Intelligent monitoring method and system for high-voltage fuse

Abstract

The invention relates to the technical field of fuse monitoring, in particular to an intelligent monitoring method and a preparation system for a high-voltage fuse, and the method comprises the steps: obtaining a spectral image and color change according to a temperature liquid crystal material coated on the surface of a fuse melt, and receiving a sound wave signal through piezoelectric ceramic pieces integrated at the two ends of an acoustic waveguide structure implanted into the melt; and establishing a color recognition algorithm and an acoustic signal analysis model, inverting the temperature and fault diagnosis of the fuse, monitoring mechanical stress accumulation, predicting the residual life, and displaying the temperature, acoustic frequency spectrum and residual life parameters of the fuse in real time. According to the method, multi-dimensional information such as temperature, sound waves and mechanical stress is fused, through cooperative verification of color recognition and sound wave signal analysis, environmental interference such as electromagnetic noise and illumination fluctuation is effectively resisted, the false alarm rate is reduced, and accurate judgment of fault types is achieved.

Description

Technical Field

[0001] This invention belongs to the field of fuse monitoring technology, specifically relating to a method and system for intelligent monitoring of high-voltage fuses. Background Technology

[0002] High-voltage fuses are critical protective components in power systems, and their operating status directly affects the safety and stability of the power grid. Existing fuse monitoring technologies have many limitations: traditional contact sensors need to be connected to the circuit, which may not only interfere with the normal operation of the fuse but also easily cause insulation failure due to the high-voltage environment, and are complex to install and maintain; non-contact monitoring mostly relies on a single physical quantity, such as temperature or vibration, and is significantly affected by environmental interference, such as electromagnetic noise and changes in light, which can easily lead to false alarms or missed alarms.

[0003] Existing technologies struggle to accurately assess the accumulated mechanical stress and remaining life of fuses, often only able to respond passively after a fault occurs, failing to meet the predictive and early warning requirements of smart grids. Furthermore, some monitoring solutions rely on external power supplies, limiting their applicability in remote or harsh environments, further restricting the continuity and reliability of monitoring. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to provide a method for intelligent monitoring of high-voltage fuses.

[0005] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is: a method for intelligent monitoring of high-voltage fuses, the method comprising: acquiring spectral images and color changes based on a temperature-sensitive liquid crystal material coated on the surface of the fuse fusible element; receiving acoustic signals by piezoelectric ceramic sheets integrated at both ends of an acoustic waveguide structure embedded inside the fusible element; establishing a color recognition algorithm and an acoustic signal analysis model; reversing the temperature and fault diagnosis of the fuse; and monitoring the accumulation of mechanical stress and predicting the remaining life by using the power attenuation curve output by a thermopile structure deployed on the fuse casing and the fusible element; and displaying the fuse temperature, acoustic spectrum, and remaining life parameters in real time. The method also includes the following sub-steps:

[0006] S101, Implanting temperature-sensitive liquid crystal material, sensing material, and thermopile structure for monitoring, and completing the acquisition of spectral images and acoustic signals;

[0007] S102. The spectral image and acoustic signal are fused to obtain the fuse temperature, acoustic spectrum and fault diagnosis results.

[0008] S103. Obtain the power decay curve of the thermopile output, acquire the mechanical stress accumulation of the melt during long-term operation, and predict the remaining life.

[0009] In some embodiments, the temperature-sensitive liquid crystal material is formed on the surface of the fuse melt using plasma etching technology to create a rough structure, enhancing material adhesion. The bottom layer is a 3-5 μm thick PDMS elastomer buffer layer, and the top layer is a 1-3 μm thick temperature-sensitive liquid crystal-graphene composite film. The rough structure on the fuse melt surface is covered by spin coating. When the melt temperature exceeds a threshold, the material color changes from transparent to red, which is identified by a hyperspectral camera to obtain a spectral image. The acoustic waveguide structure is etched into a binary tree branching structure using an AlN piezoelectric thin film. The binary tree branching structure includes the following specific settings: 3-level branching, straight tube length ≥ 3 cm, and branch angle 60°. The binary tree branching structure is integrated into the melt using MEMS technology. After using a SiC substrate, piezoelectric ceramic sheets are integrated at both ends of the acoustic waveguide structure. The receiving end uses a differential amplifier circuit to acquire the acoustic signal.

