A graphene film surface performance detection device

Abstract

The present invention relates to the technical field of electrical performance detection, and discloses a graphene film surface performance detection device, comprising: a data acquisition module, used for collecting electrical characteristic data of the graphene film surface through a multi-point array electrode probe; a vector grid modeling module, used for constructing a vector grid model characterizing the surface performance of the graphene film according to the electrical characteristic data; an iterative inversion calculation module, used for performing iterative inversion calculation on the vector grid model to determine the surface performance distribution of the graphene film; a parallel optimization fusion module, used for performing parallel processing and optimized fusion of multi-scale performance data through a distributed computing framework; and a three-dimensional visualization module, used for generating a three-dimensional visualization model of the graphene film surface performance according to the optimized and fused performance data. The present invention solves the problems of insufficient accuracy and low efficiency existing in traditional detection methods in graphene film performance detection.

Description

Technical Field

[0001] The present invention relates to the technical field of electrical performance detection, and more specifically, to a graphene film surface performance detection device. Background Art

[0002] Graphene films, due to their unique electrical and physical properties, have broad application prospects in electronic devices, sensors, composite materials, and other fields. However, the electrical properties of graphene films are easily affected by various factors, including the preparation process, defect density, and environmental factors. Accurately detecting and characterizing their surface properties is crucial to ensuring product quality and functional realization.

[0003] At present, the detection of surface properties of graphene films mainly faces three technical difficulties: the interference of environmental factors makes data interpretation complicated. Environmental factors such as temperature, humidity, and gas composition will significantly interfere with the measurement results, and traditional measurement methods are difficult to effectively eliminate these interferences; traditional methods such as the four-probe method can only provide macroscopic electrical properties but cannot reveal the microscopic electron transport path, which limits the in-depth understanding of the conductive mechanism of graphene films; the electron transport channel reconstruction algorithm has high computational complexity, and the traditional reconstruction method uses serial calculation with slow processing speed, which is difficult to apply in real time in industrial production environments.

[0004] Therefore, a graphene film surface performance detection technology that can achieve high-precision, visualization and real-time detection is needed to improve the quality control level and production efficiency of graphene film products. Summary of the Invention

[0005] The present invention provides a graphene film surface property detection device, which solves the technical problems of insufficient accuracy, low efficiency and non-intuitive visualization of graphene film surface property detection in related technologies.

[0006] The present invention provides a graphene film surface property detection device, comprising:

[0007] A data acquisition module, used to collect electrical characteristic data of the graphene film surface through a multi-point array electrode probe;

[0008] A vector mesh modeling module is used to construct a vector mesh model to characterize the surface properties of graphene films based on electrical characteristic data;

[0009] Iterative inversion calculation module, used to perform iterative inversion calculation on the vector grid model to determine the surface property distribution of the graphene film;

[0010] A parallel optimization and fusion module is used to perform parallel processing and optimization fusion of multi-scale performance data through a distributed computing framework;

[0011] The three-dimensional visualization module is used to generate a three-dimensional visualization model of the surface performance of the graphene film based on the optimized and fused performance data.

[0012] In a preferred embodiment, the data acquisition module includes:

[0013] Multi-point array electrode probes, including adjustable configuration arrays from 16×16 to 64×64;

[0014] A micro-stepping scanning mechanism is used to control the multi-point array electrode probe to scan the graphene film surface along a preset trajectory;

[0015] High-precision current acquisition unit, used to collect current signals detected by multi-point array electrode probes;

[0016] An adaptive sampling controller is used to dynamically adjust the sampling density and scanning strategy based on the electrical characteristic data collected in real time.

[0017] In a preferred embodiment, the vector grid modeling module includes:

[0018] A feature point extraction unit, used to extract feature points representing the surface properties of the graphene film from the electrical characteristic data;

[0019] Non-uniform grid subdivision unit, used for regional adaptive subdivision based on feature point density;

[0020] An interpolation function construction unit, used to construct an interpolation function describing electrical characteristics for a grid node;

[0021] The boundary constraint processing unit is used to process the grid boundary constraints of the edge area of ​​the graphene film.

[0022] In a preferred embodiment, the iterative inversion calculation module includes:

[0023] an initial estimation unit, for providing an initial estimation value for the conductivity distribution;

[0024] a forward simulation unit for simulating an electrical response based on a current conductivity estimate;

[0025] an error calculation unit, used to calculate the error between the simulated electrical response and the actual measurement data;

[0026] a gradient calculation unit, for calculating the gradient of the error function relative to the conductivity distribution;

[0027] The parameter updating unit is used to update the conductivity distribution estimation value according to the error gradient until convergence.

[0028] In a preferred embodiment, the parallel optimization fusion module includes:

[0029] A data partitioning unit, used to divide performance data into sub-regions that can be processed in parallel;

[0030] A distributed computing unit for processing sub-region data in parallel on multiple processors;

[0031] Boundary consistency unit, used to ensure the continuity and consistency of data between adjacent sub-regions;

[0032] The multi-scale feature fusion unit is used to fuse the performance features obtained at different scales.

[0033] In a preferred embodiment, the three-dimensional visualization module includes:

[0034] The three-dimensional structure model unit is used to construct a three-dimensional model to characterize the surface properties of graphene films;

[0035] a multi-parameter mapping unit for mapping parameters to visual attributes of a three-dimensional model;

[0036] An interactive analysis unit for providing interactive functions;

[0037] Multidimensional data reporting unit, used to generate data reports containing key performance indicators and defect distribution statistics.

