Thermal infrared imager based on multilevel thermal image analysis and intelligent fault diagnosis method

Abstract

The invention discloses a thermal infrared imager based on multi-level thermal image analysis and an intelligent fault diagnosis method, and relates to the technical field of thermal infrared imagers, and the thermal infrared imager comprises a multi-level temperature data acquisition module, an intelligent temperature data processing module and an intelligent fault diagnosis and early warning module. According to the invention, the surface layer, deep layer and transient temperature data of the monitored object are collected at the same time, so that more comprehensive and accurate temperature information can be obtained, and the monitoring accuracy is improved. The intelligent temperature data processing module effectively fuses the temperature data of different levels, generates a more accurate comprehensive temperature map, and provides a reliable data basis for subsequent analysis. The intelligent fault diagnosis and early warning module automatically identifies and matches fault features by establishing a fault mode model, thereby reducing the dependence on artificial experience and improving the efficiency and accuracy of fault diagnosis. The temperature change can be monitored in real time, potential faults can be early warned, a more effective maintenance plan can be formulated by accurately diagnosing the faults, and unnecessary downtime and maintenance cost are reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of infrared thermal imagers, and in particular to an infrared thermal imager and an intelligent fault diagnosis method based on multi-level thermal image analysis. Background Art

[0002] Currently, infrared thermal imagers are increasingly widely used in the fields of equipment temperature monitoring, fault diagnosis, etc. However, traditional infrared imaging data is mostly processed based on single-layer temperature data, making it difficult to obtain accurate temperature change information.

[0003] At the same time, most of the existing fault diagnosis methods rely on manual identification and cannot meet the requirements of real-time monitoring and automatic fault warning. Therefore, a high-precision infrared thermal imager based on multi-level temperature data analysis and automated fault diagnosis has important application value. Summary of the Invention

[0004] In order to solve the above technical problems, the present invention provides an infrared thermal imager and an intelligent fault diagnosis method based on multi-level thermal image analysis. The following technical solutions are adopted: An infrared thermal imager based on multi-level thermal image analysis includes a multi-level temperature data acquisition module, an intelligent temperature data processing module, and an intelligent fault diagnosis and warning module. The multi-level temperature data acquisition module includes a surface temperature acquisition unit, a deep temperature detection unit, and a transient temperature analysis unit. The surface temperature acquisition unit captures the surface temperature data of the monitored object through a high-resolution infrared sensor; the deep temperature detection unit obtains the deep temperature data below the surface of the monitored object based on pulse thermal excitation and thermal diffusion analysis technology; the data input end of the transient temperature analysis unit is respectively communicatively connected to the data output ends of the surface temperature acquisition unit and the deep temperature detection unit, acquires the surface temperature data and deep temperature data of the monitored object surface, analyzes the rate of temperature rise or fall through time series analysis of temperature changes, and outputs transient temperature data; The intelligent temperature data processing module is communicatively connected to the transient temperature analysis unit, analyzes and fuses the surface temperature data, deep temperature data, and transient temperature data through a multi-level data processing algorithm, outputs the analysis result of multi-level temperature data, and generates a comprehensive temperature map based on the analysis result of multi-level temperature data; The intelligent fault diagnosis and warning module is communicatively connected to the intelligent temperature data processing module, extracts temperature features related to faults from the analysis result of multi-level temperature data. The temperature features include mutation points, temperature abnormal regions, and temperature difference fluctuations. A pattern model of known fault types is established using machine learning algorithms to identify and match the fault feature data. If a successfully identified and matched fault exists, a fault alarm is issued.

[0005] By adopting the above technical solutions, more comprehensive and accurate temperature information can be obtained by simultaneously collecting the surface, deep and transient temperature data of the monitored object, thereby improving the accuracy of monitoring. High-resolution infrared sensors ensure the resolution and details of surface temperature data, which helps to identify small temperature differences. Pulse thermal excitation and thermal diffusion analysis technology can detect the temperature conditions inside the object, which is particularly important for diagnosing internal defects or faults. The intelligent temperature data processing module can quickly identify temperature anomalies by analyzing the temperature change rate, providing important clues for fault diagnosis. It can effectively fuse temperature data at different levels to generate a more accurate comprehensive temperature map, providing a reliable data basis for subsequent analysis.

[0006] The intelligent fault diagnosis and early warning module can automatically identify and match fault characteristics by establishing a fault mode model, reducing dependence on manual experience and improving the efficiency and accuracy of fault diagnosis.

