Detection method and system of graphene far-infrared heating sheet system

Abstract

The invention relates to the technical field of image data processing, and in particular to a detection method and system for a graphene far-infrared electric heating sheet system. Nanoparticles responsive to lasers of specific wavelengths are uniformly injected into a graphene material to generate interference patterns. Interference image data are generated by multi-wavelength laser scanning, and pre-processing, fringe detection, feature extraction and multi-frequency feature fusion are performed on the interference image data to generate multi-frequency interference fringe feature data. Multi-frequency spectrum response analysis is performed based on the data, a dynamic nanoparticle response model is established, and a comprehensive response feature vector is generated. An adaptive fault detection model is trained by using the vector, and model parameters are adjusted in real time through a dynamic correction mechanism to perform real-time fault detection and classification. A fault handling suggestion is generated in combination with an intelligent decision engine, a fault report is automatically generated and an alarm system is triggered, and the model is optimized through an online learning and feedback mechanism. The invention significantly improves the accuracy and robustness of fault detection.

Description

Technical Field

[0001] The present invention relates to the technical field of image data processing, and in particular to a detection method and system for a graphene far-infrared electric heater system. Background Art

[0002] Graphene far-infrared heater system uses the special properties of graphene materials and the principle of far-infrared radiation for heating and temperature control, and is widely used in heat treatment, heating and temperature control. In order to ensure the performance and reliability of the equipment, the graphene far-infrared heater system needs to be fault-checked before leaving the factory, among which uneven working temperature is a common fault.

[0003] The graphene far-infrared heater system has the following problems during the detection process:

[0004] The temperature distribution of the equipment is uneven when it is working, which leads to the chaotic distribution of gray values ​​in the infrared image;

[0005] Due to the temperature difference in different areas of the equipment, ghosting is likely to occur in infrared images, affecting the accuracy of fault detection.

[0006] The existing technology (Chinese invention patent, publication number: CN116883412 B, title: A graphene far-infrared electric heating equipment fault detection method) mainly uses traditional image processing algorithms to detect faults of graphene far-infrared electric heating sheets, but has the following shortcomings:

[0007] The existing technology uses traditional infrared imaging technology. Due to the uneven temperature distribution, the gray value difference in the infrared image is significant, making it difficult to accurately identify the temperature distribution. This method cannot effectively remove ghost images in the infrared image, resulting in inaccurate temperature distribution detection.

[0008] The existing technology uses a simple image processing algorithm, which cannot accurately detect temperature non-uniformity due to the ghost phenomenon in the image, resulting in a high false detection rate and unreliable detection results. Summary of the invention

[0009] In view of the many problems existing in the above-mentioned prior art, the present invention provides a detection method and system for a graphene far-infrared electric heater system. The present invention uniformly injects nanoparticles with laser responsiveness at a specific wavelength into a graphene material, so that the graphene material generates an interference pattern under laser irradiation; uses a multi-wavelength laser scanner to scan the graphene far-infrared electric heater containing nanoparticles to generate interference image data; pre-processes the interference image data and fuses multi-frequency features, performs multi-frequency spectral response analysis based on multi-frequency interference fringe feature data, and establishes a dynamic nanoparticle response model; trains an adaptive fault detection model through a comprehensive response feature vector, and adjusts the model parameters in real time to perform fault detection and classification; finally, generates fault handling suggestions in combination with an intelligent decision-making engine to achieve continuous optimization and normal operation of the system; the present invention effectively solves the problem of inaccurate detection caused by temperature non-uniformity, and improves the accuracy and reliability of fault detection.

[0010] The detection method of the graphene far-infrared heating sheet system comprises the following steps:

[0011] Nanoparticles responsive to lasers of a specific wavelength are uniformly injected into graphene materials, so that the materials can generate interference patterns under laser irradiation, and graphene far-infrared electric heating sheets containing nanoparticles are manufactured;

[0012] A multi-wavelength laser scanner is used to scan the graphene far-infrared electric heating sheet containing nanoparticles to generate interference image data of different wavelengths;

[0013] Preprocess the interference image data, perform fringe detection and segmentation, feature extraction and multi-frequency feature fusion to generate multi-frequency interference fringe feature data;

[0014] Based on the multi-frequency interference fringe feature data, multi-frequency spectral response analysis is performed to establish a dynamic nanoparticle response model, and the standardized feature data is fused with the response model to generate a comprehensive response feature vector;

[0015] The adaptive fault detection model is trained using the comprehensive response feature vector, and the model parameters are adjusted in real time during operation through a dynamic correction mechanism. The real-time interference fringe feature data is input into the model for real-time fault detection and classification, and the fault classification results are generated;

[0016] Based on the fault classification results, combined with the intelligent decision-making engine, fault handling suggestions are generated, fault reports are automatically generated and the alarm system is triggered. Through the online learning and feedback mechanism, fault handling records and feedback data are used to optimize the fault detection model and intelligent decision-making engine to ensure continuous optimization and normal operation of the system.

[0017] Preferably, the nanoparticles are gold nanoparticles or silver nanoparticles, which have laser responsiveness to wavelengths of 532 nanometers and 635 nanometers. The nanoparticles are prepared by chemical or physical methods to ensure that their size and dispersibility meet the requirements.

[0018] Preferably, the scanning path, speed and resolution of the multi-wavelength laser scanner are set according to the size and structure of the graphene far-infrared electric heater, and the specific steps include:

[0019] Calibrate the wavelength, power and stability of laser sources;

[0020] Set the scan path, speed, and resolution to ensure coverage of the entire heater surface;

[0021] Perform an initial scan and collect interferometric image data at different wavelengths.

[0022] Preferably, the preprocessing of the interference image data includes denoising using a high-precision image processing algorithm, and the specific steps include:

[0023] Use filters to remove image noise;

[0024] Improve the contrast of interference fringes to enhance image clarity.

[0025] Preferably, the fringe detection and segmentation includes using a Canny edge detection algorithm to perform fringe detection on the denoised interference image and segmenting the interference fringe area, and the specific steps include:

[0026] Extract the edges of interference fringes through edge detection algorithm;

[0027] The interference fringe area is segmented according to the extracted edge information.

[0028] Preferably, the feature extraction includes extracting the fringe spacing, fringe intensity variation and fringe shape of the fringe from the interference fringe image data, and the specific steps include:

[0029] Calculate the spacing of interference fringes;

[0030] Analyze the intensity changes of interference fringes;

[0031] Evaluate the shape characteristics of the interference fringes.

[0032] Preferably, the multi-frequency feature fusion includes fusing standardized interference fringe feature data of different wavelengths to generate multi-frequency interference fringe feature data, and the specific steps include:

[0033] Standardize the interference fringe characteristic data of different wavelengths;

[0034] A multimodal data fusion algorithm is used to fuse data of different wavelengths.

[0035] Preferably, the multi-frequency spectral response analysis includes frequency division processing of interference fringes generated by lasers of different wavelengths and extracting interference features of each frequency band, and the specific steps include:

[0036] Conduct frequency division analysis on interference image data;

[0037] The interference characteristic data of each frequency band are extracted, including fringe spacing, fringe intensity and fringe shape.

[0038] Preferably, the dynamic correction mechanism dynamically adjusts the parameters of the fault detection model by comparing the real-time interference fringe feature data with the comprehensive response feature vector to ensure the accuracy and robustness of the fault detection model under different working conditions. The specific steps include:

[0039] Compare the real-time interference fringe characteristic data with the comprehensive response characteristic vector;

[0040] Dynamically adjust model parameters based on comparison results;

[0041] Ensure the detection accuracy of the model under different working conditions.