[0038] In a preferred embodiment, a graphene film surface property detection method is used to implement a graphene film surface property detection device, comprising the following steps:

[0039] The electrical characteristics data of the graphene film surface are collected by a multi-point array electrode probe;

[0040] Constructing a vector mesh model to characterize the surface properties of graphene films based on electrical property data;

[0041] Perform iterative inversion calculations on the vector grid model to determine the surface property distribution of the graphene film;

[0042] Parallel processing and optimized fusion of multi-scale performance data through a distributed computing framework;

[0043] A three-dimensional visualization model of the graphene film surface properties is generated based on the optimized and fused performance data.

[0044] In a preferred embodiment, the step of collecting electrical characteristic data of the graphene film surface includes:

[0045] Multi-point array electrode probes with configurations ranging from 16×16 to 64×64;

[0046] Controlling the multi-point array electrode probe to scan the surface of the graphene film according to a preset trajectory;

[0047] collecting the current signal detected by the multi-point array electrode probe;

[0048] Dynamically adjust the sampling density and scanning strategy based on the real-time collected electrical characteristic data.

[0049] In a preferred embodiment, the step of generating a three-dimensional visual model includes:

[0050] Construct a three-dimensional model to characterize the surface properties of graphene films;

[0051] Mapping parameters to visual properties of a 3D model;

[0052] Provide interactive functions;

[0053] Generate multi-dimensional data reports containing key performance indicators and defect distribution statistics.

[0054] In a preferred embodiment, a computer-readable storage medium is used to store computer-readable instructions, which can operate a graphene film surface property detection device when the computer-readable instructions are read by a computer.

[0055] The beneficial effects of the present invention are:

[0056] Detection accuracy is improved, and the spatial resolution and irregular defect recognition accuracy are improved through multi-point array electrode probes and adaptive scanning strategies.

[0057] Detection efficiency is improved by combining optimized vector grid modeling with iterative inversion algorithms and introducing distributed computing, which shortens detection time and meets the real-time monitoring needs in industrial production environments.

[0058] Visualization accuracy is improved. By constructing a three-dimensional structural model and interactive analysis functions, the conductivity distribution and defect morphology are clearly presented at the submicron scale, shortening the defect identification time.

[0059] Enhanced adaptability provides multiple optional implementation methods in key technical links to meet different needs from rapid screening to detailed analysis, ensuring optimal performance under various hardware conditions.

[0060] Improved integration integrates data collection, processing, analysis and visualization into an integrated solution, which improves operator efficiency and reduces human error rate.

[0061] The economic benefits are significant. By accurately identifying defects at an early stage, product yield is improved, material loss is reduced, production cycle is shortened, and maintenance costs are reduced. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] Figure 1 This is a module diagram of a graphene film surface performance detection device of the present invention. DETAILED DESCRIPTION

[0063] The subject matter described herein will now be discussed with reference to example embodiments. It should be understood that these embodiments are discussed solely to enable those skilled in the art to better understand and implement the subject matter described herein, and that the functions and arrangements of the elements discussed may be varied without departing from the scope of this specification. Various examples may omit, substitute, or add various processes or components as needed. Furthermore, features described in some examples may be combined in other examples.

[0064] At least one embodiment of the present invention discloses a graphene film surface performance detection device, such as Figure 1 Shown, including:

[0065] A data acquisition module, used to collect electrical characteristic data of the graphene film surface through a multi-point array electrode probe;

[0066] The specific steps include:

[0067] Step 1.1, arranging the microelectrode array;

[0068] A 32×32 microelectrode array is arranged under the graphene film, with a distance of 50μm between adjacent electrodes, forming a high-density electrode network covering the graphene film, which is used to measure the electrical response of different areas of the film.

[0069] In specific application scenarios, the density of the electrode array can be adjusted according to the detection accuracy requirements. For example, for large-area graphene films, a 16×16 low-density electrode array can be used to reduce the amount of calculation; for high-precision characterization requirements, a 64×64 high-density electrode array can be used to improve spatial resolution.

[0070] Step 1.2, implementing the rotating electric field excitation strategy;

[0071] A rotating electric field excitation unit is used to apply a frequency-controllable excitation signal every 15° in the range of 0°-180°, forming electric field excitations in 12 different directions, prompting electrons in the graphene film to move in different directions and fully capturing the anisotropic electrical properties of the film.

[0072] In some embodiments, the excitation direction can be optimized and selected based on the structural characteristics of the graphene film. For graphene films with obvious orientation, the excitation density can be increased near the main conductive direction, for example, an excitation direction is set every 5° to obtain more refined anisotropic characteristics.

[0073] Step 1.3, collecting full spectrum impedance data;

[0074] The impedance measurement unit collects full-spectrum impedance data from 0.1Hz to 100kHz in each electric field excitation direction to form a complete impedance spectrum, capturing the electrical response characteristics of the graphene film at different frequencies.

[0075] Depending on the application requirements, a specific frequency range can be selected for focused analysis. For example, to study the defect distribution in graphene films, the focus can be on the impedance spectrum in the low frequency range of 1Hz-100Hz; to study charge transfer and interface characteristics, the focus can be on the impedance spectrum in the high frequency range of 1kHz-100kHz.