[0007] The system can monitor temperature changes in real time and warn of potential faults, helping to take timely measures to prevent accidents. By accurately diagnosing faults, more effective maintenance plans can be developed, reducing unnecessary downtime and maintenance costs.

[0008] Infrared thermal imagers based on multi-level thermal image analysis are suitable for temperature monitoring and fault diagnosis of various industrial equipment. They can be used to detect heat loss areas in buildings, optimize energy use, provide real-time temperature monitoring in high-risk environments, and enhance operational safety.

[0009] Optionally, it also includes a visual multi-layer temperature image generation module, which includes a visual data storage device, a graphics processing unit and a display. The visual data storage device is communicated with the intelligent temperature data processing module to collect surface temperature data, deep temperature data and transient temperature data. The graphics processing unit outputs multi-layer temperature image data, and the display is communicated with the graphics processing unit.

[0010] By adopting the above technical solution, the graphics processing unit generates a comprehensive multi-layer temperature image by fusing temperature data at different levels, thereby achieving an intuitive display of temperature distribution.

[0011] Color coding and layered display function: Different temperature levels are displayed with different color codes, and the temperature distribution is displayed through the changes in the depth of the image color.

[0012] Color coding can help users quickly identify abnormal temperature areas, and the layered display function can show surface and deep temperature information separately, allowing users to observe the temperature distribution and changes more clearly.

[0013] Zoom and focus viewing functions: Users can zoom in and out of the temperature image and select the focus to observe the temperature changes in a specific area.

[0014] The zoom and focus functions help users to view the detailed temperature changes in the fault hot spot or potential risk area in depth, so as to more effectively locate the fault source and improve the analysis efficiency.

[0015] Optionally, the surface temperature acquisition unit includes an infrared lens, an infrared sensor, and a sensor drive circuit. The infrared lens receives the infrared radiation emitted by the object to be monitored, focuses and transmits the infrared radiation to the infrared sensor. The data output end of the infrared sensor is communicatively connected to the transient temperature analysis unit, and the sensor drive circuit is communicatively connected to the infrared sensor for controlling the working state of the infrared sensor.

[0016] By adopting the above technical solution, the infrared lens receives the infrared radiation emitted by the object to be monitored, focuses and transmits the infrared radiation to the infrared sensor. The infrared sensor adopts a high-resolution infrared sensor to capture the temperature data on the surface of the object to be monitored in real time and generate a clear surface thermal map. The surface temperature acquisition unit can quickly respond to surface temperature changes and is suitable for real-time monitoring of scenarios with high temperature differences or sudden temperature fluctuations. For example, when the surface temperature of mechanical equipment rises rapidly, this unit can capture the thermal change and generate an alarm.

[0017] Optionally, the deep temperature detection unit includes a laser pulse generator, a high-sensitivity thermocouple, and a signal amplification and filtering circuit. The laser pulse generator is used to thermally excite the object to be monitored, the high-sensitivity thermocouple is used to detect the temperature change of the object to be monitored after thermal excitation, and is communicatively connected to the signal input end of the signal amplification and filtering circuit. The signal output end of the signal amplification and filtering circuit is communicatively connected to the transient temperature analysis unit.

[0018] By adopting the above technical solution, using the technology of pulsed thermal excitation and thermal diffusion analysis, the laser pulse generator is used to thermally excite the object to be monitored, and the high-sensitivity thermocouple is used to detect the temperature change of the object to be monitored after thermal excitation. The pulsed thermal excitation heats the surface layer through a short thermal pulse to detect the process of heat transfer and analyze the deep temperature data.

[0019] The deep temperature detection unit can detect potential hidden dangers (such as hot spots caused by corrosion or cracks) that may exist below the surface, and is particularly suitable for scenarios such as industrial equipment and building structures. By analyzing the thermal diffusion process, this unit can discover early deep faults and give early warnings of potential risks.

[0020] Optionally, the transient temperature analysis unit includes a high-speed multi-channel data acquisition card, a temperature data buffer, and a time series analysis processor. The data input ends of the high-speed multi-channel data acquisition card are respectively communicatively connected to the infrared sensor and the signal amplification and filtering circuit. The temperature data buffer is communicatively connected to the data output end of the high-speed multi-channel data acquisition card, and the time series analysis processor is communicatively connected to the temperature data buffer. The time series analysis processor outputs transient temperature data based on the surface temperature data and the deep temperature data of the monitored object by analyzing the time series of the temperature change.

[0021] By adopting the above technical solution, the time series analysis processor calculates the rate of temperature rise or fall by analyzing the time series of the temperature change, and identifies normal and abnormal temperature change patterns.

