Overheat hidden danger early warning method based on functional coating

Abstract

The invention relates to a functional coating-based overheating hidden danger early warning method. The method comprises the following steps of S1, acquiring image data, gas concentration data and nano-particle concentration data; s2, obtaining a color feature value based on the image data; s3, comprehensive characteristic parameters are obtained based on the gas concentration data and the nanoparticle concentration data; s4, judging the state of the equipment based on a preset color threshold system and a monitoring threshold system; s5, when the color characteristic value is lower than a color minimum early warning threshold value and the comprehensive characteristic parameter is lower than a monitoring minimum early warning threshold value, determining that the state is a normal state; otherwise, entering early warning state judgment; s6, in early warning state judgment, early warning levels are judged based on the combination of the color feature values and the comprehensive feature parameters; and S7, outputting a final monitoring report. A multi-dimensional overheating state early warning system is established through fusion analysis of color change characteristics of the functional coating and characteristics of gas and nano-particles, and the credibility of monitoring data is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of electrical equipment overheating early warning, and in particular to an overheating hidden danger early warning method based on functional coatings. Background Art

[0002] With the continuous advancement of industrial automation and intelligentization, the application scope of electrical equipment in public infrastructure continues to expand, especially in key places such as substations, distribution rooms, and data centers. The deployment of large-scale electrical equipment has become an important part of infrastructure construction. The operating status of these devices directly affects the safety and stability of social infrastructure. Among them, equipment overheating, as the most prominent safety hazard, may lead to equipment performance degradation, system failures, and even major safety accidents such as electrical fires, which will not only cause huge economic losses, but also endanger public safety. Therefore, establishing an efficient and reliable equipment overheating status monitoring and early warning mechanism has become a key technical problem that needs to be solved urgently.

[0003] At present, the two main solutions are infrared temperature measurement and thermocouple temperature measurement. Although infrared temperature measurement can be measured non-contact, the equipment cost is high, professional operation is required, and it is easily disturbed by the environment. Thermocouple temperature measurement requires the installation of sensors and cables, which is complex to deploy and easily disturbed. These traditional methods all use a single detection mechanism, which is difficult to meet the needs of complex environments. Temperature monitoring coatings have gradually become a research hotspot in the field of overheating monitoring due to their convenience and economic advantages. This technology realizes visual monitoring of temperature status through the color change of the coating, and has the characteristics of simple deployment and high cost-effectiveness. However, the existing coating technology still faces severe technical challenges in practical applications. The primary problem lies in the reliability of image acquisition and recognition: due to the dynamic changing characteristics of the lighting environment in public places, including the light intensity and color temperature fluctuations caused by factors such as day and night changes and weather changes, the recognition accuracy of the paint color change is seriously affected, which in turn affects the assessment of the risk of equipment overheating. Secondly, the existing coating monitoring systems mostly use a single color change criterion, lack a collaborative analysis mechanism for multi-dimensional data, and it is difficult to provide reliable overheating warnings in complex environments. Especially when the device temperature is close to the critical value, relying solely on color changes as a basis for judgment will not accurately assess the overheating risk status of the device. Summary of the invention

[0004] In view of this, the present invention proposes an overheating hazard early warning method based on functional coatings to solve the problem that the overheating risk assessment method in the prior art is single and has poor accuracy.

[0005] The technical solution of the present invention is implemented as follows: The present invention provides an overheating hazard early warning method based on functional coatings, comprising the following steps:

[0006] S1. Coating the surface of the electrical equipment to be monitored with a thermochromic coating and a thermovolatile coating, collecting image data of the thermochromic coating through a video monitoring device, and collecting gas concentration data released by the thermovolatile coating and nanoparticle concentration data released by the electrical equipment to be monitored through a monitoring device;

[0007] S2, extracting features from the image data to obtain color feature values;

[0008] S3, fusing the gas concentration data and the nano-particle concentration data to obtain comprehensive characteristic parameters;

[0009] S4. Determine the device status based on a preset color threshold system and a monitoring threshold system; wherein the color threshold system includes a color minimum warning threshold, a second color threshold and a first color threshold, and the monitoring threshold system includes a monitoring minimum warning threshold and a monitoring threshold;

[0010] S5. When the color feature value is lower than the color minimum warning threshold and the comprehensive feature parameter is lower than the monitoring minimum warning threshold, it is determined to be in a normal state; otherwise, it enters the warning state determination;

[0011] S6. In the early warning state determination, the early warning level is determined based on the combination of the color feature value and the comprehensive feature parameter; when the two determination results are inconsistent, the double verification mechanism is triggered to improve the determination reliability through adaptive cyclic sampling;

[0012] S7. Output the final monitoring report.

[0013] On the basis of the above technical solution, preferably, step S2 specifically includes the following steps:

[0014] S21, collecting RGB images of the thermochromic coating;

[0015] S22, extracting HSV color space features of the image;

[0016] S23, calculating the hue (H), saturation (S) and value (V) values ​​in the HSV space;

[0017] S24. Based on a preset color change characteristic curve, map the HSV value to a standardized color characteristic value.

[0018] Based on the above technical solution, preferably, step S21 further includes preprocessing the RGB image:

[0019] S211, normalize and extract features of the RGB image to obtain a feature map F conv ;

[0020] S212, extracting the feature map F convPerform nonlinear feature transformation to obtain the transformed feature map F trans ;

[0021] S213, the transformed feature map F trans Perform adaptive color correction to obtain the corrected image Icorr;

[0022] S214, corrected image I corr Structural similarity optimization is performed to obtain the final preprocessed RGB image.

[0023] Based on the above technical solution, preferably, the adaptive color correction in step S213 specifically includes: trans Convert to HSV color space for adaptive color correction to obtain the corrected HSV image, and then convert the corrected HSV image back to RGB space to obtain the corrected image I corr .