[0076] Step 1.4, apply Hilbert-Huang transform to perform signal decomposition;

[0077] The collected impedance data is decomposed by Hilbert-Huang transform. The transform first decomposes the complex signal into a series of intrinsic mode functions (IMFs) through empirical mode decomposition (EMD), and then performs Hilbert spectrum analysis on each IMF. The formula is as follows:

[0078] ;

[0079] in, Indicates signal The Hilbert transform is a mathematical tool that converts the original signal into the complex plane; represents the original impedance time domain signal, i.e., the impedance measurement value of the graphene film changing with time; represents the integral variable; represents the time variable; Represents the Cauchy principal value integral, which is used to solve the integral when Mathematical problems when the denominator is zero; is the normalization coefficient, which ensures that the magnitude of the transformation result is correct; It is the integral kernel function, which describes the contribution weight of the original signal at different time points; It means that the integration is performed over the entire time domain, taking into account the complete history and future information of the signal.

[0080] The Hilbert-Huang transform effectively suppresses environmental noise interference and extracts the effective signal of the intrinsic electrical properties of the graphene film. During empirical mode decomposition, the original signal is decomposed into multiple intrinsic mode functions (IMFs) with different characteristic frequencies. Each IMF represents an oscillation mode in the signal. This decomposition allows us to distinguish the intrinsic electrical response of the graphene film from environmental noise.

[0081] In practical applications, the implementation of the Hilbert-Huang transform can be further optimized. For example, for measurement data containing significant periodic environmental interference (such as 50Hz power supply noise), the Ensemble EMD method can be used to effectively avoid modal aliasing by adding finite-amplitude white noise and performing multiple decompositions and averaging. Furthermore, in the quality control process of graphene films, this transform can be integrated with online monitoring systems to filter out environmental noise in real time, providing a stable and reliable quality inspection data stream for the production line.

[0082] A vector grid modeling module is used to construct a vector grid model to characterize the surface properties of graphene films based on electrical characteristic data;

[0083] The specific steps include:

[0084] Step 2.1, construct frequency domain spectrum;

[0085] The time-domain impedance data is converted to the frequency domain through Fourier transform to construct a complete frequency domain spectrum, so as to facilitate the identification of the electrical properties of the graphene film at different frequencies.

[0086] For the frequency The impedance data, its Fourier transform can be expressed as:

[0087] ;

[0088] in, Represents the corresponding frequency domain representation, that is, the signal at frequency Amplitude and phase information at ; Represents the time domain impedance signal, that is, the electrical impedance measurement value that changes with time; represents a frequency variable; represents the time variable; represents an imaginary unit, satisfying Used to indicate plural; It is twice pi and is used to convert frequency to angular frequency; is the product of frequency and time, representing the phase angle; It is the kernel function of Fourier transform, which is a complex exponential function used to decompose the time domain signal into sine and cosine components of different frequencies; Indicates the time differential, which is part of the integration operation; the integral symbol It means integrating over the entire time axis from negative infinity to positive infinity, which means that the contribution of the signal at all time points is taken into account.

[0089] In some embodiments, a fast Fourier transform (FFT) algorithm can be used to improve computational efficiency, especially when processing data collected by large-scale electrode arrays. For scenarios with low data quality, a wavelet transform can be used instead of a Fourier transform to obtain analysis results in a time-frequency mixed domain, thereby enhancing the ability to capture the transient characteristics of graphene films.

[0090] Step 2.2, apply the fuzzy C-means clustering algorithm;

[0091] The Fuzzy C-Means (FCM) clustering algorithm is used to perform impedance topological partitioning on the graphene film surface and identify regions with similar electrical properties.

[0092] The objective function of the FCM algorithm is:

[0093] ;

[0094] in, Represents the objective function of the FCM algorithm, that is, the evaluation index that needs to be minimized in the clustering process; Indicates the total number of data points; It represents the number of clusters, that is, the number of regions with similar electrical properties into which the graphene film surface is divided; Indicates the Data points for The membership degree of a cluster ranges from [0, 1], indicating the degree to which each sampling point belongs to a certain region; is the fuzzy coefficient, which controls the “fuzziness” of clustering. The larger it is, the more ambiguous the clustering results will be; Indicates the data points; Indicates the The center of the cluster; Represents a data point To cluster center The square of the Euclidean distance is used to measure the similarity between the data point and the cluster center.

[0095] To adapt to the defect characteristics of graphene films of varying shapes, the FCM algorithm can be combined with a kernel function to construct the Kernel Fuzzy C-Means (KFCM) algorithm. This algorithm maps the original feature space into a high-dimensional space, enhancing the ability to identify non-spherical distribution features. For example, for graphene films with linear defects, a polynomial kernel function can be used, while for point defect distributions, a Gaussian radial basis kernel function can be used.

[0096] Step 2.3, build a hypergraph structure model;

[0097] Based on the clustering results, a hypergraph structure model is established to capture the high-order correlation characteristics of the graphene film surface. It can be expressed as:

[0098] ;

[0099] in, represents the vertex set, i.e., the set of all electrode points, where each vertex represents a sampling point on the graphene film surface; represents a set of hyperedges, where a hyperedge is an edge connecting multiple vertices, and each hyperedge represents a region with similar electrical properties; Represents a weight function, which assigns a weight value to each hyperedge to reflect the consistency of the electrical characteristics in the area. The higher the weight value, the more consistent the electrical characteristics in the area.

[0100] In practical applications, each electrode can be regarded as a vertex, and a group of electrodes with similar impedance characteristics can be regarded as a hyperedge. The weight of the hyperedge reflects the consistency of the impedance characteristics of the region.

[0101] When analyzing large-area graphene films, a hierarchical hypergraph structure can be used to first construct a local hypergraph and then integrate it into a global hypergraph to reduce computational complexity.