[0042] A system for implementing the detection method of the graphene far-infrared electric heater system, comprising:

[0043] A nanoparticle injection device is used to uniformly inject nanoparticles responsive to lasers of a specific wavelength into a graphene material, so that the material can generate interference patterns under laser irradiation, and to produce a graphene far-infrared electric heating sheet containing nanoparticles;

[0044] A multi-wavelength laser scanner is used to scan the graphene far-infrared electric heating sheet containing nanoparticles to generate interference image data of different wavelengths;

[0045] An image processing unit is used to perform preprocessing, fringe detection and segmentation, feature extraction and multi-frequency feature fusion on interference image data to generate multi-frequency interference fringe feature data;

[0046] A spectral response analysis unit, used to perform multi-frequency spectral response analysis based on multi-frequency interference fringe feature data, establish a dynamic nanoparticle response model, and fuse the standardized feature data with the response model to generate a comprehensive response feature vector;

[0047] A fault detection unit is used to train an adaptive fault detection model using a comprehensive response feature vector, and adjust model parameters in real time during operation through a dynamic correction mechanism, input real-time interference fringe feature data into the model for real-time fault detection and classification, and generate a fault classification result;

[0048] The intelligent decision-making and feedback unit is used to generate fault handling suggestions based on the fault classification results in combination with the intelligent decision-making engine, automatically generate fault reports and trigger the alarm system, and optimize the fault detection model and intelligent decision-making engine using fault handling records and feedback data through online learning and feedback mechanisms to ensure continuous optimization and normal operation of the system.

[0049] Compared with the prior art, the advantages and beneficial effects of the present invention are:

[0050] The present invention uniformly injects nanoparticles with laser responsiveness at a specific wavelength into the graphene material, so that it can generate interference patterns under laser irradiation, solving the problem that the traditional method cannot accurately detect temperature distribution;

[0051] The present invention utilizes a multi-wavelength laser scanner to generate interference image data, which greatly improves the accuracy and analyzability of the data;

[0052] The present invention performs preprocessing, fringe detection and segmentation, feature extraction and multi-frequency feature fusion on interference image data, thus ensuring the accuracy and reliability of data processing;

[0053] The present invention performs multi-frequency spectral response analysis based on multi-frequency interference fringe feature data, establishes a dynamic nanoparticle response model, and solves the core problem of temperature non-uniformity detection;

[0054] The present invention uses the comprehensive response feature vector to train the adaptive fault detection model, and adjusts the model parameters in real time during operation through a dynamic correction mechanism to ensure the accuracy and robustness of the fault detection model;

[0055] The present invention combines an intelligent decision-making engine to generate fault handling suggestions, and optimizes the model through online learning and feedback mechanisms, thereby achieving continuous optimization and normal operation of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 It is a schematic diagram of the process of the present invention;

[0057] Figure 2 A schematic diagram of nanoparticles injected into graphene material in the present invention;

[0058] Figure 3 A schematic diagram of multi-wavelength laser scanning and interference image generation in the present invention;

[0059] Figure 4 is a flow chart of preprocessing of interference image data in the present invention;

[0060] Figure 5 It is a schematic diagram of multi-frequency spectral response analysis and feature fusion in the present invention;

[0061] Figure 6A schematic diagram of the adaptive fault detection and dynamic correction mechanism in the present invention;

[0062] Figure 7 A schematic diagram of the intelligent decision-making engine and feedback mechanism in the present invention;

[0063] Figure 8 It is a structural block diagram of the system of the present invention. DETAILED DESCRIPTION

[0064] Hereinafter, embodiments of the present disclosure will be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of the present disclosure. In the following detailed description, for ease of explanation, many specific details are set forth to provide a comprehensive understanding of the embodiments of the present disclosure. However, it is obvious that one or more embodiments may also be implemented without these specific details. In addition, in the following description, descriptions of known structures and technologies are omitted to avoid unnecessary confusion of the concepts of the present disclosure.

[0065] The terms used herein are only for describing specific embodiments and are not intended to limit the present disclosure. The terms "comprise", "include", etc. used herein indicate the existence of the features, steps, operations and / or components, but do not exclude the existence or addition of one or more other features, steps, operations or components.

[0066] All terms (including technical and scientific terms) used herein have the meanings commonly understood by those skilled in the art unless otherwise defined. It should be noted that the terms used herein should be interpreted as having a meaning consistent with the context of this specification and should not be interpreted in an idealized or overly rigid manner.

[0067] In the case of using expressions such as "at least one of A, B, and C, etc.", it should generally be interpreted in accordance with the meaning of the expression generally understood by those skilled in the art (for example, "a system having at least one of A, B, and C" should include but is not limited to a system having A alone, B alone, C alone, A and B, A and C, B and C, and / or A, B, C, etc.). In the case of using expressions such as "at least one of A, B, or C, etc.", it should generally be interpreted in accordance with the meaning of the expression generally understood by those skilled in the art (for example, "a system having at least one of A, B, or C" should include but is not limited to a system having A alone, B alone, C alone, A and B, A and C, B and C, and / or A, B, C, etc.).

[0068] Some block diagrams and / or flow charts are shown in the accompanying drawings. It should be understood that some boxes or combinations thereof in the block diagrams and / or flow charts can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that these instructions can create a device for implementing the functions / operations described in these block diagrams and / or flow charts when executed by the processor. The technology of the present disclosure can be implemented in the form of hardware and / or software (including firmware, microcode, etc.). In addition, the technology of the present disclosure can take the form of a computer program product on a computer-readable storage medium storing instructions, which can be used by an instruction execution system or in combination with an instruction execution system.

[0069] like Figure 1 As shown, the detection method of the graphene far-infrared electric heater system comprises the following steps:

[0070] like Figure 2 As shown, nanoparticles responsive to lasers of a specific wavelength are uniformly injected into a graphene material, so that it can generate interference patterns under laser irradiation, and a graphene far-infrared electric heating sheet containing nanoparticles is manufactured;

[0071] Nanoparticles responsive to lasers of specific wavelengths are uniformly injected into the graphene material. The key to selecting nanoparticles is their responsiveness to lasers of specific wavelengths, which usually include gold nanoparticles or silver nanoparticles, which have strong responsiveness to lasers of 532 nanometers and 635 nanometers. These nanoparticles are prepared by chemical or physical methods to ensure that their size and dispersibility meet the requirements. The specific injection method can use chemical deposition, electrochemical deposition or physical vapor deposition techniques to evenly distribute the nanoparticles in the graphene material. The principle of this step is to control the distribution and density of the nanoparticles so that they can produce a significant optical response under laser irradiation.

[0072] After the nanoparticles are uniformly injected into the graphene material, the next step is to use the optical properties of these nanoparticles to produce interference patterns under laser irradiation. When the laser irradiates the graphene heater containing nanoparticles, the nanoparticles will scatter and absorb the laser, resulting in the formation of interference fringes on the surface of the heater. The characteristics of these interference fringes (such as fringe spacing, intensity changes, and fringe shapes) reflect the physical state and working state of the heater. The formation principle of the interference pattern is based on the wave characteristics of light. When the light wave passes through the nanoparticle distribution area, an interference effect occurs, forming a striped optical pattern on the surface of the material.

[0073] After these graphene materials containing nanoparticles are made into far-infrared electric heaters, these electric heaters will produce clear interference patterns under laser irradiation, providing a data basis for subsequent image processing and analysis. The clarity and contrast of the interference fringes directly affect the accuracy and sensitivity of the detection. Therefore, by reasonably selecting the type, size and distribution of nanoparticles, the quality of the interference pattern can be significantly improved.

[0074] For example, in practical applications, gold nanoparticles are evenly distributed in graphene materials by electrochemical deposition, and then made into far-infrared electric heaters. When the electric heater is irradiated with lasers of 532 nm and 635 nm wavelengths, the interference fringe patterns produced are very clear and have high contrast. These fringe patterns can be captured and analyzed by high-resolution imaging equipment. The experimental results show that the interference patterns produced by the nanoparticle-containing graphene far-infrared electric heater prepared by this method under laser irradiation have good stability and repeatability, providing reliable data support for subsequent fault detection and performance monitoring.

[0075] In summary, by uniformly injecting nanoparticles responsive to lasers of a specific wavelength into graphene materials, causing them to generate interference patterns under laser irradiation, and making graphene far-infrared electric heaters containing nanoparticles, the present invention effectively utilizes the optical properties of nanoparticles to generate interference patterns that are easy to detect and analyze. This method not only improves the sensitivity and accuracy of detection, but also provides a high-quality data basis for intelligent fault detection, ensuring efficient monitoring and reliable operation of the graphene far-infrared electric heater system.

[0076] Preferably, the nanoparticles are gold nanoparticles or silver nanoparticles, which have laser responsiveness to wavelengths of 532 nanometers and 635 nanometers. The nanoparticles are prepared by chemical or physical methods to ensure that their size and dispersibility meet the requirements.