[0024] On the basis of the above technical solution, preferably, the adaptive color correction is processed by an adaptive color correction formula, and the adaptive color correction formula is:

[0025]

[0026] Among them, I HSV is the representation of the input image in the HSV color space; λ is a global adjustment parameter used to control the correction strength; α is an adaptive parameter used to adjust the steepness of the sigmoid function; is the sigmoid function, used to generate adaptive weights; log(1+|F trans |) is a logarithmic function used to smooth the correction effect.

[0027] On the basis of the above technical solution, preferably,

[0028] CF=w1·f(ΔH')+w2·g(ΔS')+w3·h(ΔV')

[0029] Among them, f(x), g(x), h(x) are nonlinear mapping functions of each component, w1, w2, w3 are weight coefficients, and they satisfy w1+w2+w3=1, and the value range of each weight coefficient is [0,1].

[0030] On the basis of the above technical solution, preferably, step S3 specifically includes the following steps:

[0031] S31, perform standardization preprocessing on the gas concentration data and nano-particle concentration data at each time point to obtain the gas concentration standardization feature F gas and nanoparticle concentration normalization characteristic Fparticle ;

[0032] S32, based on gas concentration normalization feature F gas and nanoparticle concentration normalization characteristic F particle Calculate and construct time series and calculate the rate of change of features respectively;

[0033] S33, calculating the mutual information coefficient and the synergistic variation coefficient by using the gas concentration standardized characteristics and the nano-particle concentration standardized characteristics and their change rates;

[0034] S34. Through the mutual information coefficient and the co-variation index, combined with the Euclidean norm of the standardized features, the fusion feature parameters are obtained through nonlinear mapping.

[0035] Based on the above technical solution, preferably, the calculation formula of the fusion feature parameter in step S34 is as follows:

[0036]

[0037] Among them, P is the fusion feature parameter, I is the mutual information coefficient, the value range is [0,1], S is the coordinated change index, the value range is [0,1], F gas is the gas concentration normalized characteristic, F particle Normalize features for nanoparticle concentration.

[0038] On the basis of the above technical solution, preferably, the specific steps of the double verification mechanism include:

[0039] When the color feature value determination result is inconsistent with the comprehensive feature parameter determination result, return to step S1 to re-collect data and perform analysis;

[0040] Carry out data collection and analysis three times in succession. If two or more of the three determination results are consistent, the majority determination result is adopted; if the three determination results are inconsistent, conduct equipment error troubleshooting.

[0041] On the basis of the above technical solution, preferably, the equipment state includes a normal state and a warning state, and the warning state is divided into a first-level warning, a second-level warning and a third-level warning, wherein:

[0042] Normal state: color feature value < color minimum warning threshold, and comprehensive feature parameter < monitoring minimum warning threshold;

[0043] Level 1 warning: Color feature value ≥ first color threshold, or comprehensive feature parameter ≥ monitoring threshold, indicating extreme overheating;

[0044] Level 2 warning: The second color threshold ≤ color feature value < first color threshold, and 0.7× monitoring threshold ≤ comprehensive feature parameter < monitoring threshold, indicating moderate overheating;

[0045] Level 3 warning: The lowest color warning threshold is less than the color characteristic value and less than the second color threshold, and the lowest monitoring warning threshold is less than the comprehensive characteristic parameter and less than 0.7× the monitoring threshold, indicating mild overheating.

[0046] Compared with the prior art, the overheating hazard warning method based on functional coatings of the present invention has the following advantages:

[0047] Beneficial effects:

[0048] (1) The present invention provides an overheating hazard warning method based on functional coatings. By coating the surface of electrical equipment with thermochromic coatings and thermovolatile coatings, combined with video monitoring and gas concentration monitoring, real-time monitoring and warning of the overheating state of the equipment can be achieved. The fusion judgment method uses the dual signals of color change and gas release to more comprehensively reflect the thermal state of the equipment, avoiding the misjudgment that may be caused by a single signal, and can provide more accurate and reliable overheating warnings in complex environments, significantly improving the safety and stability of equipment operation;

[0049] (2) In the image preprocessing stage, an adaptive color correction algorithm based on the HSV color space is used. Through nonlinear feature transformation and adaptive weight adjustment, the influence of ambient light changes on the color recognition of thermochromic paint is effectively overcome, and the accuracy of color feature extraction under different lighting conditions is significantly improved, making the recognition of paint color changes more stable and reliable.

[0050] (3) When the color feature value is inconsistent with the judgment result of the comprehensive feature parameter, the double verification mechanism is triggered, and multiple data collection and analysis are performed through adaptive cyclic sampling to ensure the reliability of the judgment result, effectively reduce the misjudgment rate, ensure the stability and accuracy of the early warning system in complex environments, and enhance the robustness of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0052] Figure 1 It is a flow chart of the overheating hazard early warning method based on functional coatings of the present invention;

[0053] Figure 2It is a flow chart of the early warning method of the double verification mechanism in the present invention. DETAILED DESCRIPTION

[0054] The following will be combined with the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0055] The technical solution of the present invention is achieved as follows: Figure 1 As shown, the present invention provides an overheating hazard early warning method based on functional coatings, comprising the following steps:

[0056] S1. Coating the surface of the electrical equipment to be monitored with a thermochromic coating and a thermovolatile coating, collecting image data of the thermochromic coating through a video monitoring device, and collecting gas concentration data released by the thermovolatile coating and nanoparticle concentration data released by the electrical equipment to be monitored through a monitoring device;

[0057] S2, extracting features from the image data to obtain color feature values;

[0058] S3, fusing the gas concentration data and the nano-particle concentration data to obtain comprehensive characteristic parameters;

[0059] S4. Determine the device status based on a preset color threshold system and a monitoring threshold system; wherein the color threshold system includes a color minimum warning threshold, a second color threshold and a first color threshold, and the monitoring threshold system includes a monitoring minimum warning threshold and a monitoring threshold;

[0060] S5. When the color feature value is lower than the color minimum warning threshold and the comprehensive feature parameter is lower than the monitoring minimum warning threshold, it is determined to be in a normal state; otherwise, it enters the warning state determination;

[0061] S6. In the early warning state determination, the early warning level is determined based on the combination of the color feature value and the comprehensive feature parameter; when the two determination results are inconsistent, the double verification mechanism is triggered to improve the determination reliability through adaptive cyclic sampling;

[0062] S7. Output the final monitoring report.