[0022] The transient temperature analysis unit identifies sudden temperature changes, such as rapid temperature rise caused by motor overload or severe temperature fluctuations caused by chemical reactions, by calculating the rate of temperature change. The transient temperature data, combined with the surface and deep temperature data, provides a reference in the time dimension for accurate diagnosis.

[0023] Optionally, the intelligent temperature data processing module includes a data memory, a data processor, and a graphics processor. The data memory is communicatively connected to the time series analysis processor, and the data processor and the graphics processor are respectively communicatively connected to the data memory.

[0024] By adopting the above technical solution, the intelligent temperature data processing module performs intelligent analysis and fusion on temperature data at different levels through a multi-level data processing algorithm to generate a high-precision comprehensive temperature map. The multi-layer temperature separation algorithm performs hierarchical separation processing on the surface, deep, and transient temperature data to extract temperature characteristics at different depths and times. The multi-layer temperature separation algorithm can effectively extract and separate the temperature data of each layer to ensure that the temperature information of each layer is not confused. For example, for the surface temperature change caused by a deep fault, the algorithm can separate it to avoid misjudgment. Based on the heat diffusion data of the deep temperature detection unit, the thermal diffusion coefficient is calculated to analyze the transmission rate of temperature in the monitored object. The thermal diffusion coefficient can help judge the diffusion trend and speed of temperature. For example, when the thermal diffusion coefficient increases sharply within a certain range, it may mean deep temperature anomalies or material problems. This function is particularly suitable for monitoring the internal temperature changes of power equipment and industrial machinery. The surface, deep, and transient temperature data are intelligently synthesized to generate a comprehensive temperature map that combines different temperature information. The energy data synthesis unit effectively fuses the multi-layer temperature data, so that the finally generated comprehensive temperature map contains both surface details and deep temperature change information, which is suitable for high-precision monitoring in complex environments.

[0025] Optionally, the intelligent fault diagnosis and early warning module includes an AI processor, an early warning signal output interface, and a communication interface. The AI processor is communicatively connected to the data memory, the early warning signal output interface, and the communication interface respectively.

[0026] By adopting the above technical solution, through the analysis of multi-level temperature data, temperature characteristics related to faults are extracted, including mutation points, temperature abnormal regions, temperature difference fluctuations, etc.

[0027] The fault feature extraction unit identifies abnormal temperature change points, such as sudden temperature rise or abnormal heat diffusion rate, according to the temperature characteristics of multi-level data, providing effective preliminary data for fault mode recognition.

[0028] Machine learning algorithms such as support vector machine (SVM) and neural network are used to establish a pattern model of known fault types to identify and match the fault feature data.

[0029] The fault mode recognition unit can quickly determine the fault type according to the machine learning model, automatically identify whether it is a common fault (such as temperature rise caused by short circuit, mechanical friction, etc.), and improve the diagnosis speed. The fault mode library of this unit can be learned and updated through historical data to further improve the accuracy and generalization ability of recognition.

[0030] According to the recognized fault mode and characteristics, the early warning trigger unit sets multi-level early warning levels, including prompt, warning, and emergency alarm.

[0031] When the early warning trigger unit recognizes potential fault characteristics, it triggers alarm signals of corresponding levels and can be connected to a remote monitoring system or a smartphone APP to realize real-time push of fault information, ensuring that users can take corresponding measures in time.

[0032] A fault diagnosis method based on multi-level thermal image analysis uses an infrared thermal imager based on multi-level thermal image analysis to diagnose faults of the monitored object, including the following steps: Step 1, the time series analysis processor outputs transient temperature data by analyzing the surface temperature data and the deep temperature data through time series analysis of temperature changes; Step 2, the AI processor analyzes the temperature characteristics of each layer of the transient temperature data; Step 3, the AI processor calculates the thermal diffusion coefficient and analyzes the transmission rate of temperature in the monitored object; Step 4, the AI processor extracts the fault characteristics of the monitored object based on the temperature characteristics of each layer and the thermal diffusion coefficient; Step 5, use machine learning algorithms to perform pattern recognition on the fault characteristics, match the extracted fault characteristics with the known fault mode database, and output the fault mode result; Step 6: According to the fault mode recognition result, output a preset warning action instruction through the warning signal output interface and the communication interface.

[0033] Optionally, the calculation formula for the transient temperature data in Step 1 is: ; where is the transient temperature component, t is time, x is position, is the temperature data after applying a low-pass filter; The formula for calculating the thermal diffusivity in Step 3 is: ; where is the thermal diffusivity, is the temperature change, is the time change; In Step 4, the fault feature extraction formula is: ; where G is the temperature gradient, and are the partial derivatives of temperature in the x and y directions respectively.

