shortest path length between effective nodes in the network. A larger value indicates better network connectivity and more uniform energy dissipation; a smaller value indicates that the network tends to be fragmented and energy dissipation is localized.

[0080] S323. Cross-correlation identification of transition critical points, specifically: A co-analysis is performed on the time series sequence of the fluctuation amplitude of the non-equilibrium energy dissipation intensity index generated by S321 and the connectivity time series sequence generated by S322 to identify the critical point of the transition of the dissipation structure from the near-equilibrium linear region to the bifurcation region. Specifically, the cross-correlation function of the two time series is calculated on the time axis; that is, Pearson correlation analysis is performed on the two series at different time delays to obtain the curve of the cross-correlation function value changing with time delay. The peak position of this cross-correlation function indicates the moment when the correlation between the two series is strongest, and the peak amplitude indicates the correlation strength.

[0081] A preset critical correlation threshold is set, which is determined based on the statistical distribution of the cross-correlation function of the non-destructive reference sample, typically taken as the mean plus two standard deviations. The time point at which the cross-correlation function first shows a positive peak near zero time delay and exceeds the preset critical correlation threshold is tracked. This time point characterizes the phase transition moment when the exponential fluctuation of non-equilibrium energy dissipation intensity and the evolution of connectivity change from weak coupling to strong cooperation, and is marked as the critical point for the dissipative structure to transition from the near-equilibrium linear region to the bifurcation region. Before this critical point, the system response satisfies the principle of linear superposition, and the dissipative structure is stable; after this critical point, the system enters the nonlinear bifurcation region, where small perturbations may trigger abrupt changes or instability in the dissipative structure.

[0082] S324. Extrapolation calculation of the service limit cycle count, specifically: Based on the number of cycles corresponding to the critical point identified by S323, and combined with the conversion relationship between the preset accelerated aging factor and the measured operating conditions, the service limit cycle count of the thin film under the coupled conditions of cyclic breathing-lighting-electric drive is extrapolated. Specifically, the following implementation method is used: Record the number of wet partial pressure modulation cycles or the number of detection cycles corresponding to the critical point, denoted as the critical cycle count; determine the accelerated aging factor, which is the stress intensity ratio between the detection condition and the actual service condition, determined through a comprehensive evaluation of the temperature difference, wet partial pressure difference, electric field strength difference, and cycle frequency difference between the detection condition and the actual condition, typically calculated using an inverse power law model or the Alling model; multiply the critical cycle count by the power or exponential function of the accelerated aging factor to obtain the estimated service limit cycle count. This estimated value characterizes the cycle life of the thin film from commissioning to dissipative structural instability under actual service conditions, serving as a quantitative basis for reliability design and quality assurance strategies.

[0083] S33. Generation of Personalized Adaptive Compensation Strategy: Based on the distribution density of interface debonding precursors in the subcritical defect entropy state matrix encoded by S1 and the spatial clustering degree of topological defect nodes identified by S2, a personalized adaptive compensation strategy is generated for the thin film under test. This strategy avoids or suppresses the identified defect risks and extends the effective service life of the thin film by adjusting the driving conditions and structural design parameters. The specific implementation includes the following sub-steps: S331. Determination of priority compensation areas, specifically: Read the S14-encoded subcritical defect entropy state matrix and extract the spatial coordinate set and distribution density data of interface debonding precursors. The distribution density is calculated by counting the number of effective interface debonding precursor regions per unit area, or by using kernel density estimation to calculate the spatial continuous density distribution. Spatial regions with distribution densities exceeding a preset threshold are identified as high-risk areas. This preset threshold is set based on the statistical distribution of samples in the same batch, typically taking the upper quartile.

[0084] Simultaneously, the spatial coordinates and clustering data of the valid topological defect nodes confirmed by S242 are read. The clustering degree is calculated by counting the number of topological defect nodes per unit area, or by calculating the nearest neighbor distance statistics of the spatial distribution of nodes. Spatial regions with clustering degrees exceeding a preset threshold are identified as energy dissipation anomaly regions. The spatial intersection of the high-risk region and the energy dissipation anomaly region, i.e., the region that simultaneously satisfies the high interface debonding risk and high energy dissipation anomaly, is taken as the priority compensation region; if the two regions do not intersect, the union is taken or they are processed separately.