[0063] The present invention provides an overheating hazard warning method based on functional coatings. By coating the surface of electrical equipment with thermochromic coatings and thermovolatile coatings, combined with video monitoring and gas concentration monitoring, real-time monitoring and warning of the overheating state of the equipment can be achieved. The fusion judgment method uses the dual signals of color change and gas release to more comprehensively reflect the thermal state of the equipment, avoid misjudgment that may be caused by a single signal, and provide more accurate and reliable overheating warnings in complex environments, significantly improving the safety and stability of equipment operation.

[0064] Specifically, in step S1, the temperature-sensitive pigment (such as a color-changing agent) used in the thermochromic coating presents different molecular arrangement states at different temperatures. When the temperature reaches a specific threshold, the spatial configuration of the molecules changes, causing the optical properties of the coating to change, thereby presenting different colors, and then the color change is identified through video monitoring to warn of equipment overheating. For thermochromic volatile coatings, when the surface temperature of the coating exceeds a certain threshold (such as 80-120°C), the thermosensitive volatile additive will decompose and release characteristic gases, which are monitored by gas sensors; at the same time, the electrical equipment to be monitored (such as wires and cables or electronic parts) itself will release extremely small invisible thermal decomposition particles, i.e., micro-nano particles, when it is overheated. The disparity in the number of micro-nano particles under normal conditions and in the case of early overheating of the equipment can be monitored by particle sensors, thereby realizing the identification and warning of overheating hazards.

[0065] In one embodiment of the present invention, the basic composition of the thermochromic coating includes: (1) polymer matrix: the polymer matrix is ​​used to provide heat resistance and adhesion, such as 25-35wt.% polyurethane resin and / or 20-30wt.% acrylic resin; (2) color-changing agent: the color-changing agent is used to respond to temperature, such as oligomer liquid crystal (such as silicone liquid crystal) or reversible color-changing pigment, and the added amount is 15-25wt.%; (3) auxiliary agent: 10-15wt.% solvent (such as ethanol or ethyl acetate), 1-3wt.% dispersant (such as polyvinyl alcohol or polyacrylic acid), 1-5wt.% thickener (such as hydrogenated vegetable oil or amide thickener), 1-2% anti-settling agent (such as silicon dioxide, etc.), 1-3% gelling agent (such as polyvinyl alcohol).

[0066] The preparation method of the thermochromic coating is as follows:

[0067] (1) Raw material preparation: Accurately weigh the polymer, color-changing agent and additives according to the formula to ensure the quality and ratio of the raw materials.

[0068] (2) Temperature Matrix preparation: Mix the polyurethane resin and acrylic resin in proportion, heat to 50-60° C., and continue stirring for 30 minutes to ensure that they are fully mixed.

[0069] (3) Adding color-changing agent: gradually add oligomer liquid crystal or temperature-sensitive pigment to the resin mixture, and keep stirring to ensure that the pigment is evenly dispersed. The stirring time is 20-30 minutes until the pigment is completely dispersed.

[0070] (4) Adding additives: Add solvents and dispersants to the mixture and stir gently to reduce viscosity and ensure uniformity. Other additives are also added to adjust the properties of the coating.

[0071] (5) Adjust viscosity: Add thickeners and gelling agents as needed to adjust the fluidity of the paint to ensure it is suitable for spraying or brushing processes.

[0072] In one embodiment of the present invention, the basic composition of the thermosensitive volatile coating includes: (1) 25-40wt.% resin matrix: polystyrene (PS) or polyurethane (PU), which provides good adhesion and mechanical strength. Polystyrene can improve thermal stability, while polyurethane can enhance flexibility; (2) 30-50% volatile solvent low boiling point organic solvent, which can evaporate quickly when the coating dries, such as ethanol, isopropanol or ether solvent; (3) thermosensitive volatile additives: A, 5-10% characteristic gas release agent: select compounds that can dissociate and release gas at a certain temperature, such as certain organic fragrances or nitrogen compounds (such as organic fragrances or triethanolamine); B, micro-particles: 5-15% functional nanoparticles, which may form tiny bubbles or particles at high temperatures, such as nano-silicon dioxide or nano-alumina; (4) additives: 1-5% thickener, used to adjust the viscosity of the coating, such as polyurethane thickener, 1-3% anti-settling agent, used to prevent the coating components from stratifying, such as silicon dioxide.

[0073] The preparation method of the temperature-sensitive volatile coating comprises:

[0074] (1) Raw material preparation: Accurately weigh all raw materials according to the above proportions to ensure uniformity and consistency.

[0075] (2) Temperature Matrix preparation: Mix the polyurethane resin and acrylic resin in proportion, heat to 50-60° C., and continue stirring for 30 minutes to ensure that they are fully mixed.

[0076] (3) Resin mixing: Under moderate heating conditions (40-60°C), place the selected resin (polystyrene or polyurethane) into a stirrer and stir until completely melted.

[0077] (4) Add volatile solvent and microparticles: Slowly add low boiling point organic solvent to the resin and stir evenly. Then add functional nanoparticles to ensure uniform dispersion.

[0078] (5) Adding heat-sensitive additives: Next, add the characteristic gas release agent and continue stirring until uniform.

[0079] (6) Adding additives: gradually add thickening agents and anti-settling agents to ensure that the final coating has a suitable consistency when applied.

[0080] The application method of the paint is as follows:

[0081] (1) Coating: Spray, brush or dip the color-changing paint evenly on the target surface.

[0082] (2) Curing: The coating is cured at room temperature or according to the requirements of the resin, and heating (such as 80-120°C) may be required to accelerate the curing reaction.