[0034] Optionally, the core formula of the machine algorithm in Step 5 is: ; where is the Lagrange multiplier, n is the number of samples, is the class label of the i-th sample, is the kernel function, used to calculate the inner product between two sample points and .

[0035] In summary, the present invention includes at least one of the following beneficial technical effects: The present invention can provide an infrared thermal imager and an intelligent fault diagnosis method based on multi-level thermal image analysis. By simultaneously collecting the surface layer, deep layer, and transient temperature data of the object to be monitored, more comprehensive and accurate temperature information can be obtained, improving the accuracy of monitoring. The high-resolution infrared sensor ensures the resolution and details of the surface layer temperature data, helping to identify small temperature differences. Using pulsed thermal excitation and thermal diffusion analysis techniques, the temperature condition inside the object can be detected.

[0036] The intelligent temperature data processing module can effectively fuse temperature data at different levels, generate a more accurate comprehensive temperature map, and provide a reliable data basis for subsequent analysis.

[0037] The intelligent fault diagnosis and early warning module automatically identifies and matches fault characteristics by establishing a fault mode model, reducing the dependence on manual experience and improving the efficiency and accuracy of fault diagnosis.

[0038] It can monitor temperature changes in real time and give early warnings of potential faults, which helps to take timely measures to prevent accidents. By accurately diagnosing faults, more effective maintenance plans can be formulated, reducing unnecessary downtime and maintenance costs.

[0039] The infrared thermal imager based on multi-level thermal image analysis is applicable to temperature monitoring and fault diagnosis of various industrial equipment. It can be used to detect heat loss areas in buildings, optimize energy use, provide real-time temperature monitoring in high-risk environments, and enhance operation safety. Description of the Drawings

[0040] Figure 1 It is a schematic diagram of the electrical component connection principle of the infrared thermal imager based on multi-level thermal image analysis of the present invention.

[0041] Figure 2 It is a schematic flowchart of the fault diagnosis method based on multi-level thermal image analysis of the present invention.

[0042] Description of the Reference Numerals: 1 Temperature data acquisition module; 11 Surface temperature acquisition unit; 111 Infrared lens; 112 Infrared sensor; 113 Sensor drive circuit; 12 Deep temperature detection unit; 121 Laser pulse generator; 122 High-sensitivity thermocouple; 123 Signal amplification and filtering circuit; 13 Transient temperature analysis unit; 131 High-speed multi-channel data acquisition card; 132 Temperature data buffer; 133 Time series analysis processor; 2 Intelligent temperature data processing module; 21 Data memory; 22 Data processor; 23 Graphics processor; 3 Intelligent fault diagnosis and early warning module; 31 AI processor; 32 Early warning signal output interface; 33 Communication interface; 4 Visual multi-level temperature image generation module; 41 Visualization data memory; 42 Graphics processing unit; 43 Display. Detailed Description of the Invention

[0043] The following further describes the present invention in detail with reference to the drawings.

[0044] The embodiments of the present invention disclose an infrared thermal imager and an intelligent fault diagnosis method based on multi-level thermal image analysis.

[0045] Refer to Figure 1 and Figure 2Embodiment 1, an infrared thermal imager based on multi-level thermal image analysis, comprising a multi-level temperature data acquisition module 1, an intelligent temperature data processing module 2 and an intelligent fault diagnosis and early warning module 3, wherein the multi-level temperature data acquisition module 1 comprises a surface temperature acquisition unit 11, a deep temperature detection unit 12 and a transient temperature analysis unit 13, wherein the surface temperature acquisition unit 11 captures the surface temperature data of the monitored object through a high-resolution infrared sensor; the deep temperature detection unit 12 obtains the deep temperature data below the surface of the monitored object based on pulse thermal excitation and thermal diffusion analysis technology; the data input end of the transient temperature analysis unit 13 is respectively connected to the data output end of the surface temperature acquisition unit 11 and the deep temperature detection unit 12 for communication, so as to collect the surface temperature data and the deep temperature data of the surface of the monitored object, analyze the rate of temperature rise or fall through the time series analysis of the temperature change, and output the transient temperature data; The intelligent temperature data processing module 2 is in communication connection with the transient temperature analysis unit 13, analyzes and fuses the surface temperature data, the deep temperature data and the transient temperature data through a multi-level data processing algorithm, outputs the multi-level temperature data analysis results, and generates a comprehensive temperature map based on the multi-level temperature data analysis results; The intelligent fault diagnosis and early warning module 3 is communicated with the intelligent temperature data processing module 2, and the fault-related temperature characteristics are extracted from the multi-level temperature data analysis results. The temperature characteristics include mutation points, temperature abnormality areas and temperature difference fluctuations. The machine learning algorithm is used to establish a pattern model of known fault types to identify and match the fault feature data. If there is a fault that is successfully identified and matched, a fault alarm is issued.