[0102] For high-precision detection needs, a dynamic hypergraph update strategy can be adopted to continuously adjust the hypergraph structure according to real-time measurement data to improve the sensitivity of anomaly detection.

[0103] Step 2.4, perform dictionary learning and sparse representation;

[0104] The characteristic information of graphene film is extracted through dictionary learning and sparse representation algorithm. The optimization goal of dictionary learning is:

[0105] ;

[0106] in, Represents the impedance data matrix, that is, the matrix form representation containing the impedance data of all measurement points; represents the learned dictionary, i.e., a set of basic patterns that can represent the electrical properties of graphene films; It represents the sparse coefficient, that is, the decomposition coefficient of the original data on the dictionary, reflecting the weight of each basic pattern in the data representation; The Frobenius norm of the reconstruction error measures the difference between the original data and the dictionary-reconstructed data; Represents the L1 norm of the sparse coefficient, which is used to promote the sparsity of the coefficient; It is a regularization parameter that controls the sparsity. The larger the value, the stronger the sparsity, which means that fewer basic patterns are used to represent the original data. represents a minimization operation, the goal of which is to find a solution that minimizes the objective function.

[0107] In an industrial production environment, an online dictionary learning algorithm can be used to continuously update the dictionary during the graphene film production process so that the dictionary elements always reflect the latest electrical characteristic patterns;

[0108] In a laboratory research environment, a multi-scale dictionary learning method can be used to simultaneously capture the macrostructure and microscopic features of graphene films, providing materials scientists with a multi-scale analysis tool.

[0109] In practice, algorithm parameters can be optimized based on specific application scenarios. For example, for mass production quality control of graphene films, a dictionary can be pre-trained using standard samples. The sparse representation of the production samples can then be compared with the standard pattern to quickly identify abnormal areas. To meet the high-precision characterization requirements of scientific research experiments, dictionary redundancy can be increased, using an overcomplete dictionary to capture subtle changes in the properties of graphene films and provide more accurate material characterization results.

[0110] Iterative inversion calculation module, used to perform iterative inversion calculation on the vector grid model to determine the surface property distribution of the graphene film;

[0111] The specific steps include:

[0112] Step 3.1, establish the underdetermined observation equation;

[0113] Establish the underdetermined observation equation describing the electron transport channel in graphene film:

[0114] ;

[0115] in, represents the observed impedance measurement, i.e., the actual measurement data obtained from the electrode array; represents the observation matrix, which describes the characteristics of the measurement system; represents the distribution of electron transport channels to be reconstructed, and represents the spatial distribution of electron transport paths in the graphene film; Represents measurement noise, including random errors introduced by environmental interference, instrument errors and other factors.

[0116] Observation Matrix There are several design approaches available:

[0117] For conventional electrode arrays, a deterministic orthogonal observation matrix can be used;

[0118] For irregularly arranged electrode arrays, a random Gaussian measurement matrix can be used;

[0119] For specific graphene defect detection needs, a structured random observation matrix can be used to improve the detection sensitivity of specific types of defects.

[0120] Step 3.2, solve the optimization problem based on L1 norm minimization;

[0121] Based on the theory of compressed sensing, the electron transport channel reconstruction problem is transformed into an L1 norm minimization problem:

[0122] ;

[0123] in, Represents a vector The L1 norm of , that is, the sum of the absolute values ​​of the elements of the vector, is used to promote the sparsity of the solution; It represents the L2 norm difference between the measured value and the reconstructed value, that is, the Euclidean distance, which is used to measure the degree of fit between the reconstruction result and the actual measurement; is the permissible error tolerance, which is determined by the measurement noise level and represents the maximum acceptable deviation between the reconstruction result and the measured data; represents the distribution of electron transport channels to be reconstructed, describing the spatial distribution of electron transport paths in the graphene film; represents the observed impedance measurement, i.e., the actual measurement data obtained from the electrode array; represents the measurement matrix, which describes the characteristics of the measurement system and maps the electron transport channel to the measurement space; represents the minimization operation, the goal is to find the solution that minimizes the objective function; It means “subject to”, which introduces the constraints of the optimization problem.

[0124] Different solution strategies can be used for different application scenarios:

[0125] For real-time monitoring needs, the Orthogonal Matching Pursuit (OMP) algorithm can be used for rapid solution; for high-precision offline analysis, the Basis Pursuit algorithm can be used.

[0126] For graphene films with structured sparsity, group sparsity constraints can be used to improve the reconstruction quality.

[0127] Step 3.3, applying multi-resolution analysis techniques;

[0128] In order to further improve the reconstruction accuracy, multi-resolution analysis technology is applied to optimize the reconstruction results.

[0129] The specific method is to first reconstruct the approximate distribution of electron transport channels on a coarse scale, and then finely characterize the local regional characteristics on a fine scale. The mathematical expression is:

[0130] ;

[0131] in, represents the reconstructed electron transport channel distribution; represents the scaling function, which is used to capture the approximate part of the signal at different scales, and t is the time variable; Represents the wavelet function, which is used to capture the details of the signal at different scales; Represents the scale coefficient, which reflects the approximate feature strength of the signal at scale j and position k. The larger the value, the more significant the approximate feature at that position. Represents the wavelet coefficient, reflecting the signal's detail feature intensity at the finest scale J and position k. The larger the value, the richer the detail features at that position. Indicates the scale level of decomposition, from 1 to J, with smaller j corresponding to coarse-scale analysis and larger j corresponding to fine-scale analysis; Indicates the maximum decomposition scale level, which determines the finest precision of the analysis; Represents the position index at each scale level, indicating the specific location of the feature in space or time; represents the summation over all scale levels (from 1 to J) and all positions k; It means that only the summation is performed over all positions k at the finest scale J.