[0083] On the basis of the above technical solution, preferably, step S2 specifically includes the following steps:

[0084] S21, collecting RGB images of the thermochromic coating;

[0085] In one embodiment of the present invention, an industrial-grade camera is used to collect real-time images of the surface of the device coated with the thermochromic coating. The resolution of the camera is not less than 1920×1080 pixels, and the frame rate is not less than 30fps, to ensure that the image quality meets the analysis requirements. During the acquisition process, the camera maintains a fixed distance (preferably 0.5-2m) from the monitored surface, and the ambient light conditions are kept as stable as possible.

[0086] In one embodiment, step S21 further includes preprocessing the RGB image:

[0087] S211, normalize and extract features of the RGB image to obtain a feature map F conv ;

[0088] Specifically, the linear normalization method is used to normalize the pixel values ​​of the RGB image, and the pixel values ​​of the three RGB channels are mapped to the [0,1] interval to eliminate the image brightness difference caused by the change of light intensity. Subsequently, the image features are extracted through a multi-scale convolutional neural network. The network structure adopts a VGG-like architecture, which contains 5 convolution blocks. Each convolution block consists of two 3×3 convolution layers and a maximum pooling layer. Through layer-by-layer feature extraction, a feature map F containing multi-scale feature information such as texture and edge is obtained. conv , providing rich feature representation for subsequent processing.

[0089] S212, extracting the feature map F conv Perform nonlinear feature transformation to obtain the transformed feature map F trans ;

[0090] For the feature map F conv Implement nonlinear transformation and use improved ReLU activation function for feature mapping. The process can be expressed as:

[0091] F trans =max(0,F conv )+α·min(0,F conv )

[0092] Where α is a learnable parameter with a value range of [0,1]. This nonlinear transformation can effectively enhance the expressiveness of features and improve the distinguishability of features, while retaining negative information and avoiding information loss.

[0093] S213, the transformed feature map F trans Perform adaptive color correction to obtain the corrected image I corr ;

[0094] Based on the above technical solution, preferably, the adaptive color correction in step S213 specifically includes: trans Convert to HSV color space for adaptive color correction to obtain the corrected HSV image, and then convert the corrected HSV image back to RGB space to obtain the corrected image I corr .

[0095] The feature map F trans Convert to HSV color space for color correction. The advantage of HSV space is that it can separate color information from brightness information, which allows more precise adjustment of hue and saturation without affecting brightness when performing color correction. This feature of HSV space makes it an ideal choice for color correction. In HSV space, an adaptive color correction algorithm is applied to generate a corrected HSV image to adapt to different lighting conditions and environmental changes. This process ensures the color consistency of the image in different environments by adjusting hue, saturation, and brightness. The adaptive correction algorithm dynamically adjusts the correction parameters according to the local and global features of the image to achieve the best correction effect. The corrected HSV image is then converted back to RGB space to obtain the corrected image I corr , so that subsequent structural similarity optimization is performed in RGB space. Images in RGB space are more suitable for visual display and further image processing.

[0096] Among them, the adaptive color correction is processed by the adaptive color correction formula, and the adaptive color correction formula is:

[0097]

[0098] Among them, I HSVis the representation of the input image in the HSV color space; λ is a global adjustment parameter used to control the correction strength. By adjusting λ, the correction amplitude can be changed so that the corrected image is more consistent with the expected color performance; α is an adaptive parameter used to adjust the steepness of the sigmoid function. The sigmoid function is used here to generate adaptive weights to help smooth the correction effect. By adjusting α, the sensitivity of the correction, that is, the response speed to different color changes, can be controlled; is a sigmoid function, which is used to generate adaptive weights, so that the correction process can be dynamically adjusted according to the characteristics of the image, ensuring that different correction strengths are applied in different image regions; log(1+|F trans |) is a logarithmic function, which is used to smooth the correction effect. The introduction of the logarithmic function can prevent excessive changes in the correction process and ensure that the corrected image remains visually natural.

[0099] The adaptive color correction method of the present invention can dynamically adjust the color performance of the image under different lighting conditions and environmental changes, ensuring that the corrected image is more accurate and consistent in color. The adaptive color correction method can automatically adjust the correction intensity according to the changes in image features, which not only ensures the accuracy of the correction effect, but also avoids distortion caused by over-correction, providing a reliable data basis for subsequent color feature extraction.

[0100] S214, corrected image I corr Structural similarity optimization is performed to obtain the final preprocessed RGB image.

[0101] Specifically, firstly, based on the structural similarity theory, the corrected image I corr Perform structural similarity evaluation. The structural similarity evaluation uses the SSIM index, which comprehensively evaluates the image quality from three dimensions: brightness, contrast, and structure. The calculation formula of the SSIM index is:

[0102] SSIM(x,y)=[l(x,y) α ]·[c(x,y) β ]·[s(x,y) γ ]

[0103] Among them, l(x, y) represents the brightness comparison function, c(x, y) represents the contrast comparison function, s(x, y) represents the structure comparison function, and α, β, and γ are weight coefficients used to adjust the relative importance of the three components.

[0104] When performing structural similarity optimization, the system first divides the image into multiple overlapping local windows and calculates the SSIM value for each window. By sliding the window, the SSIM distribution map of the entire image is obtained. Based on the SSIM distribution map, the system adopts an iterative optimization strategy to improve the overall structural similarity by adjusting the local features of the image. The optimization process is carried out in an iterative manner. In each iteration, the SSIM map of the current image is calculated, and the areas to be optimized are determined based on the SSIM map. These areas are adaptively enhanced, and structural preservation and global consistency constraints are applied. The optimization effect is evaluated, and the iteration is stopped if the termination condition is met. The termination condition can be that the preset maximum number of iterations is reached, or the increase in the SSIM value is less than the preset threshold. Through this iterative optimization method, the system can effectively improve the visual quality of the image while maintaining the key structural features of the image. The final output preprocessed RGB image not only retains the important information of the original image, but also has better visual effects and feature expression capabilities.

