An industrial equipment operation monitoring and early warning system based on infrared temperature distribution technology

Abstract

The present invention discloses an industrial equipment operation monitoring and early warning system based on infrared temperature distribution technology, including: a temperature distribution image correction module, which calculates the compensation coefficient of heat reflection on the surface of industrial equipment and obtains a temperature distribution image after correcting the reflection effect; an image displacement correction module, which corrects the temperature distribution image offset caused by the displacement of displacement components of industrial equipment; a temperature abnormal area recognition module, which identifies the temperature abnormal area; a temperature abnormal area maintenance strategy formulation module, which formulates a maintenance strategy for the temperature abnormal area; a temperature imbalance area recognition module, which identifies the temperature imbalance area in the non-temperature abnormal area; a temperature imbalance area maintenance module, which formulates a preventive maintenance strategy for the temperature imbalance area; a maintenance strategy execution and optimization module, which continuously optimizes the maintenance strategy according to the execution results of the maintenance task. The present invention improves the efficiency and effect of industrial equipment maintenance work by combining infrared thermal imaging technology with multiple data correction and analysis methods.

Description

Technical Field

[0001] The present invention relates to the field of industrial equipment monitoring, and in particular to an industrial equipment operation monitoring and early warning system based on infrared temperature distribution technology. Background Art

[0002] In modern industrial production, with the improvement of the automation level of industrial equipment, the demand for monitoring the operating status of industrial equipment has become increasingly important. In particular, accurate monitoring and early warning of the temperature on the surface of industrial equipment are crucial to ensure the normal operation of industrial equipment and extend the service life of industrial equipment. Infrared thermal imaging technology is widely used in temperature monitoring of industrial equipment as a non-contact temperature measurement method. This technology generates temperature distribution images by capturing infrared radiation on the surface of industrial equipment to achieve real-time monitoring of the surface temperature of industrial equipment. However, infrared thermal imaging technology itself also has some inherent problems. Due to the diversity of surface materials of industrial equipment, different materials have different reflection coefficients in the infrared band, which may cause deviations in the temperature data captured by the infrared thermal imager. This deviation will affect the accuracy of temperature measurement, especially on the surface of materials with large reflection coefficients. The error will be more significant. In addition, changes in the surface roughness of industrial equipment will lead to uneven distribution of infrared radiation, thereby affecting the accuracy of infrared imaging. On the other hand, during the operation of industrial equipment, the temperature distribution on the surface of industrial equipment often changes with the movement of mechanical parts. High-speed moving parts may cause offset and blurring of thermal imaging images, which also affects the accuracy of temperature measurement. Accurately identifying temperature anomaly areas based on thermal imaging images is a key link in industrial equipment fault diagnosis and early warning, while traditional temperature monitoring often determines whether the equipment has temperature anomalies by setting simple temperature thresholds. However, the stress concentration area of ​​the equipment will change dynamically with the change of working conditions, which directly affects the temperature distribution. The existing temperature anomaly detection method is difficult to combine the stress concentration characteristics of the equipment for comprehensive analysis, resulting in insufficient recognition efficiency and accuracy of temperature anomaly areas, and prone to false alarms or missed reports. How to accurately identify temperature anomaly areas by combining the working status of the equipment and the changes in stress concentration areas is one of the difficulties of current technology. In addition, temperature imbalance is often an early signal of potential failure of industrial equipment, but due to the complexity of temperature imbalance detection, it is difficult for existing technologies to effectively monitor and warn non-temperature anomaly areas, which makes industrial equipment lack effective preventive maintenance measures before temperature anomalies occur, which easily leads to industrial equipment failures. Therefore, there are still many problems that need to be solved in temperature monitoring and fault early warning of industrial equipment. How to improve the accuracy of temperature measurement, reduce measurement errors caused by reflection effects and surface roughness, how to accurately capture temperature data in a high-speed motion environment, how to effectively identify abnormal temperature areas and achieve accurate fault diagnosis, and how to take preventive measures in a timely manner when temperature imbalance occurs, these are all key issues that need to be urgently addressed by current technology. Summary of the invention

[0003] The present invention solves the problems existing in the above-mentioned prior art and provides an industrial equipment operation monitoring and early warning system based on infrared temperature distribution technology, which mainly includes:

[0004] The temperature distribution image correction module is used to obtain the temperature distribution image of the industrial equipment through an infrared thermal imager, and determine the compensation coefficient of the heat reflection on the surface of the industrial equipment based on the reflection coefficient of different surface materials of the industrial equipment in the infrared band, so as to obtain the temperature distribution image after the reflection effect is corrected;

[0005] An image displacement correction module is used to obtain displacement data of displacement components of industrial equipment, and to correct the temperature distribution image offset caused by the displacement by comparing the pixel displacement in the temperature distribution image with the displacement data;

[0006] The temperature abnormality region identification module is used to identify potential temperature abnormality regions by setting a temperature threshold, predict regional characteristic data of stress concentration regions of industrial equipment under the current working state based on the equipment type and working state of the industrial equipment, and identify the temperature abnormality region based on the similarity between the regional characteristic data of the stress concentration region and the regional characteristic data of the potential temperature abnormality region;

[0007] The maintenance strategy formulation module for abnormal temperature areas is used to obtain the regional characteristic data of historical abnormal temperature areas, build a fault diagnosis model for abnormal temperature areas, determine the fault type and fault severity of abnormal temperature areas, and formulate maintenance strategies for abnormal temperature areas;

[0008] A temperature imbalance region identification module is used to monitor the temperature change rate of the non-temperature abnormal region and identify the temperature imbalance region in the non-temperature abnormal region;

[0009] The temperature imbalance area maintenance module is used to predict the time point and location where the temperature in the temperature imbalance area is greater than the set temperature threshold based on the historical temperature monitoring data of the temperature imbalance area, and formulate and implement preventive maintenance strategies for the temperature imbalance area;

[0010] The maintenance strategy execution and optimization module is used to generate a maintenance task schedule based on the maintenance strategy for the temperature abnormality area and the preventive maintenance strategy for the temperature imbalance area, and continuously optimize the maintenance strategy based on the maintenance task execution results and maintenance task execution feedback.