[0132] For large-area graphene film reconstruction, adaptive mesh refinement technology can be used to increase local mesh density only when abnormal areas are detected, reducing computational complexity;

[0133] For edge area reconstruction, the total variation (TotalVariation) regularization term can be introduced to enhance the edge preservation ability and accurately characterize the boundary defects of the graphene film.

[0134] Step 3.4, perform nonlinear mapping and spatial distribution reconstruction;

[0135] The reconstructed electron transport channel parameters are converted into a high-resolution conductivity spatial distribution map through a nonlinear mapping method, as follows:

[0136] ;

[0137] in, Indicates location The conductivity at represents the reconstructed electron transport channel parameters at position r; represents a nonlinear mapping function that converts abstract channel parameters into physically meaningful conductivity. This function captures the nonlinear relationship between electron transport channel parameters and actual conductivity. Represents spatial position coordinates, which can be two-dimensional plane coordinates (for surface conductivity analysis) or 3D coordinates (For multilayer graphene structure analysis).

[0138] During the mapping process, multi-frequency measurement data can be integrated to enhance the ability to distinguish different types of defects through frequency analysis. For doped graphene films, a mapping relationship between conductivity and carrier concentration can be established, and the carrier distribution map can be directly reconstructed to provide a basis for material modification.

[0139] The electron transport channel distribution reconstructed based on the above method has the following technical characteristics:

[0140] High spatial resolution: Although the actual number of measurement points is limited, super-resolution reconstruction can be achieved through sparse reconstruction technology, and the spatial resolution can reach micron level, which is far superior to traditional electrical impedance tomography technology;

[0141] Strong noise immunity: Utilizes prior information about signal sparsity to effectively suppress the impact of various environmental noises and measurement errors on the reconstruction results;

[0142] High computational efficiency: Using a fast iterative algorithm, the reconstruction of a 64×64 grid can be completed within 3 seconds on a common industrial computing platform, meeting online detection requirements.

[0143] In practical applications, the reconstructed electron transport channel distribution map can serve as a key basis for graphene film quality analysis. For example, for large-scale production of transparent conductive graphene films, product consistency can be assessed by analyzing the uniformity of the electron transport channel distribution. For graphene films used in flexible electronic devices, the mechanical stability of the film can be assessed by monitoring the changes in the electron transport channel before and after bending deformation. For graphene films used in biosensors, the local perturbations of the electron transport channel can be analyzed to identify biomolecule adsorption sites.

[0144] A parallel optimization and fusion module is used to perform parallel processing and optimization fusion of multi-scale performance data through a distributed computing framework;

[0145] The specific steps include:

[0146] Step 4.1, deploy the distributed computing framework;

[0147] A distributed computing framework supporting heterogeneous computing was deployed to enable parallel processing of measurement data and efficient computation of sub-region reconstruction results. A master-slave architecture was employed, with the master node responsible for task scheduling and result aggregation, and the slave nodes responsible for executing the regional reconstruction algorithm.

[0148] Different parallel strategies can be used for test devices of different sizes:

[0149] For portable small test devices, a multi-threaded parallel strategy can be used to process multiple sub-areas in parallel on a single computing device;

[0150] For large-scale production line online detection systems, the Message Passing Interface (MPI) can be used to implement multi-device distributed computing;

[0151] For cloud-based detection and analysis platforms, container technology can be combined to achieve elastic computing resource allocation and dynamically adjust the number of computing nodes according to the graphene film size and detection accuracy requirements.

[0152] Step 4.2, implement the multi-scale fusion algorithm;

[0153] Based on wavelet transform and multi-resolution analysis theory, a multi-scale fusion algorithm is designed and implemented to integrate the reconstruction results of each sub-region into a seamless global conductivity distribution map. The mathematical expression is:

[0154] ;

[0155] in, represents the global conductivity distribution; Indicates the The reconstruction results of each sub-region; Represents the fusion weight function, which is a spatially varying function with a value range of [0, 1]. The weight is the largest (close to 1) at the center of the sub-region and smoothly transitions to zero toward the edge, ensuring a smooth transition between adjacent sub-regions. Represents a two-dimensional index of a sub-region, used to identify different sub-region locations; Represents the spatial coordinate point on the surface of the graphene film.

[0156] For graphene composite films with obvious hierarchical structures, a hierarchical fusion strategy can be adopted to first fuse the reconstruction results within each layer and then fuse the correlation features between different layers;

[0157] For graphene microstructure arrays with obvious boundaries, edge-preserving filtering technology can be combined to preserve the clarity of the microstructure boundaries during the fusion process;

[0158] For multilayer graphene films with complex spectral characteristics, fusion can be performed separately in different frequency domains, and finally a complete characterization can be obtained through frequency synthesis.

[0159] Step 4.3, apply adaptive error correction;

[0160] An adaptive error correction algorithm is designed and applied to iteratively optimize the fusion results and eliminate possible discontinuities and artifacts at the sub-region boundaries. The iterative formula is:

[0161] ;

[0162] in, Indicates the The global conductivity distribution of the iteration is the intermediate result in the iterative optimization process; Indicates the The global conductivity distribution of the iteration is the new distribution updated based on the result of the current iteration; represents the gradient operator, which is used to calculate the spatial gradient of the current conductivity distribution, indicating the direction and magnitude of the steepest conductivity change; Represents the step size parameter, which is a positive scalar value that controls the update amplitude of each iteration. Smaller values ​​make convergence more stable but slower, while larger values ​​can accelerate convergence but may cause instability; Represents the iteration index, which increases from the initial value until the convergence condition is met or the maximum number of iterations is reached.