[0046] By simultaneously collecting the surface, deep and transient temperature data of the monitored object, more comprehensive and accurate temperature information can be obtained, improving the accuracy of monitoring. High-resolution infrared sensors ensure the resolution and details of surface temperature data, which helps to identify small temperature differences. Pulse thermal excitation and thermal diffusion analysis technology can detect the temperature conditions inside the object, which is particularly important for diagnosing internal defects or faults. The intelligent temperature data processing module 2 can quickly identify temperature anomalies by analyzing the temperature change rate, providing important clues for fault diagnosis. It can effectively integrate temperature data at different levels to generate a more accurate comprehensive temperature map, providing a reliable data basis for subsequent analysis.

[0047] The intelligent fault diagnosis and early warning module 3 can automatically identify and match fault characteristics by establishing a fault mode model, reducing dependence on manual experience and improving the efficiency and accuracy of fault diagnosis.

[0048] The system can monitor temperature changes in real time and give early warnings of potential faults, which helps to take timely measures to prevent accidents. By accurately diagnosing faults, more effective maintenance plans can be formulated, reducing unnecessary downtime and maintenance costs.

[0049] The infrared thermal imager based on multi-level thermal imaging analysis is suitable for temperature monitoring and fault diagnosis of various industrial equipment. It can be used to detect heat loss areas in buildings, optimize energy use, provide real-time temperature monitoring in high-risk environments, and enhance operation safety.

[0050] Embodiment 2 further includes a visual multi-level temperature image generation module 4. The visual multi-level temperature image generation module 4 includes a visual data memory 41, a graphics processing unit 42, and a display 43. The visual data memory 41 is communicatively connected to the intelligent temperature data processing module 2 to collect surface temperature data, deep temperature data, and transient temperature data. The graphics processing unit 42 outputs multi-level temperature image data, and the display 43 is communicatively connected to the graphics processing unit 42.

[0051] The graphics processing unit 42 generates a comprehensive multi-level temperature image through the fusion processing of temperature data at different levels, realizing an intuitive display of temperature distribution.

[0052] Color coding and layered display function: Different temperature levels are displayed with different color codings, and the temperature distribution is shown through the depth change of the image color.

[0053] The color coding can help users quickly identify abnormal temperature areas, and the layered display function can separately display surface and deep temperature information, facilitating users to more clearly observe the temperature distribution and changes.

[0054] Zooming and focus viewing function: Users can zoom in and out of the temperature image and select a focus to observe the temperature changes in a specific area.

[0055] The zooming and focus functions help users to deeply view the detailed temperature changes in the fault hot spots or potential risk areas, so as to more effectively locate the fault source and improve the analysis efficiency.

[0056] Embodiment 3, the surface temperature acquisition unit 11 includes an infrared lens 111, an infrared sensor 112, and a sensor drive circuit 113. The infrared lens 111 receives the infrared radiation emitted by the monitored object, focuses and transmits the infrared radiation to the infrared sensor 112. The data output end of the infrared sensor 112 is communicatively connected to the transient temperature analysis unit 13, and the sensor drive circuit 113 is communicatively connected to the infrared sensor 112 for controlling the working state of the infrared sensor 112.

[0057] The infrared lens 111 receives the infrared radiation emitted by the object to be monitored, focuses and transmits the infrared radiation onto the infrared sensor 112. The infrared sensor 112 uses a high-resolution infrared sensor to capture the temperature data on the surface of the object to be monitored in real time and generate a clear surface thermal map. The surface temperature acquisition unit can quickly respond to surface temperature changes and is suitable for real-time monitoring of scenarios with high temperature differences or sudden temperature fluctuations. For example, when the surface temperature of a mechanical device rises rapidly, this unit can capture the thermal change and generate an alarm.

[0058] Embodiment 4: The deep temperature detection unit 12 includes a laser pulse generator 121, a high-sensitivity thermocouple 122, and a signal amplification and filtering circuit 123. The laser pulse generator 121 is used to thermally stimulate the object to be monitored. The high-sensitivity thermocouple 122 is used to detect the temperature change of the object to be monitored after thermal stimulation and is communicatively connected to the signal input end of the signal amplification and filtering circuit 123. The signal output end of the signal amplification and filtering circuit 123 is communicatively connected to the transient temperature analysis unit 13.