[0011] Furthermore, the temperature distribution image correction module is used to obtain the temperature distribution image of the industrial equipment through an infrared thermal imager, and determine the compensation coefficient of the heat reflection on the surface of the industrial equipment based on the reflection coefficient of different surface materials of the industrial equipment in the infrared band, so as to obtain the temperature distribution image after the reflection effect is corrected, including:

[0012] The temperature distribution image of the surface of industrial equipment is obtained by infrared thermal imager; the reflectance spectrum information of different types of industrial equipment surface materials in each infrared band is obtained by hyperspectral imager, and different types of industrial equipment surface materials include but are not limited to multi-layer metal covering, chrome coating, ceramic coating, and reflective coating; the reflectance spectrum data of various types of industrial equipment surface materials are compared with the data in the standard reflectance spectrum library, and the infrared reflectance coefficient of various types of industrial equipment surface materials is determined by the least squares method; according to the roughness data of the industrial equipment surface, the infrared reflectance coefficient correction module is used to correct the infrared reflectance coefficient of the industrial equipment surface; according to the reflectance coefficient of the industrial equipment surface in the infrared band, the thermal reflection compensation coefficient formula is used Determine the heat reflection compensation coefficient C of the industrial equipment surface, where R a is the corrected infrared reflection coefficient of the surface of the industrial equipment, n is a correction factor between 0 and 1, which is determined by fitting the infrared reflection coefficient and is used to adjust the correction amplitude. According to the thermal reflection compensation coefficient of the surface of the industrial equipment, the temperature distribution image obtained by the infrared thermal imager is corrected to obtain the temperature distribution image after correcting the reflection effect.

[0013] It also includes an infrared reflection coefficient correction module, which is used to correct the infrared reflection coefficient of the surface of the industrial equipment according to the roughness data of the surface of the industrial equipment, specifically including:

[0014] Use a surface profiler to measure the three-dimensional profile of the surface of industrial equipment, obtain the roughness data of the surface of industrial equipment, and use the roughness correction coefficient formula Correct the infrared reflectance coefficient of the surface material of industrial equipment, where R a is the corrected infrared reflection coefficient, R o is the original infrared reflection coefficient, obtained from the standard reflection spectrum, k is an empirical constant, reflecting the material type and its sensitivity to roughness changes. It is obtained by measuring the reflection coefficient changes of different materials under different roughness conditions and extracting it using the data fitting method. σ is the surface roughness of the industrial equipment, and λ is the wavelength of the incident light. The corrected infrared reflection coefficient is applied to the spectral data, the spectrum is re-matched, and the spectral characteristics of the surface material of the industrial equipment are determined to obtain the infrared band spectral characteristic diagram of the industrial equipment surface after adjustment.

[0015] Furthermore, the image displacement correction module is used to obtain displacement data of displacement components of industrial equipment, and to correct the temperature distribution image offset caused by the displacement by comparing the pixel displacement in the temperature distribution image with the displacement data, including:

[0016] A high-speed camera is used to obtain the displacement data of the displacement components of industrial equipment, and the displacement data is timestamped. The displacement includes displacement, velocity, and acceleration data. By comparing the timestamps of the displacement data and the temperature distribution image data, the motion data is time-synchronized using a linear interpolation method or a Lagrange interpolation method, including determining the corresponding motion data point for each frame of the temperature distribution image, adjusting the time deviation, and matching the displacement data with the frame number and time of the temperature distribution image to obtain a time-series aligned data set. Based on the aligned displacement data and temperature distribution image, the motion vector data of each pixel in the temperature distribution image is calculated. By comparing the motion vector data with the temperature distribution image at the pixel level, the position of each pixel in the image is corrected using an inverse transformation algorithm, including performing inverse motion compensation on each frame of the image, aligning the offset pixels to the position where they should be in a non-motion state, and obtaining a corrected temperature distribution image. By comparing multiple frames of corrected images, all image frames are aligned to the same coordinate system using image registration technology to obtain a temperature distribution image after the coordinates are aligned.

[0017] Furthermore, the temperature abnormality region identification module is used to identify potential temperature abnormality regions by setting a temperature threshold, predict regional characteristic data of stress concentration regions of industrial equipment under the current working state based on the equipment type and working state of the industrial equipment, and identify temperature abnormality regions according to the similarity between the regional characteristic data of the stress concentration region and the regional characteristic data of the potential temperature abnormality region, including:

[0018] The temperature distribution image after aligning the coordinates is converted into a grayscale image, the first temperature threshold is set as the segmentation standard, the grayscale image is binarized, and the image area with a temperature higher than the set first temperature threshold is marked as a potential temperature anomaly area; the Canny edge detection algorithm is used to extract the regional contour of the potential temperature anomaly area; the area of ​​the potential temperature anomaly area is determined by calculating the number of pixels within the regional contour; the shape factor calculation formula is used Calculate the shape factor of the potential temperature anomaly area, where φ i is the area shape factor, P i is the perimeter of the potential temperature anomaly area, A iis the area of ​​the potential temperature anomaly area; the location of the potential temperature anomaly area is determined by calculating the centroid of the potential temperature anomaly area; the rate of temperature change in the potential temperature anomaly area is calculated using the central difference method to obtain the temperature gradient distribution of the potential temperature anomaly area; a statistical analysis is performed on the temperature gradient distribution of the potential temperature anomaly area to obtain the temperature gradient characteristics of the potential temperature anomaly area, including the average temperature gradient, the maximum temperature gradient and the standard deviation of the gradient direction; the area, shape factor, position and temperature gradient characteristics of the potential temperature anomaly area are standardized and combined into a feature vector to constitute the regional feature data of the potential temperature anomaly area; the temperature distribution image of the industrial equipment under different working conditions is obtained, converted into a grayscale image, the second temperature threshold is set as the segmentation standard, the grayscale image is binarized, and the image area with a temperature higher than the set second temperature threshold is marked as the area that should be segmented. The method comprises the following steps: identifying a stress concentration area and obtaining regional characteristic data of the stress concentration area, wherein the working state includes electrical parameters and working time, and the electrical parameters include current, voltage and power; constructing a stress concentration area recognition model according to the equipment type, working state and regional characteristic data of the stress concentration area of ​​the industrial equipment, and predicting the regional characteristic data of the stress concentration area of ​​the industrial equipment under different working states; using the stress concentration area recognition model according to the equipment type and working state of the industrial equipment, predicting the regional characteristic data of the stress concentration area of ​​the industrial equipment under the current working state; using the cosine similarity calculation method, determining the similarity between the regional characteristic data of the stress concentration area and the regional characteristic data of the potential temperature abnormality area; if the similarity is higher than a preset first similarity threshold, determining that there is no temperature abnormality area; and if the similarity is lower than the preset first similarity threshold, determining that the potential temperature abnormality area is a temperature abnormality area.

[0019] Furthermore, the temperature abnormal area maintenance strategy formulation module is used to obtain regional characteristic data of historical temperature abnormal areas, build a fault diagnosis model for temperature abnormal areas, determine the fault type and fault severity of the temperature abnormal area, and formulate a maintenance strategy for the temperature abnormal area, including:

[0020] Through the historical monitoring database of industrial equipment, the regional characteristic data of historical temperature anomaly areas are obtained, and the corresponding fault type labels and fault severity labels are marked for the regional characteristic data of the temperature anomaly areas. The random forest algorithm is used for model training to build a fault diagnosis model for the temperature anomaly area; the regional characteristic data of new temperature anomaly areas are obtained, and the trained fault diagnosis model of the temperature anomaly area is used to determine the fault type and fault severity of the new temperature anomaly area, and maintenance strategies for temperature anomaly areas with different fault types and severities are formulated and implemented.