[0077] The present invention selects gold nanoparticles or silver nanoparticles as nanoparticle materials. These nanoparticles are widely used in various optical and biological detection fields due to their excellent optical properties and chemical stability. Gold nanoparticles and silver nanoparticles have a strong local surface plasmon resonance (LSPR) effect under laser irradiation with a wavelength of 532 nanometers and 635 nanometers, which can significantly enhance the scattering and absorption of light, thereby forming a clear interference pattern in the graphene material. The main reason for selecting these two nanoparticles is that they have excellent optical response performance in the visible light and near-infrared spectrum range, which can significantly improve the detection sensitivity and accuracy.

[0078] These nanoparticles are prepared by chemical or physical methods to ensure that their size and dispersibility meet the requirements. Common chemical methods include sol-gel method, chemical reduction method and electrochemical deposition method. These methods can accurately control the size and morphology of nanoparticles, so that they have uniform particle size distribution and good dispersibility. For example, gold nanoparticles with uniform size and good dispersibility can be prepared by chemical reduction method. This method generates nanoparticles by reducing gold ions in solution, and the particle size can be controlled between 10 and 100 nanometers by adjusting the reaction conditions. Physical methods such as physical vapor deposition and laser evaporation vaporize the material by physical means and deposit it on the substrate to form a uniform nanoparticle film. These methods ensure the size and dispersibility of nanoparticles and avoid the problems of particle agglomeration and uneven size.

[0079] In practical applications, gold nanoparticles prepared by electrochemical deposition are uniformly dispersed in graphene materials. These nanoparticles can produce a strong LSPR effect under the irradiation of lasers with wavelengths of 532 nanometers and 635 nanometers, resulting in significantly enhanced light scattering and absorption, thereby forming a clear interference fringe pattern on the surface of the graphene heater. The experimental results show that the graphene heater containing nanoparticles prepared by this method exhibits excellent optical response characteristics under laser irradiation, and the interference fringes are clearly visible with high contrast, providing a high-quality data basis for subsequent image analysis and fault detection.

[0080] In summary, by selecting gold nanoparticles or silver nanoparticles and preparing them by chemical or physical methods, ensuring that their size and dispersibility meet the requirements, the present invention effectively utilizes the optical properties of nanoparticles to generate high-quality interference patterns. The strong response of these nanoparticles under specific wavelength laser irradiation makes the detection of graphene electric heater system more sensitive and accurate. By accurately controlling the preparation conditions of nanoparticles, ensuring its uniform distribution and stable performance in graphene materials, providing reliable technical support for intelligent fault detection, ensuring efficient monitoring and operation of the system.

[0081] like Figure 3-Figure 4 As shown, a multi-wavelength laser scanner is used to scan a graphene far-infrared electric heating sheet containing nanoparticles to generate interference image data of different wavelengths;

[0082] Preferably, the scanning path, speed and resolution of the multi-wavelength laser scanner are set according to the size and structure of the graphene far-infrared electric heater, and the specific steps include:

[0083] Calibrate the wavelength, power and stability of laser sources;

[0084] Set the scan path, speed, and resolution to ensure coverage of the entire heater surface;

[0085] Perform an initial scan and collect interferometric image data at different wavelengths.

[0086] The laser source of the multi-wavelength laser scanner needs to be calibrated to ensure that the wavelength, power and stability of the laser meet the requirements. The wavelength of the laser source should be set to 532 nanometers and 635 nanometers, which match the responsiveness of the nanoparticles. The power and stability of the laser need to be precisely adjusted to ensure that a stable and consistent optical signal can be generated during the scanning process. The wavelength of the laser source determines how the laser interacts with the nanoparticles, while the power and stability affect the clarity and repeatability of the interference pattern. Through precise calibration, it is ensured that the laser can produce stable interference fringes when irradiating the electric heater, and obtain high-quality laser source output, laying the foundation for subsequent scanning and data acquisition.

[0087] According to the size and structure of the graphene far-infrared heater, set the scanning path, speed and resolution. The scanning path should cover the entire surface of the heater to ensure that no area is missed. The scanning speed should be moderate, ensuring sufficient time for data collection while avoiding thermal effects that affect the results due to too long a time. The resolution setting should be based on the detection accuracy requirements to ensure that the details of the interference fringes can be accurately captured. The scanning path determines the area covered by the laser, and the scanning speed and resolution affect the detail of the data and the detection efficiency. Reasonable setting of these parameters can optimize the scanning effect and improve the accuracy of the data. By reasonably setting the scanning parameters, ensure that the scanning process is comprehensive and accurate, and obtain high-quality interference image data.

[0088] After completing the laser source calibration and scanning parameter setting, an initial scan is performed. During the initial scan, the laser scanner scans the graphene far-infrared electric heater according to the preset path, speed and resolution. During the scanning process, the laser interacts with the nanoparticles to generate interference fringes. These fringes are collected by imaging equipment such as high-resolution cameras to form interference image data. During the scanning process, the laser interacts with the nanoparticles on the surface of the electric heater to produce an interference effect and form a fringe pattern. The imaging device captures these fringe patterns and converts them into digital image data. Through the initial scan, interference image data of different wavelengths are generated to provide raw data support for subsequent image processing and feature extraction.

[0089] For example, in practical applications, a multi-wavelength laser scanner can be selected, and the wavelength of the laser source can be set to 532 nm and 635 nm. The laser source is calibrated by precision instruments to ensure accurate wavelength and stable power. Next, the scanning path is set according to the size of the electric heater (for example, 10 cm × 10 cm) so that the laser can cover the entire surface. The scanning speed is set to 10 mm / s to ensure sufficient exposure time while avoiding heat accumulation effects. The resolution is set to 100 pixels per millimeter to ensure that the details of the interference fringes are clearly captured. During the initial scan, the interference image data is collected by a high-resolution camera and saved to a computer for subsequent processing. The experimental results show that through such a scanning setting, high-quality interference image data can be obtained, with clear interference fringes and high contrast, which verifies the effectiveness and feasibility of this method.

[0090] Through this step, a multi-wavelength laser scanner is used to scan the graphene far-infrared electric heater containing nanoparticles to generate interference image data of different wavelengths. The specific steps include calibrating the wavelength, power and stability of the laser source, setting the scanning path, speed and resolution to ensure that the entire surface of the electric heater is covered, and performing initial scanning and data acquisition. This method uses the interaction between laser and nanoparticles to generate stable and high-quality interference fringes, collects data through imaging equipment, and provides a basis for subsequent image analysis and fault detection. This detection method not only improves the sensitivity and accuracy of detection, but also realizes comprehensive scanning and precise detection of electric heaters, and is suitable for the detection needs of electric heaters of various sizes and structures.

[0091] Preprocess the interference image data, perform fringe detection and segmentation, feature extraction and multi-frequency feature fusion to generate multi-frequency interference fringe feature data;

[0092] Preprocess the interference image data. The purpose of preprocessing is to remove noise and background interference in the image and enhance the clarity and contrast of the image. Specifically, high-precision image processing algorithms such as filters, histogram equalization and other methods can be used. Filters can effectively remove random noise in the image, while histogram equalization can enhance the contrast of the interference fringes, making them more obvious and easy to identify in the image. The principle of the preprocessing step is to adjust the pixel value distribution of the image by mathematical methods to improve the quality and usability of the image. The effect is to significantly improve the clarity of the interference image, ensuring that the subsequent steps can accurately identify and process the interference fringes.

[0093] Next is fringe detection and segmentation. This step uses edge detection algorithms, such as the Canny edge detection algorithm, to detect fringe patterns on the preprocessed interference image and segment the interference fringe regions. The Canny edge detection algorithm can effectively identify edge features in images, and is particularly suitable for structures with distinct boundaries such as interference fringes. During the fringe detection process, the algorithm identifies and marks the edge positions of the interference fringes, and then segments the interference fringe regions in the image based on this edge information. The principle of fringe detection and segmentation is to identify significant edge features by analyzing the pixel gradient changes of the image. The effect is to extract clear interference fringe regions from the preprocessed interference image, providing clear objects for subsequent feature extraction.