[0085] S332. Generation of driving electric field frequency-phase premodulation parameters, specifically: For the priority compensation region determined in S331, extract the electric field frequency-wet voltage coupling phase difference data recorded in S143 for this region. Calculate the statistical distribution of this data: if the priority compensation region contains multiple subcritical defect entropy state matrix entries, calculate the arithmetic mean of the coupling phase differences of each entry as the center value, and calculate the standard deviation or interquartile range as the dispersion; if it contains only a single entry, use the coupling phase difference of that entry as the center value, and estimate the dispersion by combining the statistical results of neighboring regions.

[0086] Based on the center value and dispersion, frequency-phase pre-modulation parameters for the driving electric field are generated: the operating frequency of the driving electric field is adjusted from the standard value to a range deviating from the peak frequency of dielectric loss, with the adjustment amount determined according to the frequency offset corresponding to the center value; the applied phase of the driving electric field is adjusted from being synchronized with the wet partial voltage modulation to an asynchronous state with a preset phase difference, which is determined based on the negative value of the center value, in order to actively avoid the frequency-phase coupling range corresponding to the peak dielectric loss. These pre-modulation parameters reduce the coupling strength between the electric field drive and the wet partial voltage breathing effect during actual operation of the thin film, slowing down the rate of interface debonding.

[0087] S333. Generation of design parameters for the thickness gradient of the local humidity barrier layer, specifically: Based on the spatial distribution characteristics of the priority compensation region at the film edge determined in S331, and combined with the energy flow convergence direction of this region in the three-band energy dissipation topology network reconstructed in S2, the thickness gradient design parameters of the local humidity barrier layer are generated. Specifically, the following implementation method is used: The portion of the priority compensation region located at the film edge or at a distance less than a preset distance threshold from the edge is identified and marked as an edge debonding risk zone; the energy flow convergence direction of these regions is extracted from the three-band energy dissipation topology network, i.e., the vector synthesis direction of the energy input path and the dissipation path, which indicates the priority propagation direction of debonding expansion.

[0088] The thickness distribution of the localized humidity barrier layer is designed as follows: at the beginning of the edge debonding risk zone, the barrier layer thickness maintains the standard design value; along the energy flow convergence direction, i.e., the debonding propagation path, the barrier layer thickness gradually increases with a linear or exponential gradient. The gradient slope is determined based on the difference between the service limit cycle number extrapolated from S324 and the target lifetime; the larger the difference, the steeper the slope. This thickness gradient design causes the water vapor permeation resistance to the interior of the membrane to increase along the propagation path, inhibiting further propagation of edge debonding.

[0089] S334. Generation of pre-tilt angle distribution parameters of orientation axis in microcrystalline media, specifically: Based on the current distribution state of the orientation axis of the microcrystalline medium in the priority compensation region determined in S331, and combined with the anisotropic evolution characteristics of the birefringence retardation in S13, the pre-tilt angle distribution parameters of the orientation axis of the microcrystalline medium are generated. Specifically, the spatial distribution of the birefringence retardation in the priority compensation region is extracted from the dielectric relaxation-photoelastic coupling field. The current distribution state of the orientation axis of the microcrystalline medium, i.e., the tilt angle and in-plane azimuth angle of the orientation axis relative to the film normal, is estimated by solving the inverse problem or using a lookup table method. The anisotropic evolution characteristics of the birefringence retardation are analyzed to identify the principal direction of moisture expansion, which is the direction most sensitive to changes in moisture partial pressure.

[0090] Calculate the compensating pre-tilt angle: This causes the orientation axis to undergo an additional deflection relative to its current state in the principal direction of moisture expansion. The deflection angle is determined based on the moisture expansion anisotropy coefficient and the target compensation amount, typically ranging from a few degrees to a dozen degrees. In the direction perpendicular to the principal direction of moisture expansion, the basic state of the orientation axis is maintained, or a reverse deflection is formed to balance the internal stress. This pre-tilt angle distribution parameter is achieved through an electric field pre-orientation process or a mechanical stretching process during thin film preparation, enabling the thin film to possess an inherent structure that counteracts moisture expansion anisotropy before service.