[0105] In the present invention, in the preprocessing stage, the RGB image is converted into the HSV space for color correction to ensure the color consistency of the image under different lighting conditions. In the feature extraction stage (S22), the hue, saturation and brightness based on the HSV space directly reflect the physical properties of the color, which is convenient for temperature-related feature analysis, and the RGB image is converted to the HSV space again for color feature extraction. Through preprocessing, the quality and consistency of the image are significantly improved, providing a reliable data basis for subsequent color feature extraction. This processing flow ensures the robustness and accuracy of the system under different environmental conditions. Through the above-mentioned preprocessing steps, the system can more accurately identify and quantify the color changes of thermochromic coatings, and provide reliable data support for the determination of the overheating state of the equipment. This multi-stage image preprocessing method not only improves the accuracy of color feature extraction, but also enhances the adaptability and reliability of the system in complex environments.

[0106] S22, extracting HSV color space features of the image;

[0107] Specifically, the collected RGB image is converted to the HSV color space, and then the HSV color space features are extracted. Compared with the RGB space, the HSV space is closer to the way the human eye perceives color and has better robustness to changes in lighting. The conversion process uses the standard RGB to HSV conversion algorithm, which specifically includes: normalizing the RGB values ​​to the [0,1] interval, calculating the maximum value (max) and minimum value (min), determining the H component based on the relationship between max and min, calculating the S component: S = (max-min) / max, and directly taking max as the V component.

[0108] S23, calculating the hue (H), saturation (S) and value (V) values ​​in the HSV space;

[0109] Specifically, the converted HSV image is segmented and the HSV value of the region of interest (ROI) is extracted, that is, the ROI area is determined, and the central area of ​​the coating surface is usually selected; the average value of each component in the ROI area is calculated, where the H value ranges from 0° to 360°, the S value ranges from 0 to 100%, and the V value ranges from 0 to 100%; the calculation results are statistically filtered to remove the influence of outliers.

[0110] S24. Based on a preset color change characteristic curve, map the HSV value to a standardized color characteristic value.

[0111] Specifically, the standard HSV values ​​of the coating at different temperatures are obtained through experiments, and the temperature-color correspondence curve is established; the measured HSV value is compared with the standard curve, and the HSV value is converted into a standardized color feature value using a nonlinear mapping function. The mapping function considers the weight influence of each component, and finally outputs the final color feature value for subsequent state judgment. Among them, the calculation formula of the color feature value is:

[0112] CF=w1·f(ΔH′)+w2·g(ΔS′)+w3·h(ΔV′)

[0113] ΔH'=|H 实测 -H 标准 | / 360°; ΔS'=|S 实测 -S 标准 | / 100%; ΔV'=|V 实测 -V 标准 | / 100%, where H 实测 , S 实测 、V 实测 is the HSV image measurement value at the current temperature, H 标准 , S 标准 、V 标准 is the corresponding HSV image standard value at the current temperature. Through normalization, the difference is in the interval [0,1]. f(x), g(x), h(x) are the nonlinear mapping functions of each component, where f(x) = 1-exp(-α H ·x), for nonlinear mapping of hue difference; g(x) = 1-exp(-α S ·x), nonlinear mapping for saturation difference; h(x) = 1-exp(-α V x), for nonlinear mapping of brightness differences, α H , α S , α Vis an adjustable attenuation coefficient with a value range of [2,5], which is used to adjust the steepness of the mapping curve. w1, w2, and w3 are weight coefficients, and satisfy w1+w2+w3=1. The value range of each weight coefficient is [0,1].

[0114] The value range of the color feature value is [0,1]. The closer the color feature value is to 0, the closer it is to the standard color. The closer the color feature value is to 1, the greater the difference from the standard color. By calculating the color feature value, the degree of color change of the paint can be comprehensively reflected, providing a reliable quantitative basis for subsequent state judgment.

[0115] In one embodiment of the present invention, step S3 specifically includes the following steps:

[0116] S31, performing standardization preprocessing on the gas concentration data and nano-particle concentration data at each time point to obtain the standardized characteristics of the gas concentration and the standardized characteristics of the nano-particle concentration. Specifically, performing minimum-maximum standardization processing on the collected gas concentration data and nano-particle data, and eliminating the dimensional differences of different data sources through standardization preprocessing;

[0117] The calculation formula for standardized features is

[0118]

[0119] Among them, F gas is the gas concentration normalized characteristic, F particle is the normalized characteristic of nanoparticle concentration, C gas Original gas concentration data, C particle is the original nanoparticle concentration data, is the historical minimum value of gas concentration, is the historical maximum value of gas concentration, This is the historical minimum value of nanoparticles. This is the historical maximum value of nanoparticles.

[0120] S32, constructing a time series based on the standardized characteristics of gas concentration and the standardized characteristics of nano-particle concentration, and calculating the change rate of the characteristics respectively;

[0121] Specifically, the change between every two adjacent time points is calculated and divided by the time interval between the two time points to obtain the change rate of the gas concentration and the nanoparticle concentration. For example, if the standardized gas concentration characteristic at the current moment is 0.8, the previous moment is 0.6, and the sampling time interval is 2 seconds, then the change rate of the gas concentration is 0.1 per second. By calculating the change rate, the dynamic change characteristics of the data can be effectively reflected.

[0122] The rate of change is calculated as follows:

[0123]

[0124]

[0125] in, is the normalized characteristic of the gas concentration at the current moment, is the normalized characteristic of the gas concentration at the previous moment, is the normalized characteristic of the nanoparticle concentration at the current moment, is the normalized characteristic of the nanoparticle concentration at the previous moment, Δt is the sampling time interval, ΔF gas is the rate of change of gas concentration, ΔF particle is the change rate of nanoparticle concentration.

[0126] S33, calculating the mutual information coefficient and the synergistic variation coefficient through the gas concentration standardized characteristic, the nano-particle concentration standardized characteristic, the gas concentration standardized characteristic change rate and the nano-particle concentration standardized characteristic change rate;

[0127] Specifically, the ranges of the standardized gas concentration characteristics and nano-particle concentration characteristics are divided into several intervals, and the frequency of data occurrence in these intervals is counted to obtain their respective probability distributions. At the same time, the frequency of gas concentration characteristics and nano-particle concentration characteristics falling into different interval combinations is counted to obtain the joint probability distribution. Then, the mutual information coefficient is obtained based on the joint probability and marginal probability. The mutual information coefficient can reflect the statistical correlation between the two characteristics.