[0059] Using the technology of pulsed thermal excitation and thermal diffusion analysis, the laser pulse generator 121 is used to thermally stimulate the object to be monitored. The high-sensitivity thermocouple 122 is used to detect the temperature change of the object to be monitored after thermal stimulation. The pulsed thermal excitation heats the surface layer through a short thermal pulse to detect the process of heat transfer and analyze the deep temperature data.

[0060] The deep temperature detection unit 12 can detect potential hidden dangers (such as hot spots caused by corrosion or cracks) that may exist below the surface, and is particularly suitable for scenarios such as industrial equipment and building structures. By analyzing the thermal diffusion process, this unit can discover early deep faults and give early warnings of potential risks.

[0061] Embodiment 5: The transient temperature analysis unit 13 includes a high-speed multi-channel data acquisition card 131, a temperature data buffer 132, and a time series analysis processor 133. The data input ends of the high-speed multi-channel data acquisition card 131 are respectively communicatively connected to the infrared sensor 112 and the signal amplification and filtering circuit 123. The temperature data buffer 132 is communicatively connected to the data output end of the high-speed multi-channel data acquisition card 131. The time series analysis processor 133 is communicatively connected to the temperature data buffer 132. The time series analysis processor 133 outputs transient temperature data based on the surface temperature data and deep temperature data of the monitored object surface by analyzing the time series of temperature changes.

[0062] The time series analysis processor 133 calculates the rate of temperature rise or fall by analyzing the time series of temperature changes and identifies normal and abnormal temperature change patterns.

[0063] The transient temperature analysis unit 13 identifies sudden temperature changes, such as rapid temperature rise caused by motor overload or severe temperature fluctuations caused by chemical reactions, by calculating the rate of temperature change. The transient temperature data, combined with the surface and deep temperature data, provides a reference in the time dimension for accurate diagnosis.

[0064] Embodiment 6: The intelligent temperature data processing module 2 includes a data memory 21, a data processor 22, and a graphics processor 23. The data memory 21 is communicatively connected to the time series analysis processor 133, and the data processor 22 and the graphics processor 23 are respectively communicatively connected to the data memory 21.

[0065] The intelligent temperature data processing module 2 performs intelligent analysis and fusion of temperature data at different levels through a multi-level data processing algorithm to generate a high-precision comprehensive temperature map. The multi-level temperature separation algorithm performs hierarchical separation processing on the surface, deep, and transient temperature data to extract temperature characteristics at different depths and times. The multi-level temperature separation algorithm can effectively extract and separate the temperature data of each layer to ensure that the temperature information of each layer is not confused. For example, for the surface temperature change caused by a deep fault, the algorithm can separate it to avoid misjudgment. Based on the heat diffusion data of the deep temperature detection unit, the heat diffusion coefficient is calculated to analyze the transmission rate of temperature in the monitored object. The heat diffusion coefficient can help judge the diffusion trend and speed of temperature. For example, when the heat diffusion coefficient increases sharply within a certain range, it may mean deep temperature anomalies or material problems. This function is particularly suitable for monitoring the internal temperature changes of power equipment and industrial machinery. The surface, deep, and transient temperature data are intelligently synthesized to generate a comprehensive temperature map that integrates different temperature information. The energy data synthesis unit effectively fuses the multi-level temperature data, so that the finally generated comprehensive temperature map contains both surface details and deep temperature change information, which is suitable for high-precision monitoring in complex environments.

[0066] Embodiment 7: The intelligent fault diagnosis and warning module 3 includes an AI processor 31, a warning signal output interface 32, and a communication interface 33. The AI processor 31 is respectively communicatively connected to the data memory 21, the warning signal output interface 32, and the communication interface 33.

[0067] By analyzing the multi-level temperature data, temperature characteristics related to faults are extracted, including mutation points, temperature anomaly regions, temperature difference fluctuations, etc.

[0068] The fault feature extraction unit identifies abnormal temperature change points, such as sudden temperature rise or abnormal heat diffusion rate, according to the temperature characteristics of the multi-level data, providing effective preliminary data for fault mode recognition.

[0069] Using machine learning algorithms such as support vector machines (SVM) and neural networks, a pattern model of known fault types is established to identify and match fault feature data.