[0091] S34. Encapsulation and embedding of quality-adaptive digital twins, specifically: The service limit cycle number extrapolated by S324, the adaptive compensation strategy parameter set generated by S332 to S334, and the sparse feature vector of the cross-physics coupling failure prediction map generated by S244 are encapsulated into a quality adaptive digital twin that can be embedded in thin-film flexible integrated circuits. Specifically, the service limit cycle number is encoded using floating-point encoding, retaining the number of significant digits; the adaptive compensation strategy parameter set is structured, including frequency premodulation offset, phase premodulation difference, barrier layer thickness gradient function coefficients, and pretilt angle distribution matrix; sparse encoding is performed on the cross-physics coupling failure prediction map, and principal component analysis or independent component analysis is used to extract the feature vector of the map, retaining the top principal components with a cumulative contribution rate exceeding 95%, reducing the data dimensionality to a storable range.

[0092] The aforementioned encoded data is packaged according to a preset protocol to generate a digital twin data packet. This data packet is written into the non-volatile memory of a thin-film flexible integrated circuit via a near-field communication interface or the physical contact point of a flexible circuit. This integrated circuit is integrated into the edge of the thin film or the electrode lead-out area and has temperature sensing, humidity sensing, and electric field monitoring functions. During the service life of the thin film, it collects operating condition data in real time and compares it with the dissipative structure evolution model stored in the digital twin. When the actual evolution trajectory deviates from the predicted path by more than a preset tolerance, it triggers the dynamic update of the adaptive compensation strategy or the output of an early warning signal. The embedding of the quality adaptive digital twin realizes the value-added extension of inspection data to the service stage, forming an intelligent quality closed loop throughout the entire life cycle of inspection, manufacturing, and service.

[0093] Example 2 like Figure 2 As shown, a detection system for a sunroof display film is used to implement a detection method for a sunroof display film, comprising: The multi-field coupling excitation and subcritical entropy state encoding module is used to apply a frequency-converting electric field to the thin film under controllable wet partial voltage modulation, and simultaneously acquire picosecond photoelastic imaging data to construct a dielectric relaxation-photoelastic coupling field, identify interface debonding precursors and encode them as a subcritical defect entropy state matrix. The multi-physics phase-locked superposition detection and energy dissipation topology reconstruction module is connected to the multi-field coupled excitation and subcritical entropy state encoding module. It is used to excite multi-physics phase-locked coherent superposition signals with the spatial coordinates in the subcritical defect entropy state matrix as the focusing point, extract the amplitude spectrum and strain energy spectrum through phase-locked demodulation, reconstruct the three-band energy dissipation topology network, identify topology defect nodes and register them with the entropy state matrix to generate cross-physics coupling failure prediction map. The dissipative structure modeling and adaptive digital twin module, connected to the multi-field coupled excitation and subcritical entropy state encoding module and the multi-physics phase-locked superposition detection and energy dissipation topology reconstruction module, is used to fuse coupled fields and topology networks to establish a service limit extrapolation model based on the self-organized criticality of dissipative structures, generate personalized adaptive compensation strategies, and encapsulate the service limit, compensation strategy and failure prediction map into a quality adaptive digital twin.

[0094] As can be seen from the above description, the embodiments of the present invention achieve the following technical effects: Existing detection technologies are limited to macroscopic performance testing or physical defect detection of thin films after stress, falling into the category of post-event verification and unable to capture the subcritical state before failure. This invention achieves dynamic capture of pre-debonding precursors at the microcrystal-ligand interface by constructing a dielectric relaxation-photoelastic coupling field: under controllable wet partial voltage modulation and frequency-coupled electric field excitation, the non-cooperative evolution region of birefringence retardation and ultrasonic surface velocity is tracked in real time, encoding the spatial coordinates, coupling phase difference, and relaxation time distribution of the pre-debonding precursors into a subcritical defect entropy state matrix. This allows the detection system to observe the initiation of microscopic interface damage hours or even days before macroscopic failure, providing an unprecedented quantitative means for early health assessment of thin films and filling the technological gap in subcritical state detection of intelligent dimming thin films.