[0128] The calculation formula of mutual information coefficient is:

[0129]

[0130] in, is the joint probability distribution of gas concentration characteristics and nanoparticle concentration characteristics, is the marginal probability distribution of gas concentration characteristics, is the marginal probability distribution of the nanoparticle concentration characteristics.

[0131] Specifically, the absolute value of the difference between the rate of change of gas concentration and the rate of change of nanoparticle concentration is calculated, and the difference is negated and then exponentially calculated. When the rates of change of the two characteristics are close, their difference is close to zero, and the synergistic change coefficient obtained after exponential calculation is close to 1; conversely, when the rate of change of the two characteristics is very different, the synergistic change coefficient obtained is close to 0. The synergistic change coefficient can effectively reflect the degree of synchronization of the changes of the two characteristics. The calculation formula of the synergistic change coefficient is:

[0132]

[0133] in, is the time rate of change of gas concentration, The time rate of change of nanoparticle concentration can be numerically calculated by the finite difference method.

[0134] S34. Through the mutual information coefficient and the co-variation index, combined with the Euclidean norm of the standardized features, the fusion feature parameters are obtained through nonlinear mapping.

[0135] The calculation formula of fusion feature parameters is as follows:

[0136]

[0137] Among them, P is the fusion feature parameter, I is the mutual information coefficient, which reflects the correlation between features and has a value range of [0,1]; S is the synergistic change index, which represents the degree of synergistic change of features and has a value range of [0,1]; F gas is the gas concentration normalized characteristic, F particle Normalized characteristics for nanoparticle concentration, It is the Euclidean norm of the standardized feature, which comprehensively represents the strength of the two features.

[0138] In another embodiment of the present invention, considering the characteristics of the monitored electrical equipment itself that it will produce gas and nano-particles in an overheated state, the system adopts a two-level fusion data processing scheme, which is as follows: when the sensor detects the gas and nano-particles generated by the equipment at the same time, the two types of data are firstly standardized and pre-processed, and then the above-mentioned fusion characteristic parameter calculation formula is used for the first-level fusion, that is, the gas concentration data and the nano-particle concentration data generated by the equipment are fused, and the fused equipment comprehensive characteristic parameters are output. Then, the equipment comprehensive characteristic parameters obtained in the first level are used as new nano-particle concentration characteristic parameters and the gas concentration characteristics volatilized by the thermochromic coating for the second level fusion, and the final fusion characteristic parameters are output.

[0139] Through this multi-source data fusion method, the system can simultaneously consider the changing characteristics of gas concentration and nanoparticle concentration, providing a more comprehensive and reliable basis for overheating state assessment. The fusion feature parameter P not only reflects the strength of each feature, but also considers the correlation and coordinated change characteristics between features, providing reliable data support for subsequent warning state judgment.

[0140] Specifically, in step S4, the color threshold system includes three levels of thresholds: the lowest color warning threshold: as the dividing point between the normal state and the warning state, when the color characteristic value is lower than the threshold, it indicates that the device temperature is within a safe range; the second color threshold: as the dividing point between mild overheating and moderate overheating, when the color characteristic value reaches or exceeds the threshold but does not reach the first color threshold, it indicates that the device may be at risk of moderate overheating; the first color threshold: as the highest level of warning threshold, when the color characteristic value reaches or exceeds the threshold, it indicates that the device is in an extremely overheated state and immediate measures need to be taken.

[0141] The monitoring threshold system includes two levels of thresholds: the minimum warning threshold: as the benchmark warning value for gas concentration and nanoparticle concentration, when the comprehensive characteristic parameters are lower than this threshold, it indicates that the equipment is operating normally; the monitoring threshold: as the highest level of monitoring and warning value, when the comprehensive characteristic parameters reach or exceed this threshold, it indicates that the equipment is at serious risk of overheating.

[0142] Considering that the monitored electrical equipment may have internal overheating but may not be transmitted to the temperature-sensitive coating outside the equipment, or the external temperature-sensitive coating may fail, or there may be equipment monitoring errors, a double verification mechanism is set up to further improve the accuracy of the early warning. Figure 2 As shown, when the color feature value determination result is inconsistent with the comprehensive feature parameter determination result, the double verification mechanism is triggered, and its specific steps include:

[0143] When the color feature value determination result is inconsistent with the comprehensive feature parameter determination result, return to step S1 to re-collect data and perform analysis;

[0144] Data collection and analysis are performed three times in a row. If two or more of the three judgment results are consistent, the majority judgment result is used. Among them, when the three measured color characteristic values ​​are all less than the color minimum warning threshold, and two or more comprehensive characteristic parameters are ≥ the monitoring minimum warning threshold, it may be that the heat has not been transferred to the surface, but abnormal gas and / or nanoparticles have been generated. The warning level is determined according to the comprehensive characteristic parameter value, and the warning level is output and marked as internal overheating of the equipment. Or, when the three measured color characteristic values ​​are all less than the color minimum warning threshold, and only one comprehensive characteristic parameter is ≥ the monitoring minimum warning threshold, further observation is required, and data collection and analysis are performed again. In other cases, the majority judgment result is used.

[0145] If the three judgment results are inconsistent, the system automatically triggers the equipment troubleshooting program to perform equipment error troubleshooting, which includes checking the working status of the sensor, checking the data acquisition equipment, checking the impact of environmental factors, etc.

[0146] Furthermore, the equipment status includes a normal status and a warning status, and the warning status is divided into a first-level warning, a second-level warning and a third-level warning, wherein:

[0147] Normal state: The color characteristic value is less than the color minimum warning threshold, and the comprehensive characteristic parameter is less than the monitoring minimum warning threshold. The normal state indicates that the equipment is operating normally and no warning is required.