[0021] Furthermore, the temperature imbalance region identification module is used to monitor the temperature change rate of the non-temperature abnormal region and identify the temperature imbalance region in the non-temperature abnormal region, including:

[0022] The temperature distribution image of the non-temperature abnormal area is acquired in real time by an infrared thermal imager to obtain the temperature data and timestamp of the non-temperature abnormal area; the temperature change rate of the non-temperature abnormal area is calculated according to the temperature data of continuous time points in the temperature data of the non-temperature abnormal area; according to the temperature change rate of the non-temperature abnormal area, a temperature change rate threshold is set as a segmentation standard, and the temperature distribution image of the non-temperature abnormal area is binarized using image processing technology, and the temperature distribution image area with a temperature change rate higher than the set temperature change rate threshold is marked as a temperature imbalance area.

[0023] Furthermore, the temperature imbalance area maintenance module is used to predict the time point and location where the temperature in the temperature imbalance area is greater than the set temperature threshold according to the historical temperature monitoring data of the temperature imbalance area, and formulate and implement a preventive maintenance strategy for the temperature imbalance area, including:

[0024] Through the historical monitoring database of industrial equipment, the historical temperature monitoring data of the temperature imbalance area is obtained, and the long short-term memory network is used for model training to predict the temperature change trend of the temperature imbalance area in the future preset time period, and judge whether the temperature of the temperature imbalance area will exceed the set temperature threshold in the future preset time period; if the temperature of the temperature imbalance area will exceed the set temperature threshold in the future preset time period, then determine the time point and position in the temperature imbalance area where the temperature is greater than the set temperature threshold; according to the time point and position in the temperature imbalance area where the temperature is greater than the set temperature threshold, formulate and implement the preventive maintenance strategy for the temperature imbalance area, which includes replacing parts, adjusting the operating parameters of industrial equipment in advance, and cleaning the temperature imbalance area.

[0025] Furthermore, the maintenance strategy execution and optimization module is used to generate a maintenance task schedule according to the maintenance strategy for the temperature abnormality area and the preventive maintenance strategy for the temperature imbalance area, and continuously optimize the maintenance strategy based on the maintenance task execution results and the maintenance task execution feedback, including:

[0026] According to the maintenance strategy for temperature anomaly areas and the preventive maintenance strategy for temperature imbalance areas, a maintenance task schedule is generated, and maintenance tasks are assigned to corresponding automated industrial equipment or operators to obtain a maintenance task execution instruction set, which includes operating steps, industrial equipment startup or shutdown commands, material preparation, and personnel scheduling; the process data of the maintenance task is monitored in real time to determine the execution result of the maintenance task. The process data of the maintenance task includes the operation log during the execution process, the temperature distribution image obtained by the infrared thermal imager, and the maintenance progress information; according to the maintenance task execution results and the maintenance task execution feedback, the maintenance strategy is optimized to generate an optimized maintenance strategy and obtain an optimized maintenance task execution instruction set; the optimized maintenance task is re-executed, the maintenance task execution process is monitored in real time, and the new maintenance process data is recorded to continuously optimize the maintenance strategy.

[0027] The technical solution provided by the embodiment of the present invention may have the following beneficial effects:

[0028] The present invention provides an industrial equipment operation monitoring and early warning system based on infrared temperature distribution technology. The present invention obtains a more accurate temperature distribution image by correcting the reflection effect and roughness influence of the surface of the industrial equipment, and solves the measurement error problem caused by the difference in reflection coefficient and surface roughness in traditional temperature monitoring. The present invention obtains the displacement data of the industrial equipment and corrects the pixel displacement in the temperature distribution image, thereby eliminating the image offset caused by the movement or vibration of the equipment and improving the stability and reliability in the monitoring process. The present invention predicts the characteristic data of the stress concentration area by combining the equipment type and the working state, and further uses similarity analysis to effectively identify potential temperature abnormality areas. This method can more accurately identify and diagnose temperature abnormalities, reduce false alarms and missed alarms, and improve the accuracy of fault warning. The present invention also accurately determines the fault type and severity of the temperature abnormality area by establishing a fault diagnosis model, and then formulates an effective maintenance strategy. For temperature changes in non-temperature abnormal areas, the present invention can monitor and identify temperature imbalance areas in real time, give early warnings and implement preventive maintenance, and reduce the risk of industrial equipment failures. The present invention greatly improves the accuracy and reliability of temperature monitoring of industrial equipment by combining infrared thermal imaging technology with multiple data correction and analysis methods, ensures the safe operation of industrial equipment, extends the service life of industrial equipment, and improves the efficiency and effectiveness of maintenance work. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 This is a flow chart of an industrial equipment operation monitoring and early warning system based on infrared temperature distribution technology of the present invention;

[0030] Figure 2 It is a schematic diagram of an industrial equipment operation monitoring and early warning system based on infrared temperature distribution technology of the present invention;

[0031] Figure 3 This is another schematic diagram of an industrial equipment operation monitoring and early warning system based on infrared temperature distribution technology according to the present invention. DETAILED DESCRIPTION

[0032] In order to make the objectives, technical strategies and advantages of the present invention more clear, the present invention is described in detail below with reference to the accompanying drawings and specific embodiments.

[0033] like Figure 1-3 In this embodiment, an industrial equipment operation monitoring and early warning system based on infrared temperature distribution technology may specifically include:

[0034] Step S101, a temperature distribution image correction module is used to obtain a temperature distribution image of industrial equipment through an infrared thermal imager, and determine a compensation coefficient for heat reflection on the surface of the industrial equipment based on the reflection coefficients of different surface materials of the industrial equipment in the infrared band, to obtain a temperature distribution image after correcting the reflection effect.

[0035] The temperature distribution image of the surface of industrial equipment is obtained by infrared thermal imager. The reflectance spectrum information of different types of industrial equipment surface materials in each infrared band is obtained by hyperspectral imager. Different types of industrial equipment surface materials include but are not limited to multi-layer metal covering, chrome coating, ceramic coating, and reflective coating. The reflectance spectrum data of various types of industrial equipment surface materials are compared with the data in the standard reflectance spectrum library, and the infrared reflectance coefficient of various types of industrial equipment surface materials is determined by the least squares method. According to the roughness data of the industrial equipment surface, the infrared reflectance coefficient correction module is used to correct the infrared reflectance coefficient of the industrial equipment surface. According to the reflectance coefficient of the industrial equipment surface in the infrared band, the thermal reflection compensation coefficient formula is used Determine the heat reflection compensation coefficient C of the industrial equipment surface, where R a is the corrected infrared reflection coefficient of the industrial equipment surface, and n is a correction factor between 0 and 1, which is determined by fitting the infrared reflection coefficient and is used to adjust the correction amplitude. According to the thermal reflection compensation coefficient of the industrial equipment surface, the temperature distribution image obtained by the infrared thermal imager is corrected to obtain the temperature distribution image after the reflection effect is corrected.