[0010] In some embodiments, a Bi2Te3-based alloy thermoelectric material is used, and at least 15 pairs of thermocouples are connected in series. The cold side of the Bi2Te3-based alloy thermoelectric material is attached to the fuse housing, and the hot side is in contact with the molten material. The temperature collected by the cold side is T1, and the temperature collected by the hot side is T2. The thermoelectric potential is converted into a stable 3.3V output through a DC boost circuit to provide power.

[0011] In some implementations, the temperature of the fuse is often acquired using sensors in the prior art. However, sensor acquisition requires contact testing, which has several drawbacks, including low measurement accuracy and affecting the actual operation of the fuse and the normal operation of the connected circuit. Therefore, the temperature-sensitive liquid crystal material used in this application changes color to red when the fuse exceeds the temperature threshold, enabling rapid monitoring and judgment of whether the fuse has exceeded the temperature threshold, thus achieving rapid monitoring. However, it has another drawback, namely, the acquisition of temperature. Therefore, the establishment of the color recognition algorithm and acoustic signal analysis model includes: using a hyperspectral camera to perform Gaussian filtering on the coating area image to remove image noise, extracting the pixel color information of the coating area, and obtaining the original three primary color values ​​R of the coating area. raw,i G raw,i B raw,i According to the average formula: Where N is the number of pixels in the coating area, after obtaining the R, G, and B values, they are normalized to obtain the values ​​r, g, and b, and then converted into HSV space parameters, including: using the maximum value of r, g, and b as the value of brightness V, i.e., the objective function is V = max(r, g, b); calculating saturation, specifically... in The ratio of the minimum to the maximum value of the three components r, g, and b represents the proportion of gray components in the color. Subtracting this ratio from 1 gives the proportion of pure color components, i.e., saturation. Based on the obtained lightness V and saturation S, the hue H is obtained. A temperature calibration model is constructed using hue H, specifically T = aH. 2 +bH+c; where T is the actual temperature of the fuse, H is the hue value, and a, b, and c are fitting coefficients, determined through experimental calibration.

[0012] In some implementations, the specific determination of the fitting coefficients in the temperature calibration model includes: collecting n sets of data (H1, T1), ..., (H... n ,T n The coefficients are solved using the least squares method, specifically as follows: After solving for a, b, and c, the fitting coefficients are obtained.

[0013] In some implementations, the acoustic signal analysis model receives the original time-domain acoustic signal x(t), converts the time-domain signal into a time-domain signal x′(t) through continuous wavelet transform, obtains the wavelet entropy C, and obtains the peak amplitude A based on the maximum absolute value and the proportion of pulse components in the time-domain signal. peak kurtosis K; through Fourier transform The time-domain signal x′(t) is transformed into a frequency-domain signal X′(f). Based on the energy distribution of different frequency components in the frequency-domain signal and the frequency with the highest energy, the power spectral density P(f) and characteristic frequency f are obtained. peak The extracted features are combined into a feature vector F = [A peak ,K,f peak [C] Mutual information is used to measure the correlation between features and fault type Z. The greater the mutual information, the higher the contribution of the feature to fault identification. Let the fault type be a discrete variable Z, where Z=1 represents melt aging and Z=2 represents poor contact. The feature vector F i The mutual information correspondence function with fault type Z is, Where p(F) i (Z) is the eigenvector F i The joint probability density with Z, p(F) iLet p(Z) be the marginal probability density of the fault type, and p(Z) be the prior probability of the fault type. The joint probability density can be estimated by collecting historical data on a large number of known fault types and their corresponding feature vectors, using non-parametric or parametric methods such as kernel density estimation and parameter estimation. For example, for a batch of high-voltage fuse fault samples, the acoustic feature values ​​and fault types corresponding to each fault are recorded to construct a two-dimensional data distribution, and then the joint probability density is estimated. The marginal probability density of the response is also estimated using methods such as kernel density estimation and parameter estimation. The prior probability can be determined based on past statistical data, expert experience, or industry data. For example, based on the high-voltage fuse fault records of a company over many years, the frequency of different fault types can be statistically analyzed and used as the prior probability of the fault type; mutual information MI(F) is selected. i The optimized feature vector F is composed of features whose Z) ≥ preset threshold θ. * As input for fault type identification.