[0148] Level 1 warning: Color feature value ≥ first color threshold, or comprehensive feature parameter ≥ monitoring threshold, indicating extreme overheating;

[0149] Level 2 warning: The second color threshold ≤ color feature value < first color threshold, and 0.7× monitoring threshold ≤ comprehensive feature parameter < monitoring threshold, indicating moderate overheating;

[0150] Level 3 warning: The lowest color warning threshold is less than the color characteristic value and less than the second color threshold, and the lowest monitoring warning threshold is less than the comprehensive characteristic parameter and less than 0.7× the monitoring threshold, indicating mild overheating.

[0151] In one embodiment of the present invention, the thermochromic paint is light yellow at room temperature, and changes from light yellow to red when heated to provide an overheating warning. The color change thresholds are divided as follows:

[0152] a. Color minimum warning threshold (60℃): The initial color is light yellow (normal temperature), the changing color is light orange (beginning to change color), HSV parameters: H (hue) is 45-50, S (saturation) is 0.6-0.7, V (brightness) is 0.8-0.9, then the color feature value is 0.3;

[0153] b. Second color threshold (80℃): the initial color is light orange, the color changes to orange-red (obvious color change), HSV parameters: H (hue) is 25-30, S (saturation) is 0.8-0.85, V (brightness) is 0.7-0.8, then the color feature value is 0.6;

[0154] c. First color threshold (100°C): The initial color is orange-red, the color changes to dark red (complete color change), HSV parameters: H (hue) is 0-5, S (saturation) is 0.9-1.0, V (brightness) is 0.6-0.7, then the color feature value is 0.9:

[0155] Based on the temperature-sensitive volatile coating, the monitoring thresholds are divided as follows:

[0156] a. Minimum warning threshold for monitoring: gas concentration is 5ppm (initial release concentration), nanoparticle concentration is 30,000 / cc, corresponding comprehensive characteristic parameter is 0.3, corresponding temperature is about 60-70℃;

[0157] b. Monitoring threshold: gas concentration is 20ppm (dangerous concentration), nanoparticle concentration is 500,000 / cc, the corresponding comprehensive characteristic parameter is 0.9, and the corresponding temperature is about 90-100℃.

[0158] When in normal state (temperature < 60°C), the color of the thermochromic coating is light yellow, the color characteristic value is < 0.3, and the visual characteristics are: the coating is uniformly light yellow, without obvious discoloration; the gas concentration is < 5ppm and the nanoparticle concentration is < 100,000 / cc, and the comprehensive characteristic parameter is < 0.3. When the system determines that the electrical equipment is in normal state, the system performs routine data collection and status judgment every 5 minutes, collects RGB images every 5 minutes, collects gas and nanoparticle data every 1 minute, and takes the average value within 5 minutes for status judgment.

[0159] When in the third-level warning state (temperature 60-80℃), the color of the thermochromic coating is light yellow → light orange → orange, and the color characteristic value is 0.3-0.6. Visual characteristics: the coating begins to show a light orange hue, and the color change area is ≤30%; the gas concentration is 5-14ppm, the nanoparticle concentration is 100,000-400,000 pieces / cc, and the comprehensive characteristic parameter is 0.3-0.6. When the system determines that the electrical equipment is in the third-level warning state, the monitoring interval is shortened to once every 2 minutes, and the image and gas data are collected synchronously. The continuous monitoring lasts for 30 minutes. If the status does not deteriorate, it will be restored to a 5-minute interval.

[0160] When in the second-level warning state (temperature 80-100℃), the color of the thermochromic coating changes from orange to orange-red to red, and the color characteristic value is 0.6-0.9. Visual characteristics: the coating shows obvious orange-red, and the color change area is 30%-70%; the gas concentration is 14-20ppm, the concentration of nanoparticles is 400,000-800,000 pieces / cc, and the comprehensive characteristic parameter is 0.6-0.9. When the system determines that the electrical equipment is in the second-level warning state, the monitoring interval is further shortened to once every minute, and data collection and analysis are continuously carried out, and a trend analysis report is generated every 15 minutes.

[0161] When in the first-level warning state (temperature ≥100℃), the color of the thermochromic coating changes from red to dark red, and the color characteristic value is ≥0.9. Visual characteristics: the coating is dark red, and the color change area is ≥70%; the gas concentration is ≥20ppm, the nanoparticle concentration is ≥800,000 / cc, and the comprehensive characteristic parameter is ≥0.9. When the system determines that the electrical equipment is in the first-level warning state, the real-time monitoring mode is activated, and the monitoring interval is shortened to 30 seconds. The system continues to collect and analyze data until the situation is under control, and automatically generates a warning status report every 5 minutes.

[0162] Through this multi-level early warning judgment and double verification mechanism, the system can accurately identify the overheating status of the equipment, and improve the reliability of the judgment through multiple sampling and analysis, effectively reducing the misjudgment rate. At the same time, when the judgment results are inconsistent, the system will automatically start the verification process to ensure the accuracy of the early warning results.

[0163] In step S7, the monitoring report includes basic information, monitoring data, early warning status evaluation, monitoring records and trend analysis, wherein the basic information includes the report number and generation time, basic information of the monitoring equipment (equipment ID, type, installation location), monitoring time period (start and end time) and report generation personnel / system information; the monitoring data includes: color feature data (current color feature value, HSV parameter measured value (hue H, saturation S, lightness V), color change area percentage and color change trend chart); gas and nano-particle data (current gas concentration value, current nano-particle concentration value, comprehensive feature parameter value and concentration change trend chart); early warning status evaluation includes: current warning level determination result, determination basis description (color feature value determination result, comprehensive feature parameter determination result and double verification execution status), early warning status (normal state, third-level warning, second-level warning and first-level warning), monitoring records (data records within the monitoring period, abnormal event records, verification mechanism trigger records and equipment error troubleshooting records); trend analysis includes: temperature change trend analysis, key parameter change curve, warning level change record and potential risk assessment.