[0036] For example, different industrial equipment has different surface materials, and their surface materials are usually multi-layer metal coverings, chrome coatings, ceramic coatings, reflective coatings, etc. An infrared thermal imager was used to obtain a temperature distribution image of the surface of a CNC machine tool coated with a reflective coating, showing that the surface temperature of the CNC machine tool was 85°C. However, due to the high reflectivity of the surface coating, this temperature may not be the true surface temperature. A hyperspectral imager was used to scan the coating material on the surface of the CNC machine tool, and it was found that the reflectance spectrum of the material was similar to that of a certain reflective coating in the standard spectral library. By comparing the reflectance spectrum data using the least squares method, the initial infrared reflection coefficient of the coating material was determined to be 0.75. According to the measurement of the roughness data of the CNC machine tool surface, the surface roughness affects the reflection characteristics, so the initial infrared reflection coefficient was corrected, and the corrected infrared reflection coefficient R a Calculated as 0.72. Using the heat reflection compensation coefficient formula Calculate and determine the heat reflection compensation coefficient C of the CNC machine tool surface, where R a is the corrected infrared reflection coefficient of the CNC machine tool surface, n is a correction factor between 0 and 1, determined by infrared reflection coefficient fitting, and used to adjust the correction amplitude. If n = 0.5, substitute the corrected infrared reflection coefficient R a =0.72, and the heat reflection compensation coefficient C≈1.56 is obtained. This heat reflection compensation coefficient C=1.56 is applied to each measurement point in the initial temperature distribution image. For example, at a point with a temperature of 85°C, the temperature value is multiplied by the compensation coefficient for correction, and the corrected temperature is 85°C×3.57≈132.6°C, thereby obtaining the actual temperature distribution image after the reflection effect is corrected, indicating that the actual surface temperature of the CNC machine tool may be much higher than the initially measured temperature.

[0037] Among them, the infrared reflection coefficient correction module is used to correct the infrared reflection coefficient of the surface of the industrial equipment according to the roughness data of the surface of the industrial equipment.

[0038] Use a surface profiler to measure the three-dimensional profile of the surface of industrial equipment, obtain the roughness data of the surface of industrial equipment, and use the roughness correction coefficient formula Correct the infrared reflectance coefficient of the surface material of industrial equipment, where R a is the corrected infrared reflection coefficient, R ois the original infrared reflection coefficient, obtained from the standard reflection spectrum, k is an empirical constant, reflecting the material type and its sensitivity to roughness changes, and is extracted from it by measuring the reflection coefficient changes of different materials under different roughness conditions and using data fitting methods, σ is the surface roughness of industrial equipment, and λ is the wavelength of incident light. The corrected infrared reflection coefficient is applied to the spectral data, the spectrum is re-matched, and the spectral characteristics of the surface material of the industrial equipment are determined, and the infrared band spectral characteristic diagram of the adjusted industrial equipment surface is obtained.

[0039] For example, a surface profiler is used to measure the three-dimensional profile of a CNC machine tool surface, and the surface roughness σ = 2μm is obtained. The surface of this CNC machine tool is covered with a specific reflective coating material. The initial infrared reflection coefficient Ro of the reflective coating material is known to be 0.8 from the standard reflection spectrum library. The roughness correction empirical constant k of this material has been experimentally determined to be 0.05 through different materials under different roughness conditions, and the wavelength λ of the infrared incident light is 10\mum. Substituting these values ​​into the roughness correction coefficient formula Correct the infrared reflection coefficient of the surface material of the CNC machine tool to obtain the corrected infrared reflection coefficient R a is 0.792, where R a is the corrected infrared reflection coefficient, R o is the original infrared reflection coefficient obtained from the standard reflection spectrum. This corrected reflection coefficient is applied to the spectral data, and the spectrum matching is re-performed to finally determine the infrared band spectral characteristics of the CNC machine tool surface, and the adjusted spectral characteristic diagram is obtained, which reflects the real optical characteristics of the CNC machine tool surface.

[0040] Step S102, an image displacement correction module is used to obtain displacement data of displacement components of industrial equipment, and correct the temperature distribution image offset caused by the displacement by comparing the pixel displacement in the temperature distribution image with the displacement data.

[0041] Use a high-speed camera to obtain the displacement data of the displacement components of industrial equipment, and mark the displacement data with a timestamp. The displacement includes displacement, velocity, and acceleration data. By comparing the timestamps of the displacement data and the temperature distribution image data, the linear interpolation method or the Lagrange interpolation method is used to perform time synchronization processing on the motion data, including determining the corresponding motion data point for each frame of the temperature distribution image, adjusting the time deviation, and matching the displacement data with the frame number and time of the temperature distribution image to obtain a time-series aligned data set. According to the aligned displacement data and temperature distribution image, the motion vector data of each pixel in the temperature distribution image is calculated. By comparing the motion vector data with the temperature distribution image at the pixel level, the inverse transformation algorithm is used to correct the position of each pixel in the image, including performing inverse motion compensation on each frame of the image, aligning the offset pixels to the position where they should be in the non-motion state, and obtaining the corrected temperature distribution image. By comparing multiple frames of corrected images, all image frames are aligned to the same coordinate system using image registration technology to obtain a temperature distribution image after the coordinates are aligned.

[0042] For example, there are many industrial equipment that are constantly moving, such as conveyor belts, compressors, and CNC machine tools. If an infrared thermal image of a CNC machine tool is obtained by an infrared thermal imager, one of the displacement components, such as a rotary table, moves back and forth at a speed of 50 times per second. A high-speed camera is used to capture the movement of the rotary table at a rate of 1000 frames per second, and the displacement data of the rotary table during operation is obtained, including displacement, speed, and acceleration, and a timestamp is marked for each set of displacement data. For example, at the 10th second, the displacement of the rotary table is 5 mm, the speed is 100 mm / s, and the acceleration is 20 mm / s2. The displacement data at this time point is timestamped with the temperature distribution image data obtained using an infrared thermal imager. The time deviation of the displacement data and the temperature distribution image is adjusted by a linear interpolation method so that the timing of the two can be accurately aligned. If at 10 seconds, the corresponding temperature distribution image is aligned, and it is determined that a pixel point in the image at this moment has a displacement of 2 pixels due to the movement of the rotary table. According to the aligned displacement data and temperature distribution image, the motion vector data of each pixel in the temperature distribution image is calculated. If the motion vector of a pixel is (2 pixels, to the right), it means that it has been displaced by 2 pixels due to the movement of the rotating table. The temperature distribution image is corrected by the inverse transformation algorithm, and each pixel is compensated for the inverse motion. The offset pixel caused by the movement of the rotating table is corrected to the position where it should be in the non-moving state, and it is moved 2 pixels to the left to return to the position where it should be, and the corrected temperature distribution image is obtained. By comparing multiple frames of corrected images, all image frames are aligned to the same coordinate system using image registration technology, such as accurately superimposing the temperature information in each frame of the image to the reference coordinate system to obtain the temperature distribution image after the coordinates are aligned. In this process, if there is a 1-pixel deviation in the image frames acquired at different time points, these deviations are corrected by the registration technology to ensure that the temperature data of all image frames are accurately reflected in the same coordinate system.