[0070] The fault pattern recognition unit can quickly determine the fault type according to the machine learning model, automatically identify whether it is a common fault (such as temperature rise caused by short circuit, mechanical friction, etc.), and improve the diagnosis speed. The fault pattern library of this unit can be learned and updated through historical data to further improve the accuracy and generalization ability of recognition.

[0071] According to the identified fault pattern and features, the warning trigger unit sets multiple warning levels, including prompt, warning, and emergency alarm.

[0072] When the warning trigger unit identifies potential fault features, it triggers alarm signals of corresponding levels and can be connected to a remote monitoring system or a smartphone APP to realize real-time push of fault information and ensure that users can take corresponding measures in time.

[0073] Embodiment 8, a fault diagnosis method based on multi-level thermal image analysis, uses an infrared thermal imager based on multi-level thermal image analysis to perform fault diagnosis on the monitored object, including the following steps: Step 1, the time series analysis processor 133 outputs transient temperature data by analyzing the surface temperature data and the deep temperature data through time series analysis of temperature changes; Step 2, the AI processor 31 analyzes the temperature characteristics of each layer of the transient temperature data; Step 3, the AI processor 31 calculates the thermal diffusivity and analyzes the transmission rate of temperature in the monitored object; Step 4, the AI processor 31 extracts the fault features of the monitored object based on the temperature characteristics of each layer and the thermal diffusivity; Step 5, use machine learning algorithms to perform pattern recognition on the fault features, match the extracted fault features with the known fault pattern database, and output the fault pattern result; Step 6, according to the fault pattern recognition result, output a preset warning action instruction through the warning signal output interface 32 and the communication interface 33.

[0074] Embodiment 9, the calculation formula of the transient temperature data in Step 1 is: ; where is the transient temperature component, t is the time, x is the position, is the temperature data after applying a low-pass filter; The formula for calculating the thermal diffusivity in Step 3 is: ; where is the thermal diffusivity, is the temperature change, is the time change; In step 4, the fault feature extraction formula is: ; where G is the temperature gradient, and are the partial derivatives of the temperature in the x and y directions respectively.

[0075] In Embodiment 10, the core formula of the machine algorithm in step 5 is: ; where is the Lagrange multiplier, n is the number of samples, is the class label of the i-th sample, is the kernel function for calculating the inner product between two sample points and .

[0076] The above are all preferred embodiments of the present invention, and the protection scope of the present invention is not limited thereby. Therefore, all equivalent changes made according to the structure, shape, and principle of the present invention should be covered within the protection scope of the present invention.

Claims

1. Infrared thermal imager based on multi-level thermal image analysis, characterized by: The invention comprises a multi-layer temperature data acquisition module (1), an intelligent temperature data processing module (2) and an intelligent fault diagnosis and early warning module (3). The multi-layer temperature data acquisition module (1) comprises a surface temperature acquisition unit (11), a deep temperature detection unit (12) and a transient temperature analysis unit (13). The surface temperature acquisition unit (11) captures the surface temperature data of the monitored object through a high-resolution infrared sensor; the deep temperature detection unit (12) obtains the temperature data of the deep layer below the surface of the monitored object based on pulse thermal excitation and thermal diffusion analysis technology; the data input end of the transient temperature analysis unit (13) is respectively connected to the data output ends of the surface temperature acquisition unit (11) and the deep temperature detection unit (12), collects the surface temperature data and the deep temperature data of the surface of the monitored object, analyzes the rate of temperature rise or fall through time series analysis of temperature changes, and outputs transient temperature data; The intelligent temperature data processing module (2) is in communication connection with the transient temperature analysis unit (13), analyzes and fuses the surface temperature data, the deep temperature data and the transient temperature data through a multi-level data processing algorithm, outputs the multi-level temperature data analysis results, and generates a comprehensive temperature map based on the multi-level temperature data analysis results; The intelligent fault diagnosis and early warning module (3) is in communication connection with the intelligent temperature data processing module (2), and extracts fault-related temperature features from the multi-level temperature data analysis results, wherein the temperature features include mutation points, temperature abnormality areas, and temperature difference fluctuations. A pattern model of known fault types is established using a machine learning algorithm to identify and match the fault feature data, and a fault alarm is issued if a fault is successfully identified and matched.

2. The infrared thermal imager based on multi-level thermal image analysis according to claim 1, characterized in that: The system also includes a visualized multi-layer temperature image generation module (4), the visualized multi-layer temperature image generation module (4) including a visualized data storage device (41), a graphics processing unit (42) and a display (43), the visualized data storage device (41) being communicatively connected to the intelligent temperature data processing module (2) to collect surface temperature data, deep temperature data and transient temperature data, the graphics processing unit (42) outputting multi-layer temperature image data, and the display (43) being communicatively connected to the graphics processing unit (42).