[0014] In some implementations, fault type identification is achieved by establishing a mapping relationship between feature vectors and fault types, and outputting the fault type. This includes classification based on an SVM model. The construction of the SVM model includes the following specific steps: training an SVM model based on feature vectors of known fault types; assuming the training samples are... Z i ∈{1,2,...,M}, where M is the total number of fault types, the SVM model finds the optimal hyperplane ω·F * +b=0; Fault identification; Extract optimized feature vector from newly acquired signals. Input the trained SVM model and determine the fault type using the hyperplane: in For the predicted fault type, such as 1 = melt aging, 2 = poor contact, 3 = partial discharge, etc. argmax is an operator that finds the independent variable that makes the following expression reach its maximum value. sign(·) is a sign function that returns different results based on the sign of the input value. ω is the weight vector, and b is the bias term.

[0015] In some embodiments, after the Bi2Te3-based alloy thermoelectric material is sampled at a cold surface temperature of T1 and a hot surface temperature of T2, a stress degradation formula is established based on the lattice distortion of the thermoelectric material caused by mechanical stress, including: R int (σ)=R0·(1+k2·σ 1.5 ); where R int (σ) represents the internal resistance of the thermopile, R0 represents the initial internal resistance of the fuse, and k2 represents the internal resistance growth coefficient obtained based on fracture mechanics. 1.5To conform to the Paris-Erdogan crack propagation law, the resistance is proportional to the 1.5th power of the stress. Based on the lattice distortion of thermoelectric materials caused by mechanical stress, a Seebeck coefficient α(σ) is introduced; α(σ) = α0 - k1·σ; where α0 = 220uV / K is the initial Seebeck coefficient, k1 is the stress sensitivity coefficient determined through experimental fitting, and σ is the mechanical stress; finally, a thermopile output power model considering the influence of stress is obtained, including: P = α(σ)·ΔT·IR int (σ)·I 2 Where ΔT is the temperature difference between the cold surface acquisition temperature T1 and the hot surface acquisition temperature T2, and I is the load current. After the thermopile output power model outputs the power of each cycle, it obtains a curve to describe the power decay based on the power difference between adjacent cycles.

[0016] In some implementations, after obtaining the curves describing power decay, a mapping relationship between power decay and stress is established through accelerated degradation tests. Power decay curves corresponding to different stress levels are simulated for the monitored high-voltage fuses, and the lifespan is predicted by calculating the critical stress of material failure.

[0017] Another technical problem to be solved by the present invention is to provide an intelligent monitoring system for high-voltage fuses.

[0018] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is: an intelligent monitoring system for high-voltage fuses, the system comprising: an acquisition module; used to acquire spectral images of the temperature-sensitive liquid crystal material on the surface of the fuse element through a hyperspectral camera, receive acoustic signals through piezoelectric ceramic sheets integrated at both ends of an acoustic waveguide structure, and acquire power output signals through a thermopile structure deployed on the fuse casing and the fuse element; a signal preprocessing module; used to preprocess the spectral images, acoustic signals, and power signals acquired by the acquisition module; and a fusion analysis module; including a color recognition submodule and an acoustic wave analysis submodule, wherein the color recognition submodule is based on the preprocessed HS... The V parameter is used to invert the fuse temperature through a calibration model of hue and temperature. The acoustic wave analysis submodule uses mutual information to filter and optimize the feature vector, and then outputs the fault diagnosis result through an SVM model. The life prediction module is used to calculate the mechanical stress accumulation of the fuse based on the power decay curve output by the thermopile structure and the stress degradation model. Based on the linear stress accumulation model and the critical stress threshold, it predicts the remaining life of the fuse. The display and interaction module is used to receive the temperature, acoustic wave spectrum, fault diagnosis result output by the fusion analysis module and the remaining life parameter output by the life prediction module in real time. It displays the monitoring data through a visual interface and supports the triggering and transmission of abnormal early warning signals.