[0043] Step S103, a temperature abnormality area identification module is used to identify potential temperature abnormality areas by setting a temperature threshold, predict the regional characteristic data of the stress concentration area of ​​the industrial equipment under the current working state based on the equipment type and working state of the industrial equipment, and identify the temperature abnormality area based on the similarity between the regional characteristic data of the stress concentration area and the regional characteristic data of the potential temperature abnormality area.

[0044] Convert the temperature distribution image after aligning the coordinates into a grayscale image, set the first temperature threshold as the segmentation standard, perform binarization on the grayscale image, and mark the image area with a temperature higher than the set first temperature threshold as a potential temperature anomaly area. Use the Canny edge detection algorithm to extract the regional contour of the potential temperature anomaly area. Determine the area of ​​the potential temperature anomaly area by calculating the number of pixels within the regional contour. Use the shape factor calculation formula Calculate the shape factor of the potential temperature anomaly area, where φ i is the area shape factor, P i is the perimeter of the potential temperature anomaly area, A i is the area of ​​the potential temperature anomaly region. The location of the potential temperature anomaly region is determined by calculating the centroid of the potential temperature anomaly region. The rate of temperature change in the potential temperature anomaly region is calculated using the central difference method to obtain the temperature gradient distribution of the potential temperature anomaly region. The temperature gradient distribution of the potential temperature anomaly region is statistically analyzed to obtain the temperature gradient characteristics of the potential temperature anomaly region, including the average temperature gradient, the maximum temperature gradient and the standard deviation of the gradient direction. The area, shape factor, position and temperature gradient characteristics of the potential temperature anomaly region are standardized and combined into a feature vector to form the regional characteristic data of the potential temperature anomaly region. The temperature distribution image of the industrial equipment under different working conditions is obtained, converted into a grayscale image, and the second temperature threshold is set as the segmentation standard. The grayscale image is binarized, and the image area with a temperature higher than the set second temperature threshold is marked as a stress concentration area, and the regional characteristic data of the stress concentration area is obtained. The working state includes electrical parameters and working time, and the electrical parameters include current, voltage and power. According to the equipment type, working state and regional characteristic data of the stress concentration area of ​​the industrial equipment, a stress concentration area recognition model is constructed to predict the regional characteristic data of the stress concentration area under different working conditions of the industrial equipment. According to the equipment type and working state of the industrial equipment, the stress concentration area identification model is used to predict the regional characteristic data of the stress concentration area of ​​the industrial equipment under the current working state. The cosine similarity calculation method is used to determine the similarity between the regional characteristic data of the stress concentration area and the regional characteristic data of the potential temperature anomaly area. If the similarity is higher than the preset first similarity threshold, it is judged that there is no temperature anomaly area. If the similarity is lower than the preset first similarity threshold, the potential temperature anomaly area is judged to be a temperature anomaly area.

[0045] For example, an infrared thermal imager is used to obtain a temperature distribution image of a CNC machine tool, which is converted into a grayscale image. The temperature threshold is set to 80°C, and the areas with a temperature higher than 80°C in the grayscale image are marked as temperature abnormal areas. The Canny edge detection algorithm is used to extract the contours of these temperature abnormal areas, and the perimeter P of one of the temperature abnormal areas is calculated. iis 50 pixels, area A i is 200 pixels. According to the shape factor calculation formula Calculate the shape factor of the temperature anomaly area and get the shape factor φ of the area i ≈0.995, which indicates that the shape of the area is close to a circle, and the centroid position of the temperature anomaly area is determined to be located at (150,120) in the image coordinate system. The temperature change rate in the area was calculated using the central difference method, and it was found that the average temperature gradient was 3℃ / pixel, the maximum temperature gradient was 5℃ / pixel, and the standard deviation of the gradient direction was 0.5℃ / pixel. The area, shape factor, position, and temperature gradient characteristics of the temperature anomaly area were standardized and combined into a feature vector. The temperature distribution images of industrial equipment under different working conditions were obtained, the second temperature threshold was set to 90℃, the grayscale image was binarized, the area with a temperature higher than 90℃ was marked as a stress concentration area, and its regional feature data were calculated, including an area of ​​180 pixels, a shape factor of 0.9, a centroid position of (155,125), an average temperature gradient of 3.5℃ / pixel, and a maximum temperature gradient of 5.5℃ / pixel. The cosine similarity calculation method is used to determine the similarity between the regional characteristic data of the stress concentration area and the characteristic data of the potential temperature anomaly area, and the similarity is 0.85. Since the similarity is lower than the preset first similarity threshold of 0.9, the potential temperature anomaly area is judged to be a temperature anomaly area.

[0046] Step S104, a temperature abnormality area maintenance strategy formulation module is used to obtain regional characteristic data of historical temperature abnormality areas, build a fault diagnosis model for temperature abnormality areas, determine the fault type and fault severity of temperature abnormality areas, and formulate a maintenance strategy for temperature abnormality areas.

[0047] Through the historical monitoring database of industrial equipment, the regional characteristic data of historical temperature anomaly areas are obtained, and the corresponding fault type labels and fault severity labels are annotated for the regional characteristic data of the temperature anomaly areas. The model is trained using the random forest algorithm to build a fault diagnosis model for the temperature anomaly area. The regional characteristic data of new temperature anomaly areas are obtained, and the fault type and fault severity of the new temperature anomaly areas are determined using the trained fault diagnosis model for the temperature anomaly areas. Maintenance strategies for temperature anomaly areas with different fault types and severities are formulated and implemented.

[0048] For example, by querying the historical monitoring database of CNC machine tools, historical feature data of similar temperature abnormality areas were obtained, and the corresponding fault type was marked as local overheating and the fault severity was medium. These historical data were input into the random forest algorithm for model training, and a fault diagnosis model for temperature abnormality areas was constructed. When a new temperature abnormality area appears, it is analyzed using the trained fault diagnosis model to determine that the fault type in the area is still local overheating, but the fault severity is judged to be severe. Based on this diagnostic result, a corresponding maintenance strategy was formulated and implemented, including immediately reducing the load of the CNC machine tool and arranging emergency maintenance to prevent further damage to the CNC machine tool.

[0049] Step S105: a temperature imbalance region identification module is used to monitor the temperature change rate of the non-temperature abnormal region and identify the temperature imbalance region in the non-temperature abnormal region.