[0095] Existing detection methods focus only on performance changes under a single physical field or simple stress superposition, failing to understand the complex evolution of failure mechanisms under multi-field coupling. This invention utilizes a phase-locked coherent superposition technique involving acoustic, optical, and electrical multi-physics fields to excite a phase-locked coherent superposition detection mechanism at the multi-physics focusing point guided by the subcritical defect entropy state matrix. After phase-locked demodulation, displacement amplitude and strain energy density spectra are extracted, reconstructing an energy dissipation topology network covering three characteristic frequency bands: orientation polarization of the microcrystalline medium, viscoelastic relaxation of the ligand material, and resonance of the entire film structure. By identifying topological defect nodes where the energy flow convergence exceeds the critical dissipation threshold and spatially registering them with the subcritical defect entropy state matrix, a complete causal chain visualization map is established, from interface debonding to energy dissipation anomalies and then to macroscopic failure. This topological analysis framework reveals the cross-scale migration path of energy flow during failure, providing a novel theoretical perspective and analytical tool for understanding the failure mechanisms of thin films under complex conditions.

[0096] Existing lifetime prediction methods rely on empirical statistical models, which cannot reflect the individual failure characteristics of each thin film, let alone generate proactive compensation strategies. This invention creates a service limit extrapolation model based on the self-organized criticality of dissipative structures. It uses the fluctuation amplitude of the non-equilibrium energy dissipation intensity index within the wet partial pressure modulation cycle as a measure of the degree of deviation from the equilibrium state, and the connectivity evolution of the three-band energy dissipation topology network as a measure of the change in the orderliness of the dissipative structure. By tracking the critical point of transition from the near-equilibrium linear region to the dissipative structure bifurcation region, it achieves a paradigm shift from empirical statistics to physical mechanism-driven lifetime prediction. Simultaneously, based on the distribution density of interface debonding precursors and the spatial aggregation degree of topological defect nodes in the subcritical defect entropy state matrix, it generates personalized adaptive compensation strategies that include electric field frequency-phase premodulation parameters, local humidity barrier layer thickness gradient design parameters, and microcrystalline medium orientation axis pre-tilt angle distribution parameters, achieving an intelligent leap from a unified standard to a one-film-one-strategy approach. Ultimately, the service limits, compensation strategies, and failure prediction maps are encapsulated into a quality-adaptive digital twin that can be embedded in thin-film flexible integrated circuits. This enables the thin film to have intelligent management capabilities throughout its entire lifecycle, including self-sensing, self-diagnosis, and self-optimization, providing a disruptive technological path for ensuring the reliability of high-end display dimming films.

[0097] The embodiments and / or implementation methods described above are merely preferred embodiments and / or implementation methods for implementing the technology of the present invention, and are not intended to limit the implementation methods of the technology of the present invention in any way. Any person skilled in the art may make some modifications or alterations to other equivalent embodiments without departing from the scope of the technical means disclosed in the present invention, but these should still be regarded as the technology or embodiments that are substantially the same as the present invention. This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. The above descriptions are only preferred embodiments of this application. It should be noted that due to the limitations of written expression, while there are objectively infinite specific structures, those skilled in the art can make several improvements, modifications, or changes without departing from the principles of this application, and can also combine the above technical features in an appropriate manner. These improvements, modifications, changes, or combinations, or the direct application of the inventive concept and technical solution to other situations without modification, should all be considered within the scope of protection of this application.

Claims

1. A method for detecting a sunroof display film, characterized in that, include: A frequency-converting electric field is applied to the thin film under test under controllable wet partial voltage modulation, and picosecond photoelastic imaging data are acquired simultaneously to construct a dielectric relaxation-photoelastic coupling field. Based on the non-cooperative evolution region of birefringence and wave velocity, interface debonding precursors are identified and encoded as a subcritical defect entropy state matrix. Using the spatial coordinates in the entropy state matrix as the focal point, photoacoustic, thermoacoustic and ultrasonic guided wave signals are simultaneously excited to form a phase-locked coherent superposition detection mechanism. The amplitude and strain energy spectrum are extracted by phase-locked demodulation, the three-band energy dissipation topology network is reconstructed, the topology defect nodes are identified and registered with the entropy state matrix, and a cross-physical field coupling failure prediction map is generated. By integrating the coupled field and the topology network, a service limit extrapolation model based on the self-organized criticality of dissipative structures is established. An adaptive compensation strategy is generated based on the entropy state matrix and the distribution of topological defect nodes. The service limit, compensation strategy and failure map are encapsulated into a quality adaptive digital twin.