[0019] The scope of this invention is not limited to technical solutions formed by specific combinations of the above-described technical features, but also includes other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents. For example, technical solutions formed by substituting the above-described features with (but not limited to) technical features with similar functions disclosed in this application.

[0020] Due to the application of the above technical solutions, this invention has the following advantages compared with the prior art: it integrates multi-dimensional information such as temperature, sound waves, and mechanical stress, and through the collaborative verification of color recognition and sound wave signal analysis, it effectively resists environmental interference such as electromagnetic noise and light fluctuations, reduces false alarm rate, and achieves accurate identification of fault types; it utilizes a thermopile structure to achieve thermoelectric power generation, which can maintain system operation without external power supply, making it particularly suitable for remote substations or harsh environments, ensuring the continuity and stability of monitoring; it uses the power decay curve to infer the accumulation of mechanical stress and combines it with material characteristic models to predict the remaining lifespan, significantly improving the initiative and economy of power grid operation and maintenance; and through non-contact sensing methods such as thermochromic materials and acoustic waveguide structures, it avoids the interference of traditional contact sensors on fuse operation, reduces the risk of insulation failure, simplifies the installation process, and is suitable for high-voltage complex environments. Attached Figure Description

[0021] Figure 1 This is a schematic diagram of the process of the present invention. Detailed Implementation

[0022] Example 1: As Figure 1The method for intelligent monitoring of high-voltage fuses, as shown, specifically includes the following steps: acquiring spectral images and color changes based on a temperature-sensitive liquid crystal material coated on the surface of the fuse fusible element; receiving acoustic signals through piezoelectric ceramic sheets integrated at both ends of an acoustic waveguide structure embedded inside the fusible element; establishing a color recognition algorithm and an acoustic signal analysis model; reversing the fuse's temperature and fault diagnosis; and monitoring mechanical stress accumulation and predicting remaining life based on the power attenuation curve output by a thermopile structure deployed on the fuse casing and fusible element; and displaying the fuse temperature, acoustic spectrum, and remaining life parameters in real time. The method also includes the following sub-steps: S101, implanting a temperature-sensitive liquid crystal material, a sensing material, and a thermopile structure for monitoring. The structure is constructed and the spectral image and acoustic signal are acquired. The temperature-sensitive liquid crystal material is based on plasma etching technology to form a rough structure on the surface of the fuse melt, which enhances the adhesion of the material. The bottom layer is a 3-5μm thick PDMS elastomer buffer layer, and the top layer is a 1-3μm thick temperature-sensitive liquid crystal-graphene composite film. The rough structure on the surface of the fuse melt is covered by spin coating. When the melt temperature exceeds the threshold, the material color changes from transparent to red. The spectral image is acquired by identifying it with a hyperspectral camera. The acoustic waveguide structure is etched into a binary tree branching structure with an AlN piezoelectric thin film. The binary tree branching structure includes the following specific settings: 3-level branching, straight tube length ≥3cm, and branch angle 60°.

[0023] The binary tree branching structure is integrated into the melt using MEMS technology. After being placed on a SiC substrate, piezoelectric ceramic sheets are integrated at both ends of the acoustic waveguide structure. The receiving end uses a differential amplifier circuit to acquire the acoustic signal.

[0024] The Bi2Te3-based alloy thermoelectric material is connected in series with at least 15 pairs of thermocouples. The cold side of the Bi2Te3-based alloy thermoelectric material is attached to the fuse shell, and the hot side is in contact with the molten material. The temperature collected by the cold side is T1, and the temperature collected by the hot side is T2. The thermoelectric potential is converted into a stable 3.3V output through a DC boost circuit to provide power.