[0050] The temperature distribution image of the non-temperature abnormal area is obtained in real time by an infrared thermal imager, and the temperature data and timestamp of the non-temperature abnormal area are obtained. The temperature change rate of the non-temperature abnormal area is calculated based on the temperature data of the temperature data of the non-temperature abnormal area at consecutive time points. According to the temperature change rate of the non-temperature abnormal area, the temperature change rate threshold is set as the segmentation standard, and the temperature distribution image of the non-temperature abnormal area is binarized using image processing technology, and the temperature distribution image area with a temperature change rate higher than the set temperature change rate threshold is marked as a temperature imbalance area.

[0051] For example, an infrared thermal imager is used to obtain a temperature distribution image of a non-temperature abnormal area on the surface of a CNC machine tool in real time. During the monitoring process, at 8 a.m. one day, the temperature of a certain area was 65°C. Subsequently, in the next four hours, we recorded the temperature data of the area every hour, which were 67°C at 9 o'clock, 69°C at 10 o'clock, 72°C at 11 o'clock, and 76°C at 12 o'clock. Therefore, it is calculated that the temperature change rate from 8 o'clock to 9 o'clock is 2°C / hour, the temperature change rate from 9 o'clock to 10 o'clock is 2°C / hour, the temperature change rate from 10 o'clock to 11 o'clock is 3°C / hour, and the temperature change rate from 11 o'clock to 12 o'clock is 4°C / hour. A temperature change rate threshold of 3°C / hour is set. The temperature distribution image of the area is binarized using image processing technology, and the area with a temperature change rate higher than 3°C / hour is marked as a temperature imbalance area. The temperature change rate from 11 o'clock to 12 o'clock is 4°C / hour, which exceeds the set threshold. Therefore, the temperature distribution image during this time period is marked as a temperature imbalance area. The CNC machine tool may have potential abnormalities during this time period and requires further inspection and maintenance.

[0052] Step S106, a temperature imbalance area maintenance module is used to predict the time point and location where the temperature in the temperature imbalance area is greater than the set temperature threshold based on the historical temperature monitoring data of the temperature imbalance area, and formulate and implement a preventive maintenance strategy for the temperature imbalance area.

[0053] Through the historical monitoring database of industrial equipment, the historical temperature monitoring data of the temperature imbalance area is obtained, and the long short-term memory network is used for model training to predict the temperature change trend of the temperature imbalance area in the future preset time period, and to determine whether the temperature of the temperature imbalance area will exceed the set temperature threshold in the future preset time period. If the temperature of the temperature imbalance area will exceed the set temperature threshold in the future preset time period, the time point and location where the temperature in the temperature imbalance area is greater than the set temperature threshold are determined. According to the time point and location where the temperature in the temperature imbalance area is greater than the set temperature threshold, a preventive maintenance strategy for the temperature imbalance area is formulated and implemented. The preventive maintenance strategy includes replacing parts, adjusting the operating parameters of industrial equipment in advance, and cleaning the temperature imbalance area.

[0054] For example, a CNC machine tool is monitored for a long time, and the historical temperature data of a temperature imbalance area is obtained through its historical monitoring database. It is found that in the monitoring data of this area in the past month, the temperature fluctuates greatly, especially when the CNC machine tool is running at high load, the temperature is often close to the set temperature threshold of 80°C. Using these historical data, a long short-term memory network model is trained to predict the temperature change trend of this temperature imbalance area in the next week. The model prediction shows that the temperature of this area will continue to rise during the daily peak operation period of the CNC machine tool from 12:00 to 14:00 in the next three days, and is expected to reach 82°C at 13:00 on the third day, exceeding the set temperature threshold of 80°C. Based on this prediction result, the time point and location where the temperature is greater than 80°C in the temperature imbalance area are determined, that is, at 13:00 on the third day, the temperature near the key component A of the CNC machine tool operation will exceed the threshold. In order to prevent the CNC machine from overheating during the peak period on the third day, a preventive maintenance strategy was developed and implemented. Specific measures included replacing key component A that may cause overheating in advance, adjusting the operating parameters of the CNC machine, reducing the operating load by 20% from 12:00 to 14:00 on the third day, and performing a comprehensive cleaning of the temperature imbalance area on the morning of the third day to ensure good heat dissipation. Through these preventive maintenance measures, the risk of failure of the CNC machine due to overheating was successfully avoided, ensuring the stable operation of the CNC machine.

[0055] Step S107, the maintenance strategy execution and optimization module is used to generate a maintenance task schedule according to the maintenance strategy for the temperature abnormality area and the preventive maintenance strategy for the temperature imbalance area, and continuously optimize the maintenance strategy based on the maintenance task execution results and maintenance task execution feedback.

[0056] According to the maintenance strategy for temperature anomaly areas and the preventive maintenance strategy for temperature imbalance areas, a maintenance task schedule is generated, and maintenance tasks are assigned to corresponding automated industrial equipment or operators to obtain a maintenance task execution instruction set, which includes operation steps, industrial equipment startup or shutdown commands, material preparation, and personnel scheduling. The process data of the maintenance task is monitored in real time to determine the execution results of the maintenance task. The process data of the maintenance task includes the operation log during the execution process, the temperature distribution image obtained by the infrared thermal imager, and the maintenance progress information. According to the maintenance task execution results and maintenance task execution feedback, the maintenance strategy is optimized, the optimized maintenance strategy is generated, and the optimized maintenance task execution instruction set is obtained. Re-execute the optimized maintenance task, monitor the maintenance task execution process in real time, and record the new maintenance process data to continuously optimize the maintenance strategy.

[0057] For example, when monitoring a large CNC machine tool, two areas that need to be processed were found, one is the temperature abnormality area, for which a corresponding maintenance strategy has been formulated, and the other is the temperature imbalance area, for which a preventive maintenance strategy has been formulated. Based on these two strategies, a maintenance task schedule is generated, which is scheduled to be completed between 9:00 and 11:00 this Thursday morning. The schedule clearly stipulates that the temperature abnormality area is first cooled by the automated CNC machine tool, and then the operator checks and replaces the parts that may be damaged, and at the same time adjusts the operating parameters of the CNC machine tool and cleans the area in the temperature imbalance area. According to the schedule, specific maintenance tasks are assigned to the automated CNC machine tool and the operator, and a maintenance task execution instruction set is generated. The instruction set specifies the operating steps of each task, including starting the cooling system at 9 o'clock to reduce the temperature of the temperature abnormality area, the operator replacing the key components of the temperature abnormality area at 9:30, adjusting the operating parameters of the temperature imbalance area at 10 o'clock, and the automated CNC machine tool performing the surface cleaning task. In addition, the instruction set also includes the preparation of required materials and personnel scheduling arrangements, and the required materials include coolants, replacement parts, etc. During the execution of the maintenance task, the process data of each operation is monitored in real time, including the operation log during the execution process, the temperature distribution image obtained in real time by the infrared thermal imager, and the progress information of the maintenance task. Through monitoring, it was found that the temperature of the temperature abnormal area dropped rapidly to a safe level after cooling, but when replacing parts, the cleaning work of the temperature imbalance area was delayed due to installation problems. According to the execution results and feedback of these maintenance tasks, the maintenance strategy was optimized. For example, for the installation problem of component replacement, the installation steps were adjusted and an additional inspection link was added. After the optimized maintenance strategy was generated, a new maintenance task execution instruction set was obtained again, and the optimized maintenance task was re-executed. In the process of re-executing the task, the entire maintenance process was continuously monitored in real time and new maintenance process data was recorded. Until the execution of the maintenance task progressed smoothly, all maintenance tasks were completed as planned, and the temperatures of the temperature abnormal area and the temperature imbalance area returned to normal levels.