2. The method for detecting a sunroof display film according to claim 1, characterized in that, The process involves applying a frequency-converted electric field to the thin film under test under controllable wet partial voltage modulation, simultaneously acquiring picosecond photoelastic imaging data, constructing a dielectric relaxation-photoelastic coupling field, and identifying interface debonding precursors based on the non-cooperative evolution region of birefringence and wave velocity, encoding them as a subcritical defect entropy state matrix. Specifically, this includes: The thin film sample to be tested is encapsulated in the detection chamber, and a controllable moisture partial pressure modulated by a sine wave is introduced into the chamber. The frequency and amplitude of the sine wave simulate the interlayer water vapor permeation-desorption cycle caused by the breathing effect of day and night temperature difference under the edge condition of the skylight seal failure. While the wet partial voltage is modulated, an alternating electric field with a frequency ranging from Hertz to kilohertz is applied to the thin film electrode, covering the response frequency band from microcrystalline medium interface polarization to ligand dipole orientation. A picosecond-level time-resolved pump-probe photoelastic imaging system was synchronously triggered to collect the spatial distribution of birefringence delay and ultrasonic surface wave velocity field distribution under each electric field frequency-wet partial voltage coupling state, and to construct a dielectric relaxation-photoelastic coupling field. Identify the interface debonding precursors in the coupled field, extract the spatial coordinates, the coupling phase difference between the current electric field frequency and the wet voltage modulation waveform, and the relaxation time distribution of the birefringence delay, and encode the three into a subcritical defect entropy state matrix.

3. The method for detecting a sunroof display film according to claim 2, characterized in that, Identify the interface debonding precursors in the coupled field, extract the spatial coordinates, the coupling phase difference between the current electric field frequency and the wet partial voltage modulation waveform, and the relaxation time distribution of the birefringence delay, and encode these three into a subcritical defect entropy state matrix, specifically including: Traverse each spatial pixel in the dielectric relaxation-photoelastic coupling field, calculate the Pearson correlation coefficient between the birefringence delay time series and the ultrasonic surface wave velocity time series of the pixel under all electric field frequency-wet partial voltage coupling states, and mark the pixels with correlation coefficients lower than a preset threshold as candidate non-cooperative evolution regions. Morphological closing operations are performed on the candidate non-cooperative evolution regions to merge adjacent pixels into connected regions and remove isolated noise points with an area smaller than the minimum identifiable defect size to obtain the spatial coordinate set of effective interface debonding precursor regions. For each effective interface debonding precursor region, extract the electric field frequency value and wet voltage phase value when the birefringence delay at its center point first deviates from the reference value by more than three times the standard deviation, and calculate the phase difference between the two and the zero point of the wet voltage modulation waveform as the coupling phase difference; An exponential fit is performed on the birefringence delay time series at the center point of the region, and the time required for it to decay from the peak to 1 / e is calculated as the relaxation time. The distribution histogram of relaxation time of all pixels in the region is then statistically analyzed. The spatial coordinate set, coupled phase difference, and relaxation time distribution histogram are encoded into a three-dimensional data structure and stored as a subcritical defect entropy state matrix.

4. The method for detecting a sunroof display film according to claim 3, characterized in that, Using the spatial coordinates in the entropy state matrix as the focusing point, photoacoustic, thermoacoustic, and ultrasonic guided wave signals are simultaneously excited to form a phase-locked coherent superposition detection mechanism. Amplitude and strain energy spectra are extracted via phase-locked demodulation, and a three-band energy dissipation topology network is reconstructed. Topological defect nodes are identified and registered with the entropy state matrix to generate a cross-physics field coupled failure prediction map, specifically including: Using the spatial coordinates in the subcritical defect entropy state matrix as the multi-physics field focusing point, the narrowband photoacoustic signal induced by Q-switched pulse laser, the dielectric thermoacoustic signal induced by electric field frequency modulation, and the broadband ultrasonic guided wave signal injected by the external piezoelectric transducer are synchronously excited at the point, so that the three signals form a phase-locked coherent superposition detection mechanism in the film thickness direction. Using the dielectric thermoacoustic signal as a phase reference, the narrowband photoacoustic signal and the broadband ultrasonic guided wave signal are demodulated by phase lock, and the displacement amplitude spectrum and strain energy density spectrum of the phase-locked coherent superposition detection mechanism are extracted. Based on the dispersion characteristics of the displacement amplitude spectrum and the attenuation characteristics of the strain energy density spectrum, the three-band energy dissipation topology of the thin film is reconstructed. Nodes in the topological network whose energy flow convergence exceeds the critical dissipation threshold are identified as topological defect nodes. The nodes are spatially registered with the subcritical defect entropy state matrix to generate a cross-physics field coupling failure prediction map.

