Power grid monitoring system and method based on portable infrared detection equipment

Abstract

The present invention provides a power grid monitoring system and method based on a portable infrared detection device, relating to the field of device monitoring technology. The system comprises: an infrared detection module that generates real-time temperature reports based on the portable infrared detection device; a determination module that determines historical device report data and historical device abnormality data for each real-time device to be detected; an analysis module that determines device abnormality type data and device type feature vectors for each abnormality type of each real-time device to be detected; a construction module that constructs an abnormality diagnosis model for each real-time device to be detected; and a diagnosis module that calculates the predicted accuracy value of each real-time device to be detected, determines real-time abnormality data, and issues real-time alarms to achieve real-time monitoring of the power grid. The system can achieve intelligent real-time monitoring of the power grid, quickly and accurately identify device abnormality types, improve the accuracy and reliability of diagnosis, and achieve dynamic optimization and self-adaptation of the abnormality diagnosis model, thereby improving the efficiency of power grid operation and maintenance and ensuring the safe and stable operation of the power grid.

Description

Technical Field

[0001] The present invention relates to the technical field of equipment monitoring, and in particular to a power grid monitoring system and method based on portable infrared detection equipment. Background Art

[0002] Traditional power grid monitoring relies primarily on manual inspections and fixed monitoring equipment. Manual inspections are inefficient, labor-intensive, and difficult to detect equipment anomalies in real time. While fixed monitoring equipment can provide some real-time data, it lacks flexibility and cannot cover all critical equipment and complex environments. With the expansion of power grids and the increasing complexity of equipment, traditional monitoring methods are no longer able to meet the high safety and reliability requirements of modern power grids. In recent years, with the development of infrared detection technology, portable infrared detection equipment has gradually been adopted for power grid monitoring, but it still lacks systematic data processing and anomaly diagnosis capabilities.

[0003] Therefore, the present invention provides a system and method for monitoring power grids based on portable infrared detection devices. Summary of the Invention

[0004] The present invention aims to solve the problems existing in the prior art and provides a power grid monitoring system and method based on a portable infrared detection device.

[0005] To achieve the above object, the technical solution adopted by the present invention is as follows:

[0006] On the one hand, the present invention provides a power grid monitoring system based on a portable infrared detection device, which mainly includes the following modules:

[0007] Infrared detection module: Based on portable infrared detection equipment, it collects real-time temperature data of the real-time detection equipment set in the power grid and generates real-time temperature reports;

[0008] Determination module: obtains historical temperature reports and historical abnormal data of the power grid, and determines historical device report data and historical device abnormal data of each real-time device to be detected in the set of real-time devices to be detected;

[0009] Analysis module: Based on the historical device report data and historical device abnormality data of each real-time device to be detected, determine the device abnormality type data of each abnormal type of each real-time device to be detected, and determine the device type feature vector of each abnormal type of each real-time device to be detected;

[0010] Construction module: Build an anomaly diagnosis model for each real-time device to be detected based on the historical device report data, historical device anomaly data, and device type feature vectors of all anomaly types of each real-time device to be detected;

[0011] Diagnostic module: Based on real-time temperature reports and abnormal diagnosis models of all real-time devices to be detected, it calculates the predicted accuracy value of each real-time device to be detected and determines the real-time abnormal data. It issues real-time alarms based on the real-time abnormal data to achieve real-time monitoring of the power grid.

[0012] According to the present invention, the power grid monitoring system based on the portable infrared detection device, the infrared detection module, includes the following units:

[0013] A real-time device set unit to be detected is used to obtain the real-time monitoring requirements of the power grid at the current monitoring time point, and determine a real-time device set to be detected based on the real-time monitoring requirements. The real-time device set to be detected includes multiple real-time devices to be detected.

[0014] Detection point unit: determines multiple detection points of each real-time device to be detected based on the device structure of each real-time device to be detected in the set of real-time devices to be detected;

[0015] Real-time temperature sub-data unit: Operation and maintenance personnel carry portable infrared detection equipment to collect real-time temperature sub-data of all detection points of each real-time device to be detected in the set of real-time devices to be detected. The portable infrared detection equipment is based on a multi-process integration and consists of at least a lens, components, infrared detectors, and printed circuit boards;

[0016] Real-time temperature data unit: determines the real-time temperature data of the power grid based on the real-time temperature sub-data of all the real-time devices to be detected in the real-time device set to be detected.