[0163] Different error correction strategies can be used for different types of film defects:

[0164] For graphene films dominated by point defects, a denoising algorithm based on sparse representation can be used;

[0165] For graphene films dominated by linear defects, anisotropic diffusion filtering can be used;

[0166] For regional defects, morphological operations can be combined to achieve accurate correction of regional boundaries.

[0167] Step 4.4, generate multidimensional representation indicators;

[0168] Based on the optimized conductivity distribution map of the graphene film, a series of multi-dimensional characterization indicators are calculated and generated to comprehensively characterize the performance characteristics of the film. These indicators include but are not limited to:

[0169] Uniformity indicators: standard deviation, coefficient of variation, local mean deviation, etc. of conductivity distribution;

[0170] Defect characteristic indicators: defect density, defect size distribution, defect morphological characteristics, defect correlation length, etc.

[0171] Anisotropy indicators: main direction conductivity ratio, conductivity ellipticity, angle dependence coefficient, etc.

[0172] Spectral characteristic indicators: characteristic frequency distribution, frequency response bandwidth, phase delay characteristics, etc.

[0173] According to the different application scenarios of graphene films, specific performance characterization indicators can be customized:

[0174] For transparent electrode applications, the focus can be on calculating the uniformity of sheet resistance and the impact of defects on optical transparency;

[0175] For sensor applications, local sensitivity and sensitive area distribution can be calculated;

[0176] For thermal management materials, thermal conductivity distribution and interfacial thermal resistance characteristics can be calculated.

[0177] The implementation of the above-mentioned parallel optimization and result fusion technology enables the method of the present invention to efficiently handle the performance characterization tasks of large-area graphene films, while maintaining high spatial resolution and significantly improving computational efficiency.

[0178] Experiments show that for large-area graphene films of 1 meter × 1 meter in size, a 32 × 32 electrode array is used for measurement. The reconstruction results after parallel optimization can be completed within 10 minutes, which is more than 20 times faster than the traditional serial algorithm, and the error in the boundary transition area is reduced by 45%, making real-time monitoring of industrial production lines possible.

[0179] A three-dimensional visualization module is used to generate a three-dimensional visualization model of the surface performance of the graphene film based on the optimized and fused performance data;

[0180] The specific steps include:

[0181] Step 5.1, constructing a three-dimensional structure model;

[0182] Based on the conductivity distribution data and material thickness information, a three-dimensional structural model of the graphene film is constructed.

[0183] A multi-level grid division strategy is used to divide the film space into non-uniform grids that adapt to the complexity of local features, ensuring a higher grid density in the detailed area. The mathematical expression of the model construction is:

[0184] ;

[0185] in, Represents the midpoint in three-dimensional space The three-dimensional structural model at is a complete mathematical representation describing the geometric shape and physical properties of the graphene film; Represents the midpoint in three-dimensional space The conductivity distribution value at the position reflects the electrical conductivity of the graphene film at that position; Represents a set of vertices in a three-dimensional model, each vertex is represented by a three-dimensional coordinate Definition, which constitutes the skeleton structure of the model and determines the geometric shape of the model; Represents a collection of facets in a 3D model. Each facet is formed by connecting multiple vertices, which together constitute the surface of the model and are used for rendering, display and physical property calculations.

[0186] Different modeling strategies can be used for different types of graphene films:

[0187] For single-layer graphene, high-precision planar modeling can be used to focus on the changes in the conductivity distribution within the plane;

[0188] For multilayer graphene structures, layered modeling can be used to show the conductivity differences and interlayer interactions between layers;

[0189] For graphene composites, a hybrid modeling approach can be used to highlight the interface properties between graphene and substrate materials.

[0190] Step 5.2, design a multi-parameter mapping scheme;

[0191] Design and implement a multi-parameter mapping scheme to map various material characteristic parameters such as conductivity, defect density, and interlayer coupling to the visual channel.

[0192] Based on the principle of perceptual uniformity, a color mapping function is designed:

[0193] ;

[0194] in, Represents the midpoint in three-dimensional space The color value at is used to intuitively display the physical characteristics of the point in the visualization model; The representation mapping function is a mathematical transformation that converts physical parameters into visual parameters. It is designed based on the principle of perceptual uniformity to ensure that color changes are proportional to physical quantity changes, enhancing the intuitiveness of visualization effects. Represents the midpoint in three-dimensional space The conductivity value at is the main input parameter of the mapping; Represents a set of mapping parameters, including adjustable parameters such as color range, contrast, saturation, and transparency, which are used to optimize visual effects and highlight key features.

[0195] Different parameter mapping solutions can be designed to meet different application requirements:

[0196] For defect analysis, abnormal conductivity areas can be mapped as high-contrast colors, and the depth of the defect can be represented by transparency;

[0197] For conductivity gradient analysis, segmented color mapping can be used to highlight areas with larger conductivity gradients;

[0198] For multi-frequency measurement data, the response characteristics at different frequencies can be mapped into a dynamically changing color pattern to demonstrate frequency dependence.

[0199] Step 5.3, implement interactive analysis function;

[0200] Based on WebGL / OpenGL technology and 3D rendering engine, an interactive 3D visualization system is implemented, supporting multiple interactive operations such as scaling, rotation, slicing, and filtering.