[0025] S102. The spectral image and acoustic signal are fused to obtain the fuse temperature, acoustic spectrum, and fault diagnosis results. Furthermore, in existing technologies, fuse temperature is often obtained using sensors. However, sensor acquisition requires contact testing, which has several drawbacks, including low measurement accuracy and impacting the actual operation of the fuse and the normal operation of the connected circuit. Therefore, the temperature-sensitive liquid crystal material used in this application changes color to red when the fuse exceeds the temperature threshold, enabling rapid monitoring and judgment of whether the fuse has exceeded the temperature threshold, achieving rapid monitoring. However, it has another drawback: temperature acquisition. Therefore, the establishment of the color recognition algorithm and acoustic signal analysis model includes: using a hyperspectral camera to perform Gaussian filtering on the coating area image to remove image noise, extracting the pixel color information of the coating area, and obtaining the original three primary color values ​​R of the coating area.raw,i G raw,i B raw,i According to the average formula: Where N is the number of pixels in the coating area, after obtaining the R, G, and B values, they are normalized to obtain the values ​​r, g, and b, and then converted into HSV space parameters, including: using the maximum value of r, g, and b as the value of brightness V, i.e., the objective function is V = max(r, g, b); calculating saturation, specifically... in The ratio of the minimum to the maximum value of the three components r, g, and b represents the proportion of gray components in the color. Subtracting this ratio from 1 gives the proportion of pure color components, i.e., saturation. Based on the obtained lightness V and saturation S, the hue H is obtained. A temperature calibration model is constructed using hue H, specifically T = aH. 2 +bH+c; where T is the actual temperature of the fuse, H is the hue value, and a, b, and c are fitting coefficients, determined through experimental calibration.

[0026] The specific determination of the fitting coefficients in the temperature calibration model includes: collecting n sets of data (H1, T1), ..., (H n ,T n The coefficients are solved using the least squares method, specifically as follows: After solving for a, b, and c, the fitting coefficients are obtained.

[0027] The acoustic signal analysis model receives the original time-domain acoustic signal x(t), converts it into a time-domain signal x′(t) through continuous wavelet transform, obtains the wavelet entropy C, and obtains the peak amplitude A based on the maximum absolute value and the proportion of pulse components in the time-domain signal. peak kurtosis K; through Fourier transform The time-domain signal x′(t) is transformed into a frequency-domain signal X′(f). Based on the energy distribution of different frequency components in the frequency-domain signal and the frequency with the highest energy, the power spectral density P(f) and characteristic frequency f are obtained. peak The extracted features are combined into a feature vector F = [A peak ,K,f peak [C] Mutual information is used to measure the correlation between features and fault type Z. The greater the mutual information, the higher the contribution of the feature to fault identification. Let the fault type be a discrete variable Z, where Z=1 represents melt aging and Z=2 represents poor contact. The feature vector F i The mutual information correspondence function with fault type Z is, Where p(F) i (Z) is the eigenvector F i The joint probability density with Z, p(F) iLet p(Z) be the marginal probability density of the fault type, and p(Z) be the prior probability of the fault type. The joint probability density can be estimated by collecting historical data on a large number of known fault types and their corresponding feature vectors, using non-parametric or parametric methods such as kernel density estimation and parameter estimation. For example, for a batch of high-voltage fuse fault samples, the acoustic feature values ​​and fault types corresponding to each fault are recorded to construct a two-dimensional data distribution, and then the joint probability density is estimated. The marginal probability density of the response is also estimated using methods such as kernel density estimation and parameter estimation. The prior probability can be determined based on past statistical data, expert experience, or industry data. For example, based on the high-voltage fuse fault records of a company over many years, the frequency of different fault types can be statistically analyzed and used as the prior probability of the fault type; mutual information MI(F) is selected. i The optimized feature vector F is composed of features whose Z) ≥ preset threshold θ. * As input for fault type identification.