[0017] The power grid monitoring system based on the portable infrared detection device provided by the present invention, the infrared detection module, further includes the following units:

[0018] Real-time temperature sub-report unit: The portable infrared detection device generates a real-time temperature sub-report for each real-time device to be detected based on the template library and the real-time temperature sub-data of each real-time device to be detected collected in real time. The template library includes multiple preset professional report templates. The real-time temperature sub-report includes at least the real-time device label to be detected, the real-time thermal image of the device to be detected, and the current monitoring time point;

[0019] Real-time temperature reporting unit: determines the real-time temperature report of the power grid based on the real-time temperature sub-reports of all the real-time devices to be detected in the real-time device set to be detected.

[0020] According to the power grid monitoring system based on the portable infrared detection device provided by the present invention, the determination module includes the following units:

[0021] Historical temperature reporting unit: obtains historical temperature reports of multiple historical monitoring time points before the current monitoring time point. The historical temperature report includes a historical temperature sub-report for each historical detection device in the historical detection device set at the historical monitoring time point. The historical temperature sub-report includes at least a historical detection device label, a historical detection device thermal image, and a historical monitoring time point.

[0022] Historical anomaly data unit: obtains historical anomaly data of multiple historical monitoring time points before the current monitoring time point. The historical anomaly data includes historical anomaly sub-data of each historical detection device in the historical detection device set at the historical monitoring time point. The historical anomaly sub-data includes the historical detection device label, the historical monitoring time point, multiple historical anomaly areas, the anomaly type of each historical anomaly area, the area of ​​the historical anomaly area and the historical anomaly level.

[0023] According to the power grid monitoring system based on the portable infrared detection device provided by the present invention, the determination module further includes the following units:

[0024] Historical device report data unit: Based on the real-time device tag of the real-time temperature sub-report of each real-time device to be detected in the real-time temperature report, extract the historical temperature sub-report corresponding to the historical detection device tag that is consistent with the real-time device tag in the historical temperature report of all historical monitoring time points, and obtain the historical device report data of each real-time device to be detected in the real-time device to be detected set;

[0025] Historical device abnormality data unit: Based on the real-time device tag of the real-time temperature sub-report of each real-time device to be detected in the real-time temperature report, the historical abnormality sub-data corresponding to the historical detection device tag that is consistent with the real-time device tag in the historical abnormal data of all historical monitoring time points are extracted to obtain the historical device abnormality data of each real-time device to be detected in the real-time device to be detected set.

[0026] According to the present invention, the power grid monitoring system based on the portable infrared detection device, the analysis module includes the following units:

[0027] Historical imaging image vector unit: performs feature extraction on the historical detection device thermal imaging image of each historical temperature sub-report in the historical device report data of each real-time device to be detected, and determines the historical imaging image vector of each historical temperature sub-report;

[0028] Historical anomaly vector unit: performs feature extraction on each historical anomaly region of each historical anomaly sub-data in the historical anomaly data of each real-time device to be detected, and determines the historical anomaly vector of each historical anomaly region of each historical anomaly sub-data in the historical anomaly data of each real-time device to be detected based on the anomaly type and anomaly level of each historical anomaly region;

[0029] A historical sub-thermal imaging unit maps and intercepts the historical detection device thermal imaging image in the historical temperature sub-report corresponding to the historical device report data of each real-time device to be detected based on each historical abnormal area of ​​each historical abnormal sub-data in the historical device abnormal data of each real-time device to be detected, and determines a historical sub-thermal imaging image for each historical abnormal area of ​​each historical abnormal sub-data in the historical device abnormal data of each real-time device to be detected;