[0058] The above description is only a preferred embodiment of the present application and an explanation of the technical principles used. Those skilled in the art should understand that the scope of the invention involved in the present application is not limited to the technical solution formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the concept of the present application. For example, the above features are replaced with (but not limited to) technical features with similar functions disclosed in the present application.

Claims

1. An industrial equipment operation monitoring and early warning system based on infrared temperature distribution technology, characterized in that: The system comprises: The temperature distribution image correction module is used to obtain the temperature distribution image of the industrial equipment through an infrared thermal imager, and determine the compensation coefficient of the heat reflection on the surface of the industrial equipment based on the reflection coefficient of different surface materials of the industrial equipment in the infrared band, so as to obtain the temperature distribution image after the reflection effect is corrected; An image displacement correction module is used to obtain displacement data of displacement components of industrial equipment, and to correct the temperature distribution image offset caused by the displacement by comparing the pixel displacement in the temperature distribution image with the displacement data; The temperature abnormality region identification module is used to identify potential temperature abnormality regions by setting a temperature threshold, predict regional characteristic data of stress concentration regions of industrial equipment under the current working state based on the equipment type and working state of the industrial equipment, and identify the temperature abnormality region based on the similarity between the regional characteristic data of the stress concentration region and the regional characteristic data of the potential temperature abnormality region; The maintenance strategy formulation module for abnormal temperature areas is used to obtain the regional characteristic data of historical abnormal temperature areas, build a fault diagnosis model for abnormal temperature areas, determine the fault type and fault severity of abnormal temperature areas, and formulate maintenance strategies for abnormal temperature areas; A temperature imbalance region identification module is used to monitor the temperature change rate of the non-temperature abnormal region and identify the temperature imbalance region in the non-temperature abnormal region; The temperature imbalance area maintenance module is used to predict the time point and location where the temperature in the temperature imbalance area is greater than the set temperature threshold based on the historical temperature monitoring data of the temperature imbalance area, and formulate and implement preventive maintenance strategies for the temperature imbalance area; The maintenance strategy execution and optimization module is used to generate a maintenance task schedule based on the maintenance strategy for the temperature abnormality area and the preventive maintenance strategy for the temperature imbalance area, and continuously optimize the maintenance strategy based on the maintenance task execution results and maintenance task execution feedback; The temperature distribution image correction module is used to obtain the temperature distribution image of the industrial equipment through an infrared thermal imager, and determine the compensation coefficient of the heat reflection on the surface of the industrial equipment based on the reflection coefficient of different surface materials of the industrial equipment in the infrared band, so as to obtain the temperature distribution image after the reflection effect is corrected, including: The temperature distribution image of the equipment surface is obtained by an infrared thermal imager; the reflectance spectrum information of different types of industrial equipment surface materials in each infrared band is obtained by a hyperspectral imager. Different types of industrial equipment surface materials include but are not limited to multi-layer metal coverings, chrome coatings, ceramic coatings, and reflective coatings; the reflectance spectrum data of various types of industrial equipment surface materials are compared with the data in the standard reflectance spectrum library, and the infrared reflectance coefficient of various types of industrial equipment surface materials is determined using the least squares method; according to the roughness data of the industrial equipment surface, the infrared reflectance coefficient correction module is used to correct the infrared reflectance coefficient of the industrial equipment surface; according to the reflectance coefficient of the industrial equipment surface in the infrared band, the thermal reflection compensation coefficient formula is used Determine the heat reflection compensation coefficient C of the industrial equipment surface, where R a is the corrected infrared reflection coefficient of the surface of the industrial equipment, n is a correction factor between 0 and 1, which is determined by fitting the infrared reflection coefficient and is used to adjust the correction amplitude; according to the thermal reflection compensation coefficient of the surface of the industrial equipment, the temperature distribution image obtained by the infrared thermal imager is corrected to obtain the temperature distribution image after the reflection effect is corrected; The method of correcting the infrared reflection coefficient of the surface of the industrial equipment by using the infrared reflection coefficient correction module according to the roughness data of the surface of the industrial equipment includes: Use a surface profiler to measure the three-dimensional profile of the surface of industrial equipment, obtain the roughness data of the surface of industrial equipment, and use the roughness correction coefficient formula Correct the infrared reflection coefficient of the equipment surface material, where R a is the corrected infrared reflection coefficient, R o is the original infrared reflection coefficient obtained from the standard reflection spectrum, k is an empirical constant, which reflects the material type and its sensitivity to roughness changes. It is obtained by measuring the reflection coefficient changes of different materials under different roughness conditions and extracting it using the data fitting method. σ is the surface roughness of the industrial equipment, and λ is the wavelength of the incident light. The corrected infrared reflection coefficient is applied to the spectral data, the spectrum is re-matched, and the spectral characteristics of the surface material of the industrial equipment are determined to obtain the infrared band spectral characteristic diagram of the industrial equipment surface after adjustment.

2. The system according to claim 1, wherein: The image displacement correction module is used to obtain the displacement data of the displacement components of the industrial equipment, and to correct the temperature distribution image offset caused by the displacement by comparing the pixel displacement in the temperature distribution image with the displacement data, including: Use a high-speed camera to obtain the displacement data of the displacement components of the industrial equipment, and mark the displacement data with a timestamp; compare the timestamps of the displacement data and the temperature distribution image data, and use a linear interpolation method or a Lagrange interpolation method to perform time synchronization processing on the motion data, including determining the corresponding motion data point for each frame of the temperature distribution image, adjusting the time deviation, and matching the displacement data with the frame number and time of the temperature distribution image to obtain a time-series aligned data set; calculate the motion vector data of each pixel in the temperature distribution image based on the aligned displacement data and temperature distribution image; compare the motion vector data with the temperature distribution image at the pixel level, and use an inverse transformation algorithm to correct the position of each pixel in the image, including performing inverse motion compensation on each frame of the image, aligning the offset pixels to the position where they should be in a non-motion state, and obtaining a corrected temperature distribution image; By comparing multiple frames of rectified images, all image frames are aligned to the same coordinate system using image registration technology to obtain a temperature distribution image after the aligned coordinates.