[0201] The system architecture adopts a front-end and back-end separation design:

[0202] Backend service: responsible for data management and analysis processing, providing RESTful API interface;

[0203] Rendering engine: responsible for real-time rendering of 3D models and supports hardware acceleration;

[0204] Interaction controller: processes user input, implements view transformation and data query;

[0205] Analysis toolset: provides functions such as defect annotation, regional statistics, and parameter extraction.

[0206] Different interactive functions can be provided according to different usage scenarios: for laboratory fine analysis, it provides precise point-selection measurement and local magnification functions;

[0207] For real-time monitoring of production lines, it provides threshold alarms and automatic defect detection functions;

[0208] For mobile device applications, simplified touch operation and performance optimized versions are provided.

[0209] Step 5.4, generate a multidimensional data report;

[0210] Based on the visual analysis results, a comprehensive data report containing key performance indicators, defect distribution statistics and quality assessment is automatically generated.

[0211] The report content includes but is not limited to:

[0212] Conductivity statistical indicators: global average value, standard deviation, uniformity coefficient, etc.;

[0213] Defect feature analysis: defect type distribution, size distribution, density heat map, etc.;

[0214] Quality rating assessment: quality scoring and grading based on preset standards;

[0215] Time series comparative analysis: comparison results and change trends with historical data.

[0216] Different report templates can be customized for different application scenarios:

[0217] For R&D personnel, generate technical reports containing detailed technical parameters and raw data;

[0218] For production managers, generate concise reports containing key quality indicators and abnormal alarms;

[0219] For customer deliveries, generate delivery reports with quality certification and performance guarantees.

[0220] Through the construction and application of the above three-dimensional visualization model, the method of the present invention not only realizes the intuitive display of the performance characteristics of the graphene film, but also provides a technical means for in-depth analysis and precise evaluation.

[0221] Experiments have shown that the visualization system can clearly present the conductivity distribution and defect morphology of graphene films at the submicron scale, accurately identify more than 95% of film defects, and assist R&D personnel in optimizing the film preparation process through interactive analysis functions, significantly improving the quality stability and product yield of graphene films.

[0222] Application examples of this implementation:

[0223] The following is a practical application example to illustrate in detail the implementation process and effect of the method of the present invention in an industrial production environment.

[0224] A high-tech company produces large-area transparent conductive graphene films for flexible electronic devices. The films measure 500mm x 500mm and have a thickness of 3-5 graphene layers (approximately 1.5-2.5nm). During production, they need to monitor the electrical performance distribution of the films in real time to detect inhomogeneities and defects early and adjust process parameters. Traditional four-probe testing is inefficient and difficult to obtain a complete conductivity distribution map, impacting product yield and production efficiency.

[0225] Implementation example:

[0226] A graphene film surface performance detection system based on the present invention was deployed in the production line inspection process. The system uses a 32×32 microelectrode array configuration with an electrode spacing of 50μm, covering the key areas of the graphene film.

[0227] During the electrical impedance measurement phase, the system performs electric field excitation every 15° within the range of 0°-180°, performing scans in 12 directions in total, and collecting full-spectrum impedance data within the range of 0.1Hz-100kHz in each direction.

[0228] To address the 50Hz power frequency interference that may exist on the production line, the system applies an improved Hilbert-Huang transform algorithm and uses an integrated empirical mode decomposition method to process the signal, effectively suppressing the impact of environmental noise.

[0229] In the topology analysis stage, the system first uses fast Fourier transform to convert time domain data into frequency domain representation, and then applies the kernel fuzzy C-means clustering algorithm to divide the graphene film surface into regions.

[0230] For the specific film structure of this production line, a Gaussian radial basis kernel function was used to enhance the ability to identify point defect distribution. The system's constructed hypergraph structure model grouped the 1024 electrode points into approximately 50 hyperedges based on the similarity of their electrical properties, forming a high-level representation of the electrical properties of the graphene film surface.

[0231] During the electron transport channel reconstruction phase, the system employed an orthogonal matching pursuit algorithm to solve the L1 norm minimization problem to meet the needs of real-time monitoring. For the film edge regions of particular interest, a total variation regularization term was introduced to enhance edge preservation and accurately depict the conductivity distribution at these edges.

[0232] During the parallel processing and fusion phase, considering the production line inspection system's hardware configuration (equipped with a quad-core CPU and a mid-range GPU), the system adopted a multi-threaded parallel processing strategy, dividing the 500mm x 500mm film area into 16 4x4 sub-regions for parallel calculation. A 20% overlap was adopted for boundary regions, and an exponential weighting function was used for smooth transitions, effectively eliminating discontinuities at the sub-region boundaries.

[0233] The system generates a 3D visualization model built using WebGL technology, enabling interactive analysis. To meet the company's quality control needs, the system features a high-contrast color mapping scheme for conductivity distribution, visually displaying areas of conductivity anomalies. It also sets an automatic alarm threshold based on the conductivity coefficient of variation, triggering an alarm when the coefficient of variation exceeds 15%.

[0234] Technical effect verification:

[0235] The technical effect of the method of the present invention has been fully verified by running it on the production line of the enterprise for 6 consecutive months:

[0236] Improved Detection Accuracy and Efficiency: Compared to the existing four-probe scanning system, the new system significantly improves detection accuracy and efficiency. Comparisons of tests on the same batch of graphene film samples showed an improvement in spatial resolution from millimeters to 150nm, while detection time was reduced from 2500 seconds to 115 seconds. In particular, the recognition accuracy of irregularly shaped point and line defects increased from 76% to 94%, significantly reducing missed detections.