[0028] Fault type identification is achieved by establishing a mapping relationship between feature vectors and fault types, and outputting the fault type. This includes classification based on an SVM model. The construction of the SVM model involves the following specific steps: training the SVM model based on feature vectors of known fault types; assuming the training samples are... Z i ∈{1,2,...,M}, where M is the total number of fault types, the SVM model finds the optimal hyperplane ω·F * +b=0; Fault identification; Extract optimized feature vector from newly acquired signals. Input the trained SVM model and determine the fault type using the hyperplane: in For the predicted fault type, such as 1 = melt aging, 2 = poor contact, 3 = partial discharge, etc. argmax is an operator that finds the independent variable that makes the following expression reach its maximum value. sign(·) is a sign function that returns different results based on the sign of the input value. ω is the weight vector, and b is the bias term.

[0029] S103. Obtain the power decay curve of the thermopile output, acquire the mechanical stress accumulation of the melt during long-term operation, and predict the remaining life.

[0030] After the Bi2Te3-based alloy thermoelectric material is sampled at a cold surface temperature of T1 and a hot surface temperature of T2, a stress degradation formula is established based on the lattice distortion of the thermoelectric material caused by mechanical stress, including: R int (σ)=R0·(1+k2·σ 1.5 ); where R int(σ) represents the internal resistance of the thermopile, R0 represents the initial internal resistance of the fuse, and k2 represents the internal resistance growth coefficient obtained based on fracture mechanics. 1.5 To conform to the Paris-Erdogan crack propagation law, the resistance is proportional to the 1.5th power of the stress. Based on the lattice distortion of thermoelectric materials caused by mechanical stress, a Seebeck coefficient α(σ) is introduced; α(σ) = α0 - k1·σ; where α0 = 220uV / K is the initial Seebeck coefficient, k1 is the stress sensitivity coefficient determined through experimental fitting, and σ is the mechanical stress; finally, a thermopile output power model considering the influence of stress is obtained, including: P = α(σ)·ΔT·IR int (σ)·I 2 Where ΔT is the temperature difference between the cold surface acquisition temperature T1 and the hot surface acquisition temperature T2, and I is the load current. After the thermopile output power model outputs the power of each cycle, it obtains a curve to describe the power decay based on the power difference between adjacent cycles.

[0031] After obtaining the curves used to describe power decay, the mapping relationship between power decay and stress is established through accelerated degradation tests. Power decay curves corresponding to different stress levels are simulated for the monitored high-voltage fuses, and the life is predicted by calculating the critical stress of material failure.

[0032] This invention also provides a system for intelligent monitoring of high-voltage fuses. The system includes: a data acquisition module for acquiring spectral images of the temperature-sensitive liquid crystal material on the surface of the fuse element using a hyperspectral camera, receiving acoustic signals through piezoelectric ceramic sheets integrated at both ends of an acoustic waveguide structure, and acquiring power output signals through a thermopile structure deployed on the fuse casing and the fuse element; a signal preprocessing module for preprocessing the spectral images, acoustic signals, and power signals acquired by the acquisition module; and a fusion analysis module including a color recognition submodule and an acoustic wave analysis submodule, wherein the color recognition submodule, based on the preprocessed HSV parameters, analyzes the hue... The system includes a temperature calibration model to invert the fuse temperature, an acoustic wave analysis submodule to optimize feature vectors using mutual information filtering, and an SVM model to output fault diagnosis results. A lifespan prediction module is used to calculate the mechanical stress accumulation of the fuse body based on the power decay curve output from the thermopile structure and a stress degradation model. Based on a linear stress accumulation model and a critical stress threshold, it predicts the remaining lifespan of the fuse. A display and interaction module receives the temperature, acoustic wave spectrum, fault diagnosis results from the fusion analysis module, and the remaining lifespan parameters from the lifespan prediction module in real time. It displays the monitoring data through a visual interface and supports the triggering and transmission of abnormal early warning signals.

[0033] It should be noted that the descriptions of each embodiment in the above embodiments have different focuses. For parts that are not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0034] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied 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.

[0035] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, as well as combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0036] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0037] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0038] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0039] The above embodiments are only for illustrating the technical concept and features of the present invention, and are intended to enable those skilled in the art to understand the content of the present invention and implement it accordingly. They should not be construed as limiting the scope of protection of the present invention. All equivalent changes or modifications made in accordance with the spirit and essence of the present invention should be covered within the scope of protection of the present invention.