[0094] After fringe detection and segmentation, feature extraction is performed. The purpose of feature extraction is to extract key parameters that describe fringe characteristics from the segmented interference fringe area, such as fringe spacing, fringe intensity variation, and fringe shape. Specific methods may include calculating the spacing of the fringe, extracting intensity features by analyzing the brightness variation curve of the fringe, and evaluating the geometric features of the fringe through shape analysis algorithms. The principle of feature extraction is to quantify different aspects of the interference fringe through statistical and geometric analysis methods, making them feature data that can be used for subsequent analysis. The effect is to generate a set of quantitative fringe feature data that can reflect the physical state and performance of the electric heater.

[0095] The fringe feature data of different wavelengths are multi-frequency fused to generate multi-frequency interference fringe feature data. The purpose of multi-frequency feature fusion is to integrate the fringe feature data generated by lasers of different wavelengths to form a comprehensive feature vector. This can be achieved through standardization processing and multimodal data fusion algorithms. Standardization processing can eliminate the scale differences between data of different wavelengths, while multimodal data fusion algorithms can effectively integrate multi-source data to generate a comprehensive feature vector containing multiple feature information. The principle of multi-frequency feature fusion is to integrate data from different wavelengths through mathematical and statistical methods to provide a more comprehensive feature description. The effect is to generate a comprehensive, multi-dimensional feature vector, which provides rich and accurate input data for subsequent fault detection.

[0096] For example, in practical applications, the interference image is denoised by the preprocessing algorithm, and the Canny edge detection algorithm is used to identify and segment the interference fringe area. Then, the feature data is extracted by calculating the fringe spacing and intensity changes. Finally, the feature data of 532 nanometers and 635 nanometers are standardized and fused to generate a comprehensive vector containing multi-frequency features. This process ensures the integrity and accuracy of the data and provides a solid foundation for subsequent fault detection.

[0097] In summary, by preprocessing the interference image data, fringe detection and segmentation, feature extraction and multi-frequency feature fusion, the present invention can generate high-quality multi-frequency interference fringe feature data. These steps use advanced image processing and data fusion technology to ensure the clarity, accuracy and comprehensiveness of the data in principle, providing reliable data support for subsequent intelligent fault detection. This method not only improves the sensitivity and accuracy of detection, but also realizes comprehensive monitoring and analysis of the status of the electric heater system.

[0098] Preferably, the preprocessing of the interference image data includes denoising using a high-precision image processing algorithm, and the specific steps include:

[0099] Use filters to remove image noise;

[0100] Improve the contrast of interference fringes to enhance image clarity.

[0101] A filter is used to remove image noise. A filter is a mathematical tool used to smooth images and reduce noise. Common filters include mean filters, Gaussian filters, and median filters. In the present invention, a Gaussian filter can be selected because it can effectively remove high-frequency noise while retaining important structural information in the image. The Gaussian filter performs a convolution operation on the image so that the value of each pixel becomes the weighted average of the pixel values ​​in its neighborhood, thereby smoothing the image and reducing the impact of random noise. The principle of denoising is to suppress random noise in the image and improve image quality through operations in the spatial domain or frequency domain. The effect is that the noise component in the image is significantly reduced, making the interference fringes more obvious and easy to identify in the image.

[0102] Secondly, improve the contrast of the interference fringes to enhance the clarity of the image. Contrast enhancement is to adjust the gray value distribution of the image so that the brightness difference between the interference fringes and the background is more obvious. This can be achieved using techniques such as histogram equalization or contrast stretching. Histogram equalization redistributes the gray values ​​of the image to make the distribution of gray values ​​more uniform, thereby improving the contrast. Contrast stretching, on the other hand, increases the dynamic range of the image and enhances the details of the image by stretching the gray value range. The principle of contrast enhancement is to make the difference between the target feature (interference fringes) and the background more prominent by adjusting the gray value of the image. The effect is to significantly improve the clarity and visibility of the interference fringes, providing a high-quality image foundation for subsequent fringe detection and feature extraction.

[0103] For example, in practical applications, the collected interference image data is first denoised using a Gaussian filter. The specific operation is to select a suitable Gaussian kernel size (such as 5x5) and smooth the image through a convolution operation. After denoising, the random noise in the image is significantly reduced, and the interference fringes are clearer. Next, the denoised image is subjected to histogram equalization, and the contrast between the interference fringes and the background is enhanced by adjusting the grayscale distribution of the image. In the processed image, the contrast of the interference fringes is significantly improved, and the details are clearer. This series of preprocessing steps ensures the quality of the image and provides a solid foundation for subsequent image analysis.

[0104] In summary, the present invention effectively improves the quality of the interference image by preprocessing the interference image data, removing image noise with a filter, and improving image clarity by enhancing contrast. These steps utilize advanced image processing technology to ensure the clarity and contrast of the image in principle, providing high-quality data support for subsequent fringe detection and feature extraction. This preprocessing method not only improves the sensitivity and accuracy of detection, but also ensures the reliability and stability of the detection process, and is suitable for the high-precision detection requirements of the graphene far-infrared electric heater system.

[0105] Preferably, the fringe detection and segmentation includes using a Canny edge detection algorithm to perform fringe detection on the denoised interference image and segmenting the interference fringe area, and the specific steps include:

[0106] Extract the edges of interference fringes through edge detection algorithm;

[0107] The interference fringe area is segmented according to the extracted edge information.

[0108] The edge of the interference fringes is extracted by the edge detection algorithm. The Canny edge detection algorithm is a commonly used and efficient image edge detection technology that can achieve a good balance between noise suppression and edge detection. Its principle is to identify the area with drastic gradient changes, that is, the edge, by calculating the gradient of the image. When applying the Canny edge detection algorithm, the denoised interference image is first Gaussian smoothed to reduce the impact of noise; then, the gradient amplitude and direction of the image are calculated; unnecessary edge responses are removed by non-maximum suppression; dual threshold detection is used to determine the true edge, and the edge fragments are connected by lagging edge tracking. Through this series of operations, the Canny edge detection algorithm can accurately extract the edges of the interference fringes.

[0109] The interference fringe area is segmented based on the extracted edge information. The segmentation process is to separate the pixel area belonging to the interference fringe from the background in the image. This can be achieved through region growing or image segmentation algorithms. The region growing algorithm expands the region according to the seed point until the edge; the image segmentation algorithm divides the image into different regions based on the edge information. The segmented interference fringe area contains all the identified interference fringes, providing a clear object for subsequent feature extraction.

[0110] For example, in practical applications, the Canny edge detection algorithm is first applied to the denoised interference image. The specific operations include selecting an appropriate Gaussian kernel size (such as 5x5) for Gaussian smoothing, then calculating the gradient amplitude and direction of the image, and setting an appropriate double threshold (such as a low threshold of 50 and a high threshold of 150) for edge detection. Through this series of steps, all significant edges in the image will be marked. Using the region growing algorithm, starting from the seed point in the edge detection result, it gradually expands until the entire interference fringe area is segmented. The segmented interference fringe area is clearly visible in the image and is completely separated from the background area. This process ensures the integrity and accuracy of the interference fringes and provides high-quality data for subsequent feature extraction.

[0111] In this way, fringe detection and segmentation uses the Canny edge detection algorithm to accurately extract the edges of interference fringes and segment the interference fringes region based on the edge information. The Canny edge detection algorithm can accurately identify edge features in the image while reducing noise interference through its unique multi-stage detection process. The segmented interference fringe region provides clear objects and high-quality data support for subsequent feature extraction and fault detection. This method not only improves the accuracy and reliability of detection, but also ensures comprehensive monitoring and analysis of the status of the graphene far-infrared electric heater system.

[0112] Preferably, the feature extraction includes extracting the fringe spacing, fringe intensity variation and fringe shape of the fringe from the interference fringe image data, and the specific steps include:

[0113] Calculate the spacing of interference fringes;

[0114] Analyze the intensity changes of interference fringes;

[0115] Evaluate the shape characteristics of the interference fringes.

[0116] Calculate the spacing of the interference fringes. The spacing of interference fringes refers to the distance between adjacent fringes. It is an important feature of the interference pattern and reflects the phase change of the light wave in the distribution of nanoparticles. The method for calculating the spacing of fringes is usually to convert the image in the spatial domain to the frequency domain through Fourier transform. In the frequency domain, the spacing of the interference fringes can be determined by the peak position of the spectrum. Fourier transform can effectively separate the periodic components in the image, making the calculation of the fringe spacing more accurate. In this way, the spacing data of each interference fringe can be obtained, which reflects the physical state of the electric heater in different areas.