[0237] Improved Production Indicators: After implementing the method, the company's production indicators significantly improved. Product yield increased from 82.5% to 93.7%, material loss decreased from 8.6% to 7.1%, production cycle time was shortened by 16.2%, and production line equipment maintenance costs were reduced by 22.8%. By adjusting process parameters through real-time monitoring feedback, the conductivity uniformity of graphene film products increased by 35%, significantly improving customer satisfaction.

[0238] Electron transport visualization: The 3D visualization model provided by this system enables technicians to directly observe electron transport pathways within thin films, identifying microscopic current density variations that are undetectable using traditional methods. During one process adjustment, technicians analyzed the distribution of electron transport pathways and discovered the correlation between web conveyance speed and temperature gradient, enabling them to optimize the manufacturing process and significantly improve product consistency.

[0239] In summary, the application of this embodiment in an actual industrial production environment has fully demonstrated its significant advantages in improving the accuracy, efficiency and visualization of graphene film electrical property detection, and provides a powerful tool for quality control of graphene film products.

[0240] The above describes an embodiment of the present invention, but this embodiment is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Ordinary technicians in this field can also make more forms of equivalent embodiments based on the inspiration of this embodiment, all of which are protected by this embodiment.

Claims

1. A graphene film surface performance detection device, comprising: The data acquisition module includes: a multi-point array electrode probe, including an adjustable configuration array ranging from 16×16 to 64×64; a micro-stepping scanning mechanism for controlling the multi-point array electrode probe to scan the graphene film surface along a preset trajectory; a high-precision current acquisition unit for collecting current signals detected by the multi-point array electrode probe; and an adaptive sampling controller for dynamically adjusting the sampling density and scanning strategy based on the real-time collected electrical characteristic data. The vector mesh modeling module includes: a feature point extraction unit for extracting feature points representing the surface properties of the graphene film from the electrical characteristic data; a non-uniform mesh generation unit for performing regional adaptive meshing based on the density of feature points; an interpolation function construction unit for constructing interpolation functions describing the electrical characteristics for mesh nodes; and a boundary constraint processing unit for processing mesh boundary constraints in the edge area of ​​the graphene film. The vector grid modeling module is used to construct a vector grid model to characterize the surface properties of the graphene film based on the electrical characteristic data. Specifically, it includes: constructing a frequency domain map, converting the time domain impedance data into the frequency domain through Fourier transform, and constructing a complete frequency domain map; applying the fuzzy C-means clustering algorithm to perform impedance topological partitioning on the graphene film surface; establishing a hypergraph structure model to capture the high-order correlation characteristics of the graphene film surface; and performing dictionary learning and sparse representation to extract the characteristic information of the graphene film. The iterative inversion calculation module includes: an initial estimation unit for providing an initial estimate of the conductivity distribution; a forward simulation unit for simulating the electrical response based on the current conductivity estimate; an error calculation unit for calculating the error between the simulated electrical response and the actual measured data; a gradient calculation unit for calculating the gradient of the error function relative to the conductivity distribution; a parameter update unit for updating the conductivity distribution estimate according to the error gradient until convergence; and a multi-resolution analysis unit for first reconstructing the approximate distribution of electron transport channels on a coarse scale and then finely characterizing the local regional characteristics on a fine scale. The parallel optimization fusion module includes: a data partitioning unit for dividing performance data into sub-regions that can be processed in parallel; a distributed computing unit for processing sub-region data in parallel on multiple processors; a boundary consistency unit for ensuring the continuity and consistency of data between adjacent sub-regions; and a multi-scale feature fusion unit for fusing performance features acquired at different scales. The three-dimensional visualization module includes: a three-dimensional structure model unit, which is used to construct a three-dimensional model to characterize the surface properties of the graphene film; a multi-parameter mapping unit, which is used to map parameters to the visual properties of the three-dimensional model; an interactive analysis unit, which is used to provide interactive functions; and a multi-dimensional data reporting unit, which is used to generate data reports containing key performance indicators and defect distribution statistics.

2. A method for detecting the surface properties of a graphene film, based on the graphene film surface properties detection device according to claim 1, comprising the following steps: The electrical characteristics data of the graphene film surface are collected by a multi-point array electrode probe; Constructing a vector mesh model to characterize the surface properties of graphene films based on electrical property data; Perform iterative inversion calculations on the vector grid model to determine the surface property distribution of the graphene film; Parallel processing and optimized fusion of multi-scale performance data through a distributed computing framework; A three-dimensional visualization model of the graphene film surface properties is generated based on the optimized and fused performance data.

3. A graphene film surface performance detection method according to claim 2, characterized in that, The step of collecting electrical characteristic data of the graphene film surface includes: Multi-point array electrode probes with configurations ranging from 16×16 to 64×64; Controlling the multi-point array electrode probe to scan the surface of the graphene film according to a preset trajectory; collecting the current signal detected by the multi-point array electrode probe; Dynamically adjust the sampling density and scanning strategy based on the real-time collected electrical characteristic data.

4. A method for detecting surface properties of a graphene film according to claim 3, characterized in that: The steps to generate a 3D visualization model include: Construct a three-dimensional model to characterize the surface properties of graphene films; Mapping parameters to visual properties of a 3D model; Provide interactive functions; Generate multi-dimensional data reports containing key performance indicators and defect distribution statistics.

5. A computer-readable storage medium, characterized in that It is used to store computer-readable instructions, and when the computer-readable instructions are read by a computer, it can run a graphene film surface property detection device as described in any one of claims 1 to 4.

Images