Claims

1. A method for intelligent monitoring of high-voltage fuses, characterized in that, The method includes: acquiring spectral images and color changes based on temperature-sensitive liquid crystal material coated on the surface of the fuse melt; receiving acoustic signals through piezoelectric ceramic sheets integrated at both ends of an acoustic waveguide structure embedded inside the melt; establishing a color recognition algorithm and an acoustic signal analysis model; reversing the fuse temperature and fault diagnosis; and monitoring mechanical stress accumulation and predicting remaining life based on the power attenuation curve output by a thermopile structure deployed on the fuse shell and melt. It also includes the following sub-steps: S101, implanting temperature-sensitive liquid crystal material, sensing material, and thermopile structure for monitoring, and acquiring spectral images and acoustic signals; S102, fusing the spectral images and acoustic signals to obtain fuse temperature, acoustic spectrum, and fault diagnosis results; S103, using the power attenuation curve output by the thermopile to obtain the mechanical stress accumulation of the melt during long-term operation and predicting remaining life.

2. The intelligent monitoring method for high-voltage fuses according to claim 1, characterized in that, The temperature-sensitive liquid crystal material is based on plasma etching technology to form a rough structure on the surface of the fuse melt. The bottom layer is a 3-5 μm thick PDMS elastomer buffer layer, and the top layer is a 1-3 μm thick temperature-sensitive liquid crystal-graphene composite film. The rough structure on the surface of the fuse melt is covered by spin coating. When the melt temperature exceeds a threshold, the material color changes from transparent to red. The spectral image is obtained by identifying the color using a hyperspectral camera. The acoustic waveguide structure is etched into a binary tree branching structure using an AlN piezoelectric thin film. The binary tree branching structure includes the following specific settings: 3-level branches, straight tube length ≥ 3 cm, and branch angle 60°. The binary tree branching structure is integrated into the melt using MEMS technology. After being placed on a SiC substrate, piezoelectric ceramic sheets are integrated at both ends of the acoustic waveguide structure. The receiving end uses a differential amplifier circuit to complete the acquisition of acoustic wave signals.

3. The intelligent monitoring method for high-voltage fuses according to claim 2, characterized in that, The thermopile structure includes: a Bi2Te3-based alloy thermoelectric material, connected in series with at least 15 pairs of thermocouples; the cold side of the Bi2Te3-based alloy thermoelectric material is attached to the fuse housing, the hot side contacts the molten metal, and the cold side collects temperature at [temperature value missing]. The temperature of the hot surface is The thermoelectric potential is converted into a stable output through a DC boost circuit to provide power.

4. The intelligent monitoring method and system for high-voltage fuses according to claim 1, characterized in that, The establishment of the color recognition algorithm and acoustic signal analysis model includes: performing Gaussian filtering on the coating area image acquired by a hyperspectral camera to remove image noise, extracting pixel color information of the coating area, and obtaining the original three primary color values ​​of the coating area. , , According to the average formula: , , ;in, The value is the number of pixels in the coating area, obtained by normalizing the R, G, and B values. , , Then, it is converted into HSV space parameters, including: , , The maximum value is used as the brightness. The value of , i.e., the objective function is ; Calculate saturation, specifically ,in for , , The ratio of the minimum to the maximum value of the three components represents the proportion of gray in the color, which is the saturation; based on the obtained lightness... and saturation After obtaining the hue ; with color tone Construct a temperature calibration model, specifically as follows: ;in The actual temperature of the fuse. For hue values, , , The fitting coefficients are determined through experimental calibration.