3. The infrared thermal imager based on multi-level thermal image analysis according to claim 2, characterized in that: The surface temperature acquisition unit (11) comprises an infrared lens (111), an infrared sensor (112) and a sensor drive circuit (113); the infrared lens (111) receives infrared radiation emitted by a monitored object, focuses the infrared radiation and transmits it to the infrared sensor (112); a data output end of the infrared sensor (112) is communicatively connected to the transient temperature analysis unit (13); and the sensor drive circuit (113) is communicatively connected to the infrared sensor (112) for controlling the working state of the infrared sensor (112).

4. The infrared thermal imager based on multi-level thermal image analysis according to claim 3, characterized in that: The deep temperature detection unit (12) comprises a laser pulse generator (121), a high-sensitivity thermocouple (122) and a signal amplifying and filtering circuit (123); the laser pulse generator (121) is used to perform thermal excitation on a monitored object; the high-sensitivity thermocouple (122) is used to detect a temperature change of the monitored object after thermal excitation and is communicatively connected to a signal input end of the signal amplifying and filtering circuit (123); and the signal output end of the signal amplifying and filtering circuit (123) is communicatively connected to a transient temperature analysis unit (13).

5. The infrared thermal imager based on multi-level thermal image analysis according to claim 4, characterized in that: The transient temperature analysis unit (13) comprises a high-speed multi-channel data acquisition card (131), a temperature data buffer (132) and a time series analysis processor (133); the data input end of the high-speed multi-channel data acquisition card (131) is respectively connected to the infrared sensor (112) and the signal amplification filter circuit (123); the temperature data buffer (132) is connected to the data output end of the high-speed multi-channel data acquisition card (131); the time series analysis processor (133) is connected to the temperature data buffer (132); the time series analysis processor (133) outputs transient temperature data based on the surface temperature data and deep temperature data of the surface of the monitored object collected and monitored, through time series analysis of temperature changes.

6. The infrared thermal imager based on multi-level thermal image analysis according to claim 5, characterized in that: The intelligent temperature data processing module (2) comprises a data storage device (21), a data processor (22) and a graphics processor (23); the data storage device (21) is communicatively connected to the time series analysis processor (133); and the data processor (22) and the graphics processor (23) are respectively communicatively connected to the data storage device (21).

7. The infrared thermal imager based on multi-level thermal image analysis according to claim 6, characterized in that: The intelligent fault diagnosis and early warning module (3) comprises an AI processor (31), an early warning signal output interface (32) and a communication interface (33), wherein the AI ​​processor (31) is communicatively connected to the data storage device (21), the early warning signal output interface (32) and the communication interface (33), respectively.

8. A fault diagnosis method based on multi-level thermal image analysis, characterized in that: Using the infrared thermal imager based on multi-level thermal image analysis as described in claim 7 to perform fault diagnosis on the monitored object includes the following steps: Step 1, the time series analysis processor (133) outputs transient temperature data by analyzing the surface temperature data and the deep temperature data of the temperature change time series; Step 2, the AI ​​processor (31) analyzes the temperature characteristics of each layer of the transient temperature data; Step 3, the AI ​​processor (31) calculates the thermal diffusion coefficient and analyzes the temperature transfer rate in the monitored object; Step 4, the AI ​​processor (31) extracts the fault characteristics of the monitored object based on the temperature characteristics and thermal diffusion coefficient of each layer; Step 5: Use machine learning algorithms to perform pattern recognition on fault features, match the extracted fault features with a known fault pattern database, and output fault pattern results; Step 6: outputting a preset early warning action instruction through the early warning signal output interface (32) and the communication interface (33) according to the fault mode recognition result.

9. The fault diagnosis method based on multi-level thermal image analysis according to claim 8 is characterized in that: The calculation formula for the transient temperature data in step 1 is: ； in is the transient temperature component, t is the time, x is the position, is the temperature data after applying a low-pass filter; The formula for calculating the thermal diffusion coefficient in step 3 is: ； in is the thermal diffusivity, is the temperature change, is the time variation; In step 4, the fault feature extraction formula is: ； where G is the temperature gradient, and are the partial derivatives of temperature in the x and y directions, respectively.

10. The fault diagnosis method based on multi-level thermal image analysis according to claim 9, characterized in that: The core formula of the machine algorithm in step 5 is: ； in is the Lagrange multiplier, n is the number of samples, is the category label of the i-th sample, is the kernel function, used to calculate two sample points and The inner product between .

Images