[0117] Analyze the intensity change of the interference fringes. The intensity change of the interference fringes refers to the brightness distribution of the fringes, which reflects the intensity and phase information of the light wave interference. The method for analyzing the intensity change of the fringes can be to convert the grayscale value distribution of the interference image into an intensity change curve through grayscale analysis. The specific steps include grayscale sampling of each interference fringe, calculating its brightness value, and drawing a brightness change curve. By analyzing the fluctuation amplitude and frequency of these curves, the intensity change characteristics of the interference fringes can be obtained. The intensity change characteristics can provide information about the uniformity of the heating plate material and the distribution of nanoparticles, which is very important for detecting material defects and inhomogeneities.

[0118] Evaluate the shape characteristics of the interference fringes. The shape characteristics of the interference fringes refer to the geometric shape of the fringes, including the width, curvature, and continuity of the fringes. These features can be extracted by shape analysis algorithms, such as using Hough transform to detect straight fringes or using morphological operations to analyze curved fringes. The specific steps include edge detection of the fringe area, extracting the contour of the fringe, and then calculating its geometric characteristic parameters. The shape characteristics can reveal the physical changes of the heater under different stress and thermal conditions, which is of great significance for evaluating the structural integrity and reliability of the heater.

[0119] For example, in practical applications, the spacing of interference fringes is calculated through Fourier transform to obtain a set of spacing data; the intensity change curve of interference fringes is drawn through grayscale analysis to extract the intensity change characteristics; the shape characteristics of fringes, such as the width and curvature of the fringes, are extracted through Hough transform and morphological operations. These characteristic parameters combined together can fully describe the working state of the electric heater and provide accurate data support for fault detection.

[0120] In summary, by extracting the fringe spacing, fringe intensity variation and fringe shape of the fringe from the interference fringe image data, the present invention can generate a set of key parameters describing the characteristics of the interference fringe. These steps utilize advanced image processing and analysis techniques to ensure the accuracy and reliability of feature extraction in principle. Through this method, the physical state and performance of the electric heater can be fully evaluated, providing high-quality data support for intelligent fault detection. This feature extraction method not only improves the sensitivity and accuracy of detection, but also realizes comprehensive monitoring and analysis of the state of the electric heater system, ensuring the reliable operation of the system.

[0121] Preferably, the multi-frequency feature fusion includes fusing standardized interference fringe feature data of different wavelengths to generate multi-frequency interference fringe feature data, and the specific steps include:

[0122] Standardize the interference fringe characteristic data of different wavelengths;

[0123] A multimodal data fusion algorithm is used to fuse data of different wavelengths.

[0124] The interference fringe feature data of different wavelengths are standardized. The purpose of standardization is to eliminate the scale differences between the feature data of different wavelengths so that the data of different wavelengths can be compared and fused on the same scale. The standardization method is usually to subtract the mean of the feature data and divide it by its standard deviation so that the distribution of the data has zero mean and unit variance. The principle of standardization is to adjust the distribution of the original data to a standard normal distribution through linear transformation to eliminate the deviation between different data sets. The effect is to make the interference fringe feature data of different wavelengths numerically consistent, providing a basis for subsequent fusion.

[0125] A multimodal data fusion algorithm is used to fuse data of different wavelengths. The purpose of the multimodal data fusion algorithm is to combine the standardized feature data from different wavelengths to form a comprehensive feature vector containing multi-frequency information. Commonly used multimodal data fusion algorithms include principal component analysis (PCA), linear discriminant analysis (LDA), and deep learning fusion methods. Specifically, PCA reduces the dimension of the data by finding the principal components of the data while retaining the main feature information of the data; LDA improves the separability of the data by maximizing the ratio of between-class variance to within-class variance; deep learning fusion methods use the nonlinear transformation ability of neural networks to fuse data of different modes. The principle of multimodal data fusion is to combine the features of multi-source data through mathematical and statistical methods to provide richer and more comprehensive information. The effect is to generate a comprehensive, multi-dimensional feature vector that contains the interference fringe features of different wavelengths, providing more comprehensive feature information for fault detection.

[0126] For example, in practical applications, the interference fringe feature data of 532 nm and 635 nm wavelengths are first standardized, and the data of each wavelength is subtracted from the mean and divided by the standard deviation. The standardized data is fused using the PCA algorithm, and the first few principal components are selected as the comprehensive feature vector. In this way, the obtained comprehensive feature vector contains the main feature information of the two wavelengths, which can better reflect the working status of the electric heater. The experimental results show that the multi-frequency interference fringe feature data generated by standardization and multimodal data fusion has high stability and recognition, providing accurate and reliable data support for fault detection.

[0127] In summary, through multi-frequency feature fusion, the standardized interference fringe feature data of different wavelengths are fused to generate comprehensive multi-frequency interference fringe feature data. Standardization eliminates the scale difference between feature data of different wavelengths. The multimodal data fusion algorithm integrates the characteristics of multi-source data, and the generated comprehensive feature vector provides comprehensive and accurate information for fault detection. This method not only improves the quality and consistency of feature data, but also provides rich feature information support for intelligent fault detection, ensuring the reliability and effectiveness of the detection process, and is suitable for the high-precision detection needs of graphene far-infrared electric heating sheet systems.

[0128] like Figure 5 As shown, based on the multi-frequency interference fringe feature data, multi-frequency spectral response analysis is performed to establish a dynamic nanoparticle response model, and the standardized feature data is fused with the response model to generate a comprehensive response feature vector;

[0129] Multi-frequency spectral response analysis is performed based on multi-frequency interference fringe characteristic data. The purpose of multi-frequency spectral response analysis is to use the interference fringe characteristic data of different wavelengths to analyze the optical response characteristics of nanoparticles at different frequencies. This analysis method can reveal the dynamic behavior of nanoparticles under laser irradiation of different wavelengths, including their scattering and absorption characteristics. Specifically, the characteristic information of each frequency band can be separated and extracted by performing Fourier transform or other frequency domain analysis methods on the multi-frequency interference fringe characteristic data. The principle of multi-frequency spectral response analysis is to use frequency domain analysis technology to convert time domain or space domain data into frequency domain, identify and extract the response characteristics of each frequency band. The effect is to accurately capture the optical response of nanoparticles at different wavelengths, and provide detailed data support for the establishment of subsequent models.

[0130] Establish a dynamic nanoparticle response model. The purpose of the dynamic nanoparticle response model is to establish a model that can reflect the dynamic response of nanoparticles in different environments and working conditions by analyzing multi-frequency spectral response data. This model can be trained using machine learning algorithms (such as support vector machines, random forests, neural networks, etc.). The specific steps include: collecting and organizing multi-frequency spectral response data; secondly, selecting a suitable machine learning algorithm and dividing the data into a training set and a test set; establishing a model through the training algorithm and using the test set to verify the accuracy of the model. The principle of the dynamic response model is to learn and extract the dynamic response laws of nanoparticles from a large amount of multi-frequency spectral response data through a machine learning algorithm. The effect is to generate a model that can accurately predict the response of nanoparticles and improve the detection capability of the system.

[0131] The standardized feature data is fused with the dynamic nanoparticle response model to generate a comprehensive response feature vector. The purpose of this step is to fuse the standardized feature data obtained previously with the prediction results of the dynamic response model to form a comprehensive feature vector that contains multi-frequency features and dynamic response information. The specific method can use feature-level fusion technology, such as weighted averaging, feature splicing, etc., to combine feature data from different sources. The principle of fusion is to integrate different types of feature data into a unified feature vector through mathematical operations to provide more comprehensive information. The effect is to generate a comprehensive response feature vector that contains multi-frequency spectral response and dynamic response characteristics, providing high-quality input data for subsequent fault detection.