5. The intelligent monitoring method for high-voltage fuses according to claim 4, characterized in that, Also includes: The acoustic signal analysis model receives the original time-domain acoustic signal. The time-domain signal is converted into a time-domain signal through continuous wavelet transform. And obtain wavelet entropy The peak amplitude is obtained based on the maximum absolute value and the proportion of pulse components in the time-domain signal. , cliff Through Fourier transform Time domain signal Convert to frequency domain signal The power spectral density is obtained based on the energy distribution of different frequency components in the frequency domain signal and the frequency with the highest energy. Characteristic frequencies Combine the extracted features into a feature vector. Mutual information is used to measure characteristics and fault types. Correlation, feature vector With fault type The mutual information correspondence function is, ,in For feature vectors and The joint probability density, The marginal probability density of the feature Prior probabilities of fault types; mutual information selection. The optimized feature vector is composed of features. As input for fault type identification.

6. The intelligent monitoring method for high-voltage fuses according to claim 5, characterized in that, Fault type identification is achieved by establishing a mapping relationship between feature vectors and fault types, and outputting the fault type. This includes classification based on an SVM model. The construction of the SVM model involves the following specific steps: training the SVM model based on feature vectors of known fault types; assuming the training samples are... ,in , Given the total number of fault types, the SVM model finds the optimal hyperplane. Fault identification; Extraction of optimized feature vectors from newly acquired signals. Input the trained SVM model and determine the fault type using the hyperplane: ;in For the predicted fault type, argmax is an operator that finds the argument that maximizes the following expression. A sign function that returns different results depending on the sign of the input value. For the weight vector, This is a bias term.

7. The intelligent monitoring method for high-voltage fuses according to claim 3, characterized in that, Bi2Te3-based alloy thermoelectric materials have a cold-surface collection temperature of [temperature value missing]. The temperature of the hot surface is Subsequently, based on the lattice distortion of hotspot materials caused by mechanical stress, a stress degradation formula was established, including: ;in For the internal resistance of the thermopile, The initial internal resistance of the fuse. The internal resistance growth coefficient is obtained based on fracture mechanics. To conform to the Paris-Erdogan crack propagation law, the resistance is proportional to the 1.5th power of the stress; based on the lattice distortion of thermoelectric materials caused by mechanical stress, a Seebeck coefficient is introduced. The final thermopile output power model considering stress effects is obtained, including: ;in, The temperature of the cold surface is The temperature of the hot surface is Temperature difference, Given the load current, the thermopile output power model outputs the power for each cycle, and then obtains a curve describing the power decay based on the power difference between adjacent cycles.

8. The intelligent monitoring method for high-voltage fuses according to claim 7, characterized in that, After obtaining the curves used to describe power decay, the mapping relationship between power decay and stress is established through accelerated degradation tests. Power decay curves corresponding to different stress levels are simulated for the monitored high-voltage fuses, and the life is predicted by calculating the critical stress of material failure.

9. A system for intelligent monitoring of high-voltage fuses, applied to the intelligent monitoring method for high-voltage fuses as described in any one of claims 1-8, characterized in that, The system includes: an acquisition module for acquiring spectral images of the temperature-sensitive liquid crystal material on the surface of the fuse element using a hyperspectral camera, receiving acoustic signals through piezoelectric ceramic sheets integrated at both ends of an acoustic waveguide structure, and acquiring power output signals through a thermopile structure deployed on the fuse housing and the fuse element; a signal preprocessing module for preprocessing the spectral images, acoustic signals, and power signals acquired by the acquisition module; and a fusion analysis module including a color recognition submodule and an acoustic wave analysis submodule, wherein the color recognition submodule, based on the preprocessed HSV parameters, inverts the fuse element using a hue and temperature calibration model. The acoustic wave analysis submodule uses mutual information to filter and optimize feature vectors, and then outputs fault diagnosis results through an SVM model. The lifespan prediction module calculates the mechanical stress accumulation of the melt based on the power decay curve output from the thermopile structure, combined with a stress degradation model. Based on a linear stress accumulation model and a critical stress threshold, it predicts the remaining lifespan of the fuse. The display and interaction module receives the temperature, acoustic wave spectrum, fault diagnosis results output from the fusion analysis module, and the remaining lifespan parameters output from the lifespan prediction module in real time. It displays the monitoring data through a visual interface and supports the triggering and transmission of abnormal early warning signals.

Images