5. The method for detecting a sunroof display film according to claim 4, characterized in that, The reconstruction of the three-band energy dissipation topology network of the thin film based on the dispersion characteristics of the displacement amplitude spectrum and the attenuation characteristics of the strain energy density spectrum specifically includes: From the displacement amplitude spectrum, the dispersion curves within the three characteristic frequency bands corresponding to the orientation polarization frequency band of the microcrystalline medium, the viscoelastic relaxation frequency band of the ligand material, and the resonance frequency band of the whole film structure are separated; from the strain energy density spectrum, the attenuation coefficient spectra corresponding to the three characteristic frequency bands are separated. Using the microstructural units of the thin film sample as nodes and the energy transfer efficiency between adjacent units as edges, the vibration mode participation coefficients of each node in different frequency bands are calculated based on the dispersion curve, and the energy dissipation weights of each edge are calculated based on the attenuation coefficient spectrum, thus constructing an initial three-band energy dissipation topology network. Spectral clustering analysis is performed on the initial three-band energy dissipation topology network to identify the main and side paths of energy dissipation in each frequency band. Redundant edges with weights below a preset threshold are removed to obtain the optimized three-band energy dissipation topology network.

6. The method for detecting a sunroof display film according to claim 5, characterized in that, Nodes in the topology network whose energy convergence exceeds the critical dissipation threshold are identified as topological defect nodes. Spatial registration of these nodes with the subcritical defect entropy state matrix is ​​performed to generate a cross-physics field coupled failure prediction map, specifically including: Traverse all nodes in the three-band energy dissipation topology network, calculate the energy flow convergence degree of each node, and mark nodes whose energy flow convergence degree exceeds the preset critical dissipation threshold as candidate topology defect nodes. The spatial coordinates in the subcritical defect entropy state matrix are mapped to the node coordinate system of the three-band energy dissipation topology network. The Euclidean distance between each candidate topological defect node and the spatial coordinates of the nearest subcritical defect entropy state matrix is ​​calculated. Candidate nodes whose distance is less than a preset spatial association threshold are identified as valid topological defect nodes. For each valid topological defect node, trace its energy input path and dissipation path in the three-band energy dissipation topology network, and mark all nodes and edges on the path as potential failure propagation channels. The effective topological defect nodes, potential failure propagation channels, and their corresponding energy flow convergence and spatial coordinate registration information are superimposed onto the spatial structure map of the thin film sample to generate a cross-physical field coupled failure prediction map.

7. The method for detecting a sunroof display film according to claim 6, characterized in that, By integrating the coupled field and the topological network, a service limit extrapolation model based on the self-organized criticality of dissipative structures is established. An adaptive compensation strategy is generated based on the entropy state matrix and the distribution of topological defect nodes. The service limit, compensation strategy, and failure map are encapsulated into a quality adaptive digital twin, specifically including: By integrating the dielectric relaxation-optical elastic coupling field with the three-band energy dissipation topology network, a service limit extrapolation model is established with the self-organization criticality of the dissipation structure as the criterion. By tracking the critical point of the transition from the near-equilibrium linear region to the dissipative structure bifurcation region, the service limit number of thin films under the cyclic breathing-light-electric drive coupling condition is extrapolated. Based on the distribution density of interface debonding precursors and the spatial aggregation degree of topological defect nodes in the subcritical defect entropy state matrix, a personalized adaptive compensation strategy for the thin film is generated. The service limit cycle number, the parameter set of the adaptive compensation strategy, and the sparse feature vector of the cross-physical field coupling failure prediction spectrum are encapsulated into a quality adaptive digital twin that can be embedded in a thin-film flexible integrated circuit.