[0132] For example, in practical applications, the multi-frequency interference fringe feature data is first Fourier transformed to separate the response features of 532 nanometers and 635 nanometers; then, the dynamic nanoparticle response model is trained using the support vector machine algorithm to learn the response law of nanoparticles through a large amount of training data; the standardized feature data and the prediction results of the dynamic response model are weighted averaged to generate a comprehensive response feature vector. This comprehensive feature vector contains multi-frequency features and dynamic response information, and can fully reflect the working status of the electric heater. The experimental results show that the comprehensive response feature vector generated by the fusion of multi-frequency spectral response analysis and dynamic response model has high accuracy and stability, providing reliable data support for fault detection.

[0133] In summary, a dynamic nanoparticle response model is established by performing multi-frequency spectral response analysis based on multi-frequency interference fringe feature data, and the standardized feature data is fused with the response model to generate a comprehensive response feature vector. The present invention uses advanced frequency domain analysis and machine learning technology to ensure the comprehensiveness and accuracy of feature data in principle. This method not only improves the sensitivity and accuracy of detection, but also realizes dynamic monitoring and analysis of the state of the graphene far-infrared electric heater system, providing strong technical support for intelligent fault detection.

[0134] Preferably, the multi-frequency spectral response analysis includes frequency division processing of interference fringes generated by lasers of different wavelengths and extracting interference features of each frequency band, and the specific steps include:

[0135] Conduct frequency division analysis on interference image data;

[0136] The interference characteristic data of each frequency band are extracted, including fringe spacing, fringe intensity and fringe shape.

[0137] Perform frequency division analysis on the interference image data. The purpose of frequency division analysis is to separate and process the interference image data according to different laser wavelengths in order to identify and extract the interference feature information at each wavelength. The specific steps include Fourier transforming the interference image and converting the time domain or space domain data to the frequency domain. Fourier transform can separate the periodic features in the image into different frequency components, so that the interference fringes of different wavelengths appear as different frequency peaks in the spectrum. By analyzing these frequency peaks, the interference fringe data of each wavelength can be separated. The principle of frequency division analysis is to use frequency domain analysis technology to deconstruct complex spatiotemporal data into frequency components that are easy to identify and process. The effect is that the interference fringe features at different wavelengths can be accurately separated and extracted, providing a clear data basis for subsequent feature extraction.

[0138] Extract the interference feature data of each frequency band, including fringe spacing, fringe intensity and fringe shape. Fringe spacing refers to the distance between adjacent interference fringes, reflecting the phase change of light waves in the distribution of nanoparticles. Fringe intensity refers to the brightness distribution of interference fringes, reflecting the intensity and phase information of light wave interference. Fringe shape refers to the geometric characteristics of interference fringes, including the width, curvature and continuity of fringes. The specific steps include calculating the fringe spacing of each frequency band, analyzing the fringe intensity change and evaluating the fringe shape characteristics from the interference image data after frequency division. The method of calculating the fringe spacing can be achieved by measuring the peak distance between adjacent fringes; the analysis of fringe intensity can be achieved by statistical analysis of grayscale values; and the evaluation of fringe shape can be achieved by edge detection and morphological analysis. The principle of extracting these feature data is to quantify different aspects of interference fringes from image data through mathematical and geometric analysis methods, so that they can become feature data for subsequent analysis. The effect is to generate a set of quantitative and multi-dimensional interference feature data, which can fully reflect the physical state and performance of the electric heater.

[0139] For example, in practical applications, the interference image data is converted to the frequency domain through Fourier transform, and the frequency peaks of 532 nanometers and 635 nanometers are identified and separated. The stripe spacing is extracted from the divided image data, and the spacing data is obtained by measuring the peak distance of adjacent stripes; the stripe intensity change is analyzed, and the intensity change data is obtained by calculating the gray value distribution of the stripes; the stripe shape characteristics are evaluated, and the geometric feature data of the stripes is extracted through edge detection and morphological operations. These feature data combined together can fully describe the response characteristics of the electric heater under different wavelength laser irradiation, and provide key data for the subsequent establishment of a dynamic response model.

[0140] In summary, by performing frequency division processing on the interference fringes generated by lasers of different wavelengths and extracting the interference feature data of each frequency band, the present invention realizes a comprehensive analysis and quantification of the interference fringes features. Frequency division analysis uses frequency domain analysis technology to separate complex interference image data into easy-to-process frequency components; feature extraction uses mathematical and geometric analysis methods to extract key parameters such as fringe spacing, fringe intensity and fringe shape from image data. This method not only improves the quality and consistency of feature data, but also provides comprehensive and accurate data support for intelligent fault detection, ensuring the reliable operation of the graphene far-infrared electric heater system.

[0141] like Figure 6 As shown, the comprehensive response feature vector is used to train the adaptive fault detection model, and the model parameters are adjusted in real time during operation through a dynamic correction mechanism. The real-time interference fringe feature data is input into the model for real-time fault detection and classification, and the fault classification results are generated;

[0142] The adaptive fault detection model is trained using the comprehensive response feature vector. The comprehensive response feature vector is generated by multi-frequency spectral response analysis and dynamic nanoparticle response model, which integrates the multi-dimensional feature data of the system under different working conditions. The fault detection model is trained using these comprehensive feature data so that it can accurately identify and classify different types of faults. The specific training process includes data preprocessing, feature selection, model selection and training. Commonly used models include support vector machines, random forests and neural networks. The principle of training is to learn the characteristic patterns of the normal working state and various fault states of the system from the comprehensive response feature vector through machine learning algorithms, so as to establish a model that can accurately classify faults. The effect is to generate a high-precision and high-robustness fault detection model that can accurately classify faults for real-time data.

[0143] The model parameters are adjusted in real time during operation through a dynamic correction mechanism. The purpose of the dynamic correction mechanism is to dynamically adjust the parameters of the fault detection model according to the interference fringe feature data collected in real time during the actual operation of the system to adapt to changes in the environment and working conditions. The specific steps include inputting the real-time interference fringe feature data into the model, comparing it with the comprehensive response feature vector, and identifying the difference between the current system state and the expected state. According to the comparison results, the model parameters are dynamically adjusted, such as adjusting the learning rate, weight, etc., to ensure that the model can accurately reflect the current system state. The principle of dynamic adjustment is to continuously optimize and adjust the model parameters through real-time feedback and adaptive algorithms to maintain high accuracy and high robustness. The effect is to ensure the detection accuracy of the fault detection model under different working conditions and to respond to changes in the system state in a timely manner.

[0144] For example, in practical applications, a neural network fault detection model is trained by comprehensive response feature vectors, which can identify the normal and faulty states of the heater under different working conditions. During the operation of the system, the interference fringe feature data of the heater is collected in real time and input into the fault detection model. If the difference between the real-time data and the comprehensive response feature vector is significant, it means that the system state has changed or there is a potential fault. At this time, the dynamic correction mechanism will adjust the parameters of the neural network model according to the comparison results, such as adjusting the learning rate or weight, to improve the adaptability of the model to the current state. Through this dynamic adjustment, the fault detection model can maintain high accuracy and high robustness under different working conditions. Experimental results show that through this dynamic correction mechanism, the fault detection model can identify and classify faults in a timely manner, effectively improving the detection accuracy and reliability of the system.

[0145] In summary, the adaptive fault detection model is trained by using the comprehensive response feature vector, and the model parameters are adjusted in real time through the dynamic correction mechanism during operation to ensure the accuracy and robustness of the fault detection model under different working conditions. The model is trained using comprehensive feature data to make it highly accurate and robust; the model parameters are adjusted in real time through the dynamic correction mechanism to enable the model to adapt to environmental changes. This method uses advanced machine learning and adaptive algorithms to ensure the accuracy and real-time nature of fault detection in principle, and significantly improves the detection performance and reliability of the graphene far-infrared electric heater system.

[0146] Preferably, the dynamic correction mechanism dynamically adjusts the parameters of the fault detection model by comparing the real-time interference fringe feature data with the comprehensive response feature vector to ensure the accuracy and robustness of the fault detection model under different working conditions. The specific steps include:

[0147] Compare the real-time interference fringe characteristic data with the comprehensive response characteristic vector;

[0148] Dynamically adjust model parameters based on comparison results;

[0149] Ensure the detection accuracy of the model under different working conditions.