[0030] A historical sub-imaging image vector unit is configured to extract features from a historical sub-thermal imaging image of each historical abnormal region of each historical abnormal sub-data in the historical device abnormal data of each real-time device to be detected, and determine a historical sub-imaging image vector of a historical sub-thermal imaging image of each historical abnormal region of each historical abnormal sub-data in the historical device abnormal data of each real-time device to be detected;

[0031] Equipment abnormality type data unit: Based on the abnormality types of all historical abnormal areas of all historical abnormal sub-data in the historical equipment abnormality data of each real-time equipment to be detected, the historical abnormality vectors and historical sub-imaging map vectors of each historical abnormal area of ​​each historical abnormal sub-data in the historical equipment abnormality data of each real-time equipment to be detected are extracted to determine the equipment abnormality type data of each abnormal type of each real-time equipment to be detected. The equipment abnormality type data includes the historical abnormality vectors and historical sub-imaging map vectors of multiple historical abnormal areas of multiple historical abnormal sub-data.

[0032] According to the present invention, the power grid monitoring system based on the portable infrared detection device, the analysis module includes the following units:

[0033] Correlation covariance matrix unit: calculates the correlation covariance matrix of each abnormal type of each real-time device to be detected based on the device abnormality type data of each abnormal type of each real-time device to be detected and the historical imaging map vectors of all historical temperature sub-reports in the historical device report data of each real-time device to be detected;

[0034] A time-region feature vector unit is configured to calculate a first weight and a region feature vector for each historical abnormality sub-data of each abnormality type of each real-time device to be detected based on the historical device abnormality data of each real-time device to be detected, the device abnormality type data of each abnormality type of each real-time device to be detected, and the historical imaging image vectors of all historical temperature sub-reports in the historical device report data of each real-time device to be detected, and calculate a time-region feature vector for each abnormality type of each real-time device to be detected;

[0035] Projection matrix unit: performs singular value decomposition on the associated covariance matrix of each abnormality type of each real-time device to be detected, extracts the first specified number of principal component directions of the left singular vector matrix after the singular value decomposition, and determines the projection matrix of each abnormality type of each real-time device to be detected;

[0036] First eigenvector unit: performs matrix vectorization on the projection matrix of each abnormality type of each real-time device to be detected, and determines the first eigenvector of each abnormality type of each real-time device to be detected;

[0037] Device type feature vector unit: determines a device type feature vector for each abnormal type of each real-time device to be detected based on the time region feature vector of each abnormal type of each real-time device to be detected and the first feature vector.

[0038] The power grid monitoring system based on the portable infrared detection device provided by the present invention comprises the following units:

[0039] Equipment anomaly diagnosis model unit: uses the historical equipment report data of each real-time device to be detected as the input of the equipment anomaly diagnosis model of each real-time device to be detected, and uses the historical equipment anomaly data of each real-time device to be detected as the output of the equipment anomaly diagnosis model of each real-time device to be detected;

[0040] Construction unit: Based on the device type feature vectors of all abnormal types of each real-time device to be detected, a device abnormality diagnosis model for each real-time device to be detected is constructed.

[0041] According to the present invention, the power grid monitoring system based on the portable infrared detection device and the diagnostic module include the following units:

[0042] Real-time abnormality sub-data unit: inputs the real-time temperature sub-report of each real-time device to be detected in the real-time temperature report into the device abnormality diagnosis model of the corresponding real-time device to be detected, and determines the real-time abnormality sub-data of each real-time device to be detected based on the output result of the abnormality diagnosis model. The real-time abnormality sub-data includes the real-time device label to be detected, the current monitoring time point, multiple real-time abnormal areas, the first abnormality type of each real-time abnormal area, the size of the real-time abnormal area, and the real-time abnormality level;

[0043] A real-time sub-thermal imaging unit maps and intercepts the real-time thermal imaging image of each device to be detected in the real-time temperature sub-report of each device to be detected in the real-time temperature report based on each real-time abnormal area in the real-time abnormal sub-data of each device to be detected, and determines a real-time sub-thermal imaging image of each real-time abnormal area in the real-time abnormal sub-data of each device to be detected;