[0164] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for early warning of overheating hazards based on functional coatings, characterized in that: The following steps are involved: S1. Coating the surface of the electrical equipment to be monitored with a thermochromic coating and a thermovolatile coating, collecting image data of the thermochromic coating through a video monitoring device, and collecting gas concentration data released by the thermovolatile coating and nanoparticle concentration data released by the electrical equipment to be monitored through a monitoring device; S2, extracting features from the image data to obtain color feature values; S3, fusing the gas concentration data and the nano-particle concentration data to obtain comprehensive characteristic parameters; S4. Determine the device status based on a preset color threshold system and a monitoring threshold system; wherein the color threshold system includes a color minimum warning threshold, a second color threshold and a first color threshold, and the monitoring threshold system includes a monitoring minimum warning threshold and a monitoring threshold; S5. When the color feature value is lower than the color minimum warning threshold and the comprehensive feature parameter is lower than the monitoring minimum warning threshold, it is determined to be in a normal state; otherwise, it enters the warning state determination; S6. In the early warning state determination, the early warning level is determined based on the combination of the color feature value and the comprehensive feature parameter; when the two determination results are inconsistent, the double verification mechanism is triggered to improve the determination reliability through adaptive cyclic sampling; S7. Output the final monitoring report.

2. The overheating hazard early warning method based on functional coatings according to claim 1, characterized in that: Step S2 specifically includes the following steps: S21, collecting RGB images of the thermochromic coating; S22, extracting HSV color space features of the image; S23, calculating the hue (H), saturation (S) and value (V) values ​​in the HSV space; S24. Based on a preset color change characteristic curve, map the HSV value to a standardized color characteristic value.

3. The overheating hazard early warning method based on functional coatings according to claim 2, characterized in that: Step S21 also includes preprocessing the RGB image: S211, normalize and extract features of the RGB image to obtain a feature map F conv ; S212, extracting the feature map F conv Perform nonlinear feature transformation to obtain the transformed feature map F trans ; S213, the transformed feature map F trans Perform adaptive color correction to obtain the corrected image I corr ; S214, corrected image I corr Structural similarity optimization is performed to obtain the final preprocessed RGB image.

4. The overheating hazard early warning method based on functional coatings according to claim 3, characterized in that: The adaptive color correction in step S213 specifically includes: trans Convert to HSV color space for adaptive color correction to obtain the corrected HSV image, and then convert the corrected HSV image back to RGB space to obtain the corrected image I corr .

5. The overheating hazard early warning method based on functional coatings according to claim 4, characterized in that: Adaptive color correction is handled by the adaptive color correction formula, which is: Among them, I HSV is the representation of the input image in the HSV color space; λ is a global adjustment parameter used to control the correction strength; α is an adaptive parameter used to adjust the steepness of the sigmoid function; is the sigmoid function, used to generate adaptive weights; log(1+|F trans |) is a logarithmic function used to smooth the correction effect.

6. The overheating hazard early warning method based on functional coatings according to claim 2, characterized in that: The calculation method of the color feature value is: CF=w1·f(ΔH′)+w2·g(ΔS′)+w3·h(ΔV′) Wherein, f(x), g(x), h(x) are the nonlinear mapping functions of each component, ΔH′ is the difference between the hue value of the HSV image measured at the same temperature and the standard hue value, ΔS′ is the difference between the saturation value of the HSV image measured at the same temperature and the standard saturation value, ΔV′ is the difference between the lightness value of the HSV image measured at the same temperature and the standard lightness value, w1, w2, w3 are weight coefficients, and satisfy w1+w2+w3=1, and the value range of each weight coefficient is [0,1].

7. The overheating hazard early warning method based on functional coatings according to claim 1, characterized in that: Step S3 specifically includes the following steps: S31, performing standardization preprocessing on the gas concentration data and the nano-particle concentration data at each time point to obtain the gas concentration standardization feature and the nano-particle concentration standardization feature; S32, constructing a time series based on the standardized characteristics of gas concentration and the standardized characteristics of nano-particle concentration, and calculating the change rate of the characteristics respectively; S33, calculating the mutual information coefficient and the synergistic variation coefficient through the gas concentration standardized characteristic, the nano-particle concentration standardized characteristic, the gas concentration standardized characteristic change rate and the nano-particle concentration standardized characteristic change rate; S34. Through the mutual information coefficient and the co-variation index, combined with the Euclidean norm of the standardized features, the fusion feature parameters are obtained through nonlinear mapping.

8. The method for early warning of overheating hazards based on functional coatings according to claim 7, characterized in that: The calculation formula of the fusion feature parameter in step S34 is as follows: Among them, P is the fusion feature parameter, I is the mutual information coefficient, the value range is [0,1], S is the coordinated change index, the value range is [0,1], F gas is the gas concentration normalized characteristic, F particle Normalize features for nanoparticle concentration.

9. The overheating hazard early warning method based on functional coatings according to claim 1, characterized in that: The specific steps of the dual verification mechanism include: When the color feature value determination result is inconsistent with the comprehensive feature parameter determination result, return to step S1 to re-collect data and perform analysis; Carry out data collection and analysis three times in succession. If two or more of the three determination results are consistent, the majority determination result is adopted; if all three determination results are inconsistent, conduct equipment error troubleshooting.

10. The overheating hazard early warning method based on functional coatings according to claim 1, characterized in that: The equipment status includes a normal status and a warning status, and the warning status is divided into a first-level warning, a second-level warning and a third-level warning, wherein: Normal state: color feature value < color minimum warning threshold, and comprehensive feature parameter < monitoring minimum warning threshold; Level 1 warning: Color feature value ≥ first color threshold, or comprehensive feature parameter ≥ monitoring threshold, indicating extreme overheating; Level 2 warning: The second color threshold ≤ color feature value < first color threshold, and 0.7× monitoring threshold ≤ comprehensive feature parameter < monitoring threshold, indicating moderate overheating; Level 3 warning: The lowest color warning threshold is less than the color characteristic value and less than the second color threshold, and the lowest monitoring warning threshold is less than the comprehensive characteristic parameter and less than 0.7× the monitoring threshold, indicating mild overheating.

Images