[0150] Compare the real-time interference fringe feature data with the integrated response feature vector. The real-time interference fringe feature data is collected from the actual operation of the electric heater system, while the integrated response feature vector is generated based on previous data analysis and model training. By comparing these two sets of data, the difference between the current system state and the expected state can be identified. The specific comparison method can use vector distance metrics (such as Euclidean distance, Manhattan distance, etc.) or similarity metrics (such as cosine similarity). The principle of comparison is to quantify the difference between the two sets of feature data through mathematical methods, so as to identify changes or anomalies in the system state. The effect is that the state change of the system can be detected in real time, providing a basis for subsequent parameter adjustments.

[0151] Dynamically adjust the model parameters according to the comparison results. By comparing the real-time interference fringe feature data with the comprehensive response feature vector, the difference between the current system state and the expected state is identified, and the fault detection model can dynamically adjust its parameters to adapt to the current working conditions and environmental changes. The specific adjustment method can use adaptive algorithms (such as adaptive filters, online learning algorithms, etc.) to update the model parameters in real time so that it better reflects the current system state. The principle of dynamic adjustment is to enable the model to continuously learn and adapt to new data through real-time feedback and adjustment, thereby maintaining high accuracy and high robustness. The effect is to ensure the accuracy and reliability of the fault detection model under different working conditions and to respond to changes in the system state in a timely manner.

[0152] For example, in practical applications, the interference fringe feature data of the heater is collected in real time and compared with the previously established comprehensive response feature vector. If the Euclidean distance between the real-time data and the comprehensive response feature vector is found to increase significantly, it indicates that the system state has changed or a potential fault has occurred. At this time, the fault detection model updates its parameters in real time through an adaptive filter and adjusts the weights of different features to more accurately reflect the current system state. The adjusted model parameters can better adapt to the current working conditions and improve the accuracy and sensitivity of fault detection. The experimental results show that through the dynamic correction mechanism, the fault detection model can maintain high accuracy and high robustness under different working conditions, and effectively detect and identify abnormal conditions of the system.

[0153] In summary, the dynamic correction mechanism dynamically adjusts the parameters of the fault detection model by comparing the interference fringe feature data with the comprehensive response feature vector in real time, ensuring the accuracy and robustness of the model under different working conditions. By comparing the real-time data with the comprehensive features, the system state changes are identified; the model parameters are dynamically adjusted according to the comparison results, so that the model can adapt to environmental changes. This mechanism uses advanced adaptive algorithms and real-time feedback technology to ensure the dynamic adaptability and detection accuracy of the model in principle. This method not only improves the sensitivity and reliability of fault detection, but also realizes real-time monitoring and adjustment of the state of the graphene far-infrared electric heater system to ensure the stable operation of the system.

[0154] like Figure 7 As shown, based on the fault classification results, the intelligent decision-making engine generates fault handling suggestions, automatically generates fault reports and triggers the alarm system, and optimizes the fault detection model and the intelligent decision-making engine using fault handling records and feedback data through online learning and feedback mechanisms to ensure continuous optimization and normal operation of the system.

[0155] Based on the fault classification results, the intelligent decision engine generates fault handling suggestions. The fault classification results are obtained through the previous real-time fault detection and classification model, which contains the current system status information and fault type. The intelligent decision engine uses these classification results, combined with the preset rule base and expert system, to generate handling suggestions for different fault types. The principle of the intelligent decision engine is to automatically generate the best handling solution through the rule base and knowledge base combined with real-time data. The effect is to provide fast and accurate fault handling suggestions, reduce manual intervention, and improve system response speed and processing efficiency.

[0156] Automatically generate fault reports and trigger the alarm system. The fault report contains detailed information about the fault, including fault type, time of occurrence, location, scope of impact, and handling suggestions. This information is generated by the intelligent decision-making engine and a report document is generated through the automated system. At the same time, the alarm system triggers different levels of alarm notifications based on the severity of the fault and the preset alarm rules. This can be notified to relevant personnel through SMS, email, sound and light alarms, etc., to ensure that the fault can be discovered and handled in a timely manner. The principle of automatically generating fault reports and triggering the alarm system is to quickly transmit fault information to relevant personnel through automated system integration and information transmission mechanisms. The effect is to improve the timeliness and accuracy of fault response and ensure that the fault can be handled in the shortest time.

[0157] The fault detection model and intelligent decision engine are optimized by using fault processing records and feedback data through the online learning and feedback mechanism. The purpose of the online learning and feedback mechanism is to optimize and improve the fault detection model and intelligent decision engine by continuously collecting and analyzing fault processing records and feedback data during the operation of the system. The specific steps include collecting fault processing records, analyzing processing effects and feedback data, inputting these data into the online learning system, and optimizing the model and decision engine through machine learning algorithms. The principle of the online learning and feedback mechanism is to enable the system to continuously adapt to new fault types and processing requirements through continuous data collection and learning, and maintain efficient detection and decision-making capabilities. The effect is to ensure the continuous optimization and normal operation of the system and improve the robustness and adaptability of the system.

[0158] For example, in actual applications, when the system detects a heater failure in a certain part, the intelligent decision engine generates specific processing suggestions based on the fault classification results, such as adjusting the current, checking the connection line, or replacing the heater. The automated system generates a fault report and notifies the maintenance personnel by email. The maintenance personnel handle the problem according to the suggestions and record the processing process and results. These fault handling records and feedback data are collected and input into the online learning system to optimize the fault detection model and the intelligent decision engine, so that it can provide more accurate detection and suggestions in similar faults in the future. The experimental results show that through this online learning and feedback mechanism, the system's fault detection accuracy and decision-making quality have been significantly improved, ensuring the continuous optimization and stable operation of the graphene far-infrared heater system.

[0159] In summary, the present invention generates fault handling suggestions based on the fault classification results in combination with an intelligent decision-making engine, automatically generates a fault report and triggers an alarm system, and optimizes the fault detection model and the intelligent decision-making engine using fault handling records and feedback data through an online learning and feedback mechanism to ensure continuous optimization and normal operation of the system. The intelligent decision-making engine provides fast and accurate processing suggestions, the automated system generates reports and triggers alarms, and the online learning mechanism continuously optimizes system performance. This method not only improves the timeliness and accuracy of fault handling, but also ensures continuous optimization and reliable operation of the system, and is suitable for efficient fault management of graphene far-infrared electric heating sheet systems.

[0160] like Figure 8 As shown, a system for implementing the detection method of the graphene far-infrared electric heater system comprises:

[0161] The nanoparticle injection device is used to uniformly inject nanoparticles responsive to lasers of a specific wavelength into the graphene material, so that it can generate interference patterns under laser irradiation, and make a graphene far-infrared electric heating sheet containing nanoparticles; the nanoparticle injection device is used to uniformly inject nanoparticles responsive to lasers of a specific wavelength into the graphene material. These nanoparticles can generate interference patterns under laser irradiation of a specific wavelength. When selecting nanoparticles, gold nanoparticles or silver nanoparticles are preferred because they have strong responsiveness to lasers of 532 nanometers and 635 nanometers. The uniform injection of nanoparticles ensures that the graphene material can form a stable and clear interference pattern under laser irradiation, thereby improving the accuracy of subsequent detection.

[0162] The multi-wavelength laser scanner is used to scan the graphene far-infrared electric heater containing nanoparticles and generate interference image data of different wavelengths. The multi-wavelength laser scanner is used to scan the graphene far-infrared electric heater containing nanoparticles and generate interference image data of different wavelengths. Laser scanning ensures that the entire surface of the electric heater is covered by precisely controlling the scanning path, speed and resolution. The generated interference image data can reflect the physical properties of the graphene far-infrared electric heater in detail and provide high-quality input data for subsequent data processing and analysis.

[0163] The image processing unit is used to perform preprocessing, fringe detection and segmentation, feature extraction and multi-frequency feature fusion on the interference image data to generate multi-frequency interference fringe feature data; the image processing unit performs preprocessing on the interference image data, including denoising, fringe detection and segmentation, feature extraction and multi-frequency feature fusion. High-precision image processing algorithms are used to remove image noise, and the Canny edge detection algorithm is used to extract the edges of the interference fringes, segment the interference fringe area, and extract the spacing, intensity change and shape characteristics of the fringes. Multi-frequency interference fringe feature data are generated, which have high clarity and accuracy, providing a reliable basis for spectral response analysis.