8. The method for detecting a sunroof display film according to claim 7, characterized in that, The method of extrapolating the service limit number of the thin film under cyclic breathing-light irradiation-electric drive coupling conditions by tracking the critical point of the transition from the near-equilibrium linear region to the dissipative structure bifurcation region specifically includes: From the service limit extrapolation model, the time series data of the non-equilibrium energy dissipation intensity index within the wet partial voltage modulation cycle are extracted. A sliding window variance analysis is performed on the time series data of the non-equilibrium energy dissipation intensity index to calculate the fluctuation amplitude of the non-equilibrium energy dissipation intensity index within each window, thereby generating a time series sequence of the fluctuation amplitude of the non-equilibrium energy dissipation intensity index. Simultaneously, the connectivity evolution data of the three-band energy dissipation topology network is extracted from the service limit extrapolation model to generate a connectivity time series. The time series of fluctuation amplitude of the non-equilibrium energy dissipation intensity index and the time series of connectivity are analyzed together. The cross-correlation function of the two on the time axis is calculated. The time point when the cross-correlation function value first exceeds the preset critical correlation threshold is marked as the critical point when the dissipation structure transitions from the near-equilibrium linear region to the bifurcation region. Based on the number of cycles corresponding to the critical point, and combined with the conversion relationship between the preset accelerated aging factor and the measured operating conditions, the service limit number of cycles of the thin film under the cyclic breathing-light irradiation-electric drive coupling condition is extrapolated.

9. The method for detecting a sunroof display film according to claim 8, characterized in that, Based on the distribution density of interface debonding precursors and the spatial clustering degree of topological defect nodes in the subcritical defect entropy state matrix, a personalized adaptive compensation strategy for the thin film is generated, specifically including: Read the distribution density data of interface debonding precursors in the subcritical defect entropy state matrix, and identify spatial regions with distribution density exceeding a preset threshold as high-risk areas; at the same time, read the spatial aggregation data of the topological defect nodes, identify spatial regions with aggregation degree exceeding a preset threshold as energy dissipation anomaly areas, and take the spatial intersection of the two as the priority compensation area. For the priority compensation region, the electric field frequency-wet voltage coupling phase difference of the region is extracted, the center value and dispersion of its statistical distribution are calculated, and the frequency-phase premodulation parameters of the driving electric field are generated. Based on the spatial distribution characteristics of the priority compensation region at the edge of the film, and combined with the energy flow convergence direction of the region in the three-band energy dissipation topology network, the thickness gradient design parameters of the local humidity barrier layer are generated. Based on the current distribution state of the orientation axis of the microcrystalline medium in the priority compensation region, and combined with the anisotropic evolution characteristics of the birefringence retardation, the pre-tilt angle distribution parameters of the orientation axis of the microcrystalline medium are generated.

10. A detection system for a sunroof display film, characterized in that, A method for detecting a sunroof display film according to any one of claims 1-9, comprising: The multi-field coupling excitation and subcritical entropy state encoding module is used to apply a frequency-converting electric field to the thin film under controllable wet partial voltage modulation, and simultaneously acquire picosecond photoelastic imaging data to construct a dielectric relaxation-photoelastic coupling field, identify interface debonding precursors and encode them as a subcritical defect entropy state matrix. The multi-physics phase-locked superposition detection and energy dissipation topology reconstruction module is connected to the multi-field coupled excitation and subcritical entropy state encoding module. It is used to excite multi-physics phase-locked coherent superposition signals with the spatial coordinates in the subcritical defect entropy state matrix as the focusing points, extract the amplitude spectrum and strain energy spectrum through phase-locked demodulation, reconstruct the three-band energy dissipation topology network, identify topology defect nodes and register them with the entropy state matrix to generate a cross-physics coupling failure prediction map. The dissipative structure modeling and adaptive digital twin module is connected to the multi-field coupled excitation and subcritical entropy state encoding module and the multi-physics phase-locked superposition detection and energy dissipation topology reconstruction module. It is used to fuse the coupled fields and topology network to establish a service limit extrapolation model based on the self-organized criticality of dissipative structures, generate personalized adaptive compensation strategies, and encapsulate the service limit, compensation strategy and failure prediction map into a quality adaptive digital twin.

Images