3. The system according to claim 1, wherein: The temperature abnormality region identification module is used to identify potential temperature abnormality regions by setting a temperature threshold, predict regional characteristic data of stress concentration regions of industrial equipment under the current working state based on the equipment type and working state of the industrial equipment, and identify temperature abnormality regions according to the similarity between the regional characteristic data of the stress concentration region and the regional characteristic data of the potential temperature abnormality region, including: The temperature distribution image after aligning the coordinates is converted into a grayscale image, the first temperature threshold is set as the segmentation standard, the grayscale image is binarized, and the image area with a temperature higher than the set first temperature threshold is marked as a potential temperature anomaly area; the Canny edge detection algorithm is used to extract the regional contour of the potential temperature anomaly area; the area of ​​the potential temperature anomaly area is determined by calculating the number of pixels within the regional contour; the shape factor calculation formula is used Calculate the shape factor of the potential temperature anomaly area, where φ i is the area shape factor, P i is the perimeter of the potential temperature anomaly area, A i is the area of ​​the potential temperature anomaly area; the location of the potential temperature anomaly area is determined by calculating the centroid of the potential temperature anomaly area; the rate of temperature change in the potential temperature anomaly area is calculated using the central difference method to obtain the temperature gradient distribution of the potential temperature anomaly area; a statistical analysis is performed on the temperature gradient distribution of the potential temperature anomaly area to obtain the temperature gradient characteristics of the potential temperature anomaly area, including the average temperature gradient, the maximum temperature gradient and the standard deviation of the gradient direction; the area, shape factor, position and temperature gradient characteristics of the potential temperature anomaly area are standardized and combined into a feature vector to constitute the regional feature data of the potential temperature anomaly area; the temperature distribution image of the industrial equipment under different working conditions is obtained, converted into a grayscale image, the second temperature threshold is set as the segmentation standard, the grayscale image is binarized, and the image area with a temperature higher than the set second temperature threshold is marked as the area that should be segmented. The method comprises the following steps: identifying a stress concentration area and obtaining regional characteristic data of the stress concentration area, wherein the working state includes electrical parameters and working time, and the electrical parameters include current, voltage and power; constructing a stress concentration area recognition model according to the equipment type, working state and regional characteristic data of the stress concentration area of ​​the industrial equipment, and predicting the regional characteristic data of the stress concentration area of ​​the industrial equipment under different working states; using the stress concentration area recognition model according to the equipment type and working state of the industrial equipment, predicting the regional characteristic data of the stress concentration area of ​​the industrial equipment under the current working state; using the cosine similarity calculation method, determining the similarity between the regional characteristic data of the stress concentration area and the regional characteristic data of the potential temperature abnormality area; if the similarity is higher than a preset first similarity threshold, determining that there is no temperature abnormality area; and if the similarity is lower than the preset first similarity threshold, determining that the potential temperature abnormality area is a temperature abnormality area.

4. The system according to claim 1, wherein: The temperature abnormal area maintenance strategy formulation module is used to obtain the regional characteristic data of the historical temperature abnormal area, build a fault diagnosis model for the temperature abnormal area, determine the fault type and fault severity of the temperature abnormal area, and formulate a maintenance strategy for the temperature abnormal area, including: Through the historical monitoring database of industrial equipment, the regional characteristic data of historical temperature anomaly areas are obtained, and the corresponding fault type labels and fault severity labels are marked for the regional characteristic data of the temperature anomaly areas. The random forest algorithm is used for model training to build a fault diagnosis model for the temperature anomaly area; the regional characteristic data of new temperature anomaly areas are obtained, and the trained fault diagnosis model of the temperature anomaly area is used to determine the fault type and fault severity of the new temperature anomaly area, and maintenance strategies for temperature anomaly areas with different fault types and severities are formulated and implemented.

5. The system according to claim 1, wherein: The temperature imbalance region identification module is used to monitor the temperature change rate of the non-temperature abnormal region and identify the temperature imbalance region in the non-temperature abnormal region, including: The temperature distribution image of the non-temperature abnormal area is acquired in real time by an infrared thermal imager to obtain the temperature data and timestamp of the non-temperature abnormal area; the temperature change rate of the non-temperature abnormal area is calculated according to the temperature data of continuous time points in the temperature data of the non-temperature abnormal area; according to the temperature change rate of the non-temperature abnormal area, a temperature change rate threshold is set as a segmentation standard, and the temperature distribution image of the non-temperature abnormal area is binarized using image processing technology, and the temperature distribution image area with a temperature change rate higher than the set temperature change rate threshold is marked as a temperature imbalance area.

6. The system according to claim 1, wherein: The temperature imbalance area maintenance module is used to predict the time point and location where the temperature in the temperature imbalance area is greater than the set temperature threshold according to the historical temperature monitoring data of the temperature imbalance area, and formulate and implement a preventive maintenance strategy for the temperature imbalance area, including: Through the historical monitoring database of industrial equipment, the historical temperature monitoring data of the temperature imbalance area is obtained, and the long short-term memory network is used for model training to predict the temperature change trend of the temperature imbalance area in the future preset time period, and judge whether the temperature of the temperature imbalance area will exceed the set temperature threshold in the future preset time period; if the temperature of the temperature imbalance area will exceed the set temperature threshold in the future preset time period, then determine the time point and position in the temperature imbalance area where the temperature is greater than the set temperature threshold; according to the time point and position in the temperature imbalance area where the temperature is greater than the set temperature threshold, formulate and implement the preventive maintenance strategy for the temperature imbalance area, which includes replacing parts, adjusting the operating parameters of industrial equipment in advance, and cleaning the temperature imbalance area.

7. The system according to claim 1, wherein: The maintenance strategy execution and optimization module is used to generate a maintenance task schedule according to the maintenance strategy for the temperature abnormality area and the preventive maintenance strategy for the temperature imbalance area, and continuously optimize the maintenance strategy based on the maintenance task execution results and the maintenance task execution feedback, including: Generate a maintenance task schedule based on the maintenance strategy for temperature anomaly areas and the preventive maintenance strategy for temperature imbalance areas, and assign maintenance tasks to corresponding automation equipment or operators to obtain a maintenance task execution instruction set, which includes operation steps, equipment startup or shutdown commands, material preparation, and personnel scheduling; monitor the process data of maintenance tasks in real time to determine the execution results of maintenance tasks. The process data of maintenance tasks includes the operation log during execution, the temperature distribution image obtained by the infrared thermal imager, and maintenance progress information; According to the maintenance task execution results and maintenance task execution feedback, the maintenance strategy is optimized, the optimized maintenance strategy is generated, and the optimized maintenance task execution instruction set is obtained; the optimized maintenance task is re-executed, the maintenance task execution process is monitored in real time, and the new maintenance process data is recorded to continuously optimize the maintenance strategy.

Images