[0164] The spectral response analysis unit is used to perform multi-frequency spectral response analysis based on multi-frequency interference fringe feature data, establish a dynamic nanoparticle response model, and fuse the standardized feature data with the response model to generate a comprehensive response feature vector; the spectral response analysis unit performs multi-frequency spectral response analysis based on multi-frequency interference fringe feature data to establish a dynamic nanoparticle response model. Through frequency domain analysis technology, the interference characteristics of each frequency band are extracted, and the standardized feature data are fused with the response model to generate a comprehensive response feature vector. The generated comprehensive response feature vector fully reflects the dynamic optical response characteristics of the nanoparticles, improving the accuracy and robustness of the model.

[0165] The fault detection unit is used to train the adaptive fault detection model using the comprehensive response feature vector, and adjust the model parameters in real time during operation through a dynamic correction mechanism, input the real-time interference fringe feature data into the model for real-time fault detection and classification, and generate a fault classification result; The fault detection unit uses the comprehensive response feature vector to train the adaptive fault detection model, and adjusts the model parameters in real time through a dynamic correction mechanism. The interference fringe feature data is input in real time to perform fault detection and classification, and generate a fault classification result. The model can maintain high-precision fault detection capabilities under different working conditions, identify and classify faults in real time, and ensure the reliable operation of the system.

[0166] The intelligent decision-making and feedback unit is used to generate fault handling suggestions based on the fault classification results and the intelligent decision engine, automatically generate fault reports and trigger the alarm system, and optimize the fault detection model and the intelligent decision engine using fault handling records and feedback data through the online learning and feedback mechanism to ensure continuous optimization and normal operation of the system. The intelligent decision-making and feedback unit generates fault handling suggestions based on the fault classification results and the intelligent decision engine, automatically generates fault reports and triggers the alarm system. Through the online learning and feedback mechanism, the fault handling records and feedback data are used to optimize the fault detection model and the intelligent decision engine. This unit ensures continuous optimization and normal operation of the system, improves the efficiency and accuracy of fault handling, reduces human intervention, and realizes intelligent fault management.

[0167] For example, in practical applications, gold nanoparticles are uniformly injected into graphene materials through electrochemical deposition, and then graphene far-infrared electric heaters are made. The electric heater is scanned using a multi-wavelength laser scanner to generate interference image data at wavelengths of 532 nanometers and 635 nanometers. The image processing unit performs denoising, fringe detection and feature extraction on these data to generate multi-frequency interference fringe feature data. The spectral response analysis unit performs multi-frequency spectral response analysis on these feature data, establishes a dynamic nanoparticle response model, and generates a comprehensive response feature vector. The fault detection unit uses the feature vector to train an adaptive fault detection model, adjusts the model parameters in real time, and performs fault detection and classification. Finally, the intelligent decision-making and feedback unit generates fault handling suggestions based on the classification results, automatically generates reports and triggers the alarm system, optimizes the model through an online learning mechanism, and ensures continuous optimization and normal operation of the system.

[0168] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems or computer program products. Therefore, the present application may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present application may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0169] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems) and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0170] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0171] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0172] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0173] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0174] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0175] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0176] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.

Claims

1. A method for detecting a graphene far-infrared heating sheet system, characterized in that: The following steps are involved: Nanoparticles responsive to lasers of a specific wavelength are uniformly injected into a graphene material by any one of a sol-gel method, a chemical reduction method, and an electrochemical deposition method, so that the graphene material can generate an interference pattern under laser irradiation, and a graphene far-infrared electric heating sheet containing nanoparticles is prepared; A multi-wavelength laser scanner is used to scan the graphene far-infrared electric heating sheet containing nanoparticles to generate interference image data of different wavelengths; The interference image data is preprocessed, fringe detected and segmented, feature extracted and multi-frequency feature fused to generate multi-frequency interference fringe feature data; the multi-frequency feature fusion includes fusing standardized interference fringe feature data of different wavelengths to generate multi-frequency interference fringe feature data, and the specific steps include: Standardize the interference fringe characteristic data of different wavelengths; A multimodal data fusion algorithm is used to fuse data of different wavelengths together; Based on the multi-frequency interference fringe characteristic data, a multi-frequency spectral response analysis is performed to establish a dynamic nanoparticle response model, and the standardized multi-frequency interference fringe characteristic data is integrated with the response model to generate a comprehensive response characteristic vector; the multi-frequency spectral response analysis includes frequency division processing of the interference fringes generated by lasers of different wavelengths and extracting the interference characteristics of each frequency band. The specific steps include: Conduct frequency division analysis on interference image data; Extract interference characteristic data of each frequency band, including fringe spacing, fringe intensity and fringe shape; The adaptive fault detection model is trained by using the comprehensive response feature vector, and the classification threshold, weight, and learning rate of the adaptive fault detection model are adjusted in real time during operation through a dynamic correction mechanism, and real-time interference fringe feature data are input into the model for real-time fault detection and classification to generate a fault classification result; the dynamic correction mechanism dynamically adjusts the parameters of the adaptive fault detection model by comparing the real-time interference fringe feature data with the comprehensive response feature vector, and the specific steps include: Compare the real-time interference fringe characteristic data with the comprehensive response characteristic vector; Dynamically adjust the parameters of the adaptive fault detection model according to the comparison results; Based on the fault classification results, the intelligent decision-making engine generates fault handling suggestions, automatically generates fault reports and triggers the alarm system.

2. The detection method according to claim 1, characterized in that: The nanoparticles are gold nanoparticles or silver nanoparticles, have laser responsiveness to wavelengths of 532 nanometers and 635 nanometers, and are prepared by a chemical or physical method.

3. The detection method according to claim 1, characterized in that: The scanning path, speed and resolution of the multi-wavelength laser scanner are set according to the size and structure of the graphene far-infrared electric heater, and the specific steps include: Calibrate the wavelength and power of the laser source; Set the scan path, speed, and resolution to cover the entire heater surface; Perform an initial scan and collect interferometric image data at different wavelengths.

4. The detection method according to claim 1, characterized in that: The preprocessing of the interference image data includes denoising using a high-precision image processing algorithm, and the specific steps include: Use filters to remove image noise; Improve the contrast of interference fringes to enhance image clarity.

5. The detection method according to claim 1, characterized in that: The fringe detection and segmentation includes using the Canny edge detection algorithm to perform fringe detection on the denoised interference image and segmenting the interference fringe area. The specific steps include: Extract the edges of interference fringes through edge detection algorithm; The interference fringe area is segmented according to the extracted edge information.

6. The detection method according to claim 1, characterized in that: The feature extraction includes extracting the fringe spacing, fringe intensity change and fringe shape of the fringe from the interference fringe image data, and the specific steps include: Calculate the spacing of interference fringes; Analyze the intensity changes of interference fringes; Evaluate the shape characteristics of the interference fringes.

7. A system for implementing the detection method of the graphene far-infrared heater system according to any one of claims 1 to 6, characterized in that: include: A nanoparticle injection device is used to uniformly inject nanoparticles responsive to lasers of a specific wavelength into a graphene material, so that the material can generate interference patterns under laser irradiation, and to produce a graphene far-infrared electric heating sheet containing nanoparticles; A multi-wavelength laser scanner is used to scan the graphene far-infrared electric heating sheet containing nanoparticles to generate interference image data of different wavelengths; An image processing unit is used to perform preprocessing, fringe detection and segmentation, feature extraction and multi-frequency feature fusion on interference image data to generate multi-frequency interference fringe feature data; A spectral response analysis unit is used to perform multi-frequency spectral response analysis based on multi-frequency interference fringe feature data, establish a dynamic nanoparticle response model, and fuse the standardized multi-frequency interference fringe feature data with the response model to generate a comprehensive response feature vector; A fault detection unit is used to train an adaptive fault detection model using a comprehensive response feature vector, and adjust the adaptive fault detection model parameters in real time during operation through a dynamic correction mechanism, input real-time interference fringe feature data into the model for real-time fault detection and classification, and generate a fault classification result; The intelligent decision-making and feedback unit is used to generate fault handling suggestions based on the fault classification results and the intelligent decision engine, automatically generate fault reports and trigger the alarm system, and optimize the fault detection model and intelligent decision engine using fault handling records and feedback data through online learning and feedback mechanisms.

Images