[0044] A real-time sub-imaging image vector unit is configured to extract features from a real-time sub-thermal imaging image of each real-time abnormal area in the real-time abnormal sub-data of each real-time device to be detected, and determine a real-time sub-imaging image vector of each real-time abnormal area in the real-time abnormal sub-data of each real-time device to be detected;

[0045] The sub-imaging map vector consistency value unit calculates the real-time sub-imaging map vector of each real-time abnormal area in the real-time abnormal sub-data of each real-time device to be detected and the sub-imaging map vector consistency value of the device abnormality type data of each abnormal type of each real-time device to be detected based on all device abnormality type data, real-time abnormal sub-data, and real-time sub-imaging map vectors of all real-time abnormal areas in the real-time abnormal sub-data of each real-time device to be detected;

[0046] A second abnormality type unit is configured to select, from the real-time sub-imaging image vector of each real-time abnormal area in the real-time abnormal sub-data of each real-time device to be detected and the sub-imaging image vector consistency values ​​of the device abnormality type data of each abnormal type of each real-time device to be detected, the abnormality type corresponding to the largest sub-imaging image vector consistency value as the second abnormality type of each real-time abnormal area in the real-time abnormal sub-data of each real-time device to be detected;

[0047] Prediction accuracy value unit: compares the first abnormality type and the second abnormality type of each real-time abnormal area in the real-time abnormality sub-data of each real-time device to be detected, and calculates the prediction accuracy value of each real-time device to be detected based on the comparison results of all real-time abnormal areas in the real-time abnormality sub-data of each real-time device to be detected;

[0048] Optimization unit: optimizes the device anomaly diagnosis model of each real-time device to be detected whose prediction accuracy is lower than a preset threshold based on the real-time anomaly sub-data;

[0049] Alarm unit: determines real-time abnormal data based on the real-time abnormal sub-data of all real-time devices to be detected whose prediction accuracy values ​​are higher than a preset threshold, and issues a real-time alarm based on the real-time abnormal sub-data of all real-time devices to be detected in the real-time abnormal data.

[0050] On the other hand, the present invention provides a power grid monitoring system based on a portable infrared detection device, which mainly includes the following steps:

[0051] Step 1: Using a portable infrared detection device, real-time temperature data of a set of devices to be detected in the power grid is collected in real time, and a real-time temperature report is generated;

[0052] Step 2: Obtain historical temperature reports and historical abnormal data of the power grid, and determine historical device report data and historical device abnormal data of each real-time device to be detected in the set of real-time devices to be detected;

[0053] Step 3: Based on the historical device report data and historical device abnormality data of each real-time device to be detected, determine the device abnormality type data of each abnormality type of each real-time device to be detected, and determine the device type feature vector of each abnormality type of each real-time device to be detected;

[0054] Step 4: Build an anomaly diagnosis model for each real-time device to be detected based on the historical device report data, historical device anomaly data, and device type feature vectors of all anomaly types of each real-time device to be detected;

[0055] Step 4: Based on the real-time temperature report and the abnormal diagnosis model of all real-time devices to be detected, determine the real-time abnormal data, issue a real-time alarm based on the real-time abnormal data, and realize real-time monitoring of the power grid.

[0056] Compared with the existing technology, the present invention has the following beneficial effects: using a portable infrared detection device to collect real-time temperature data of the power grid in real time, generating a real-time temperature report, determining the device abnormality type data and device type feature vector for each abnormal type of each real-time device to be detected based on the acquired historical temperature reports and historical abnormal data, constructing an abnormality diagnosis model for each real-time device to be detected, and determining the real-time abnormal data and issuing a real-time alarm based on the real-time temperature report and the abnormality diagnosis models of all real-time devices to be detected, thereby achieving real-time monitoring of the power grid. This can realize intelligent real-time monitoring of the power grid, quickly discover device abnormalities, accurately identify the type of device abnormality, reduce misjudgments and missed judgments, improve the accuracy and reliability of diagnosis, and realize dynamic optimization and adaptation of the abnormality diagnosis model, ensuring that operation and maintenance personnel can respond in a timely manner, reducing failure losses, improving power grid operation and maintenance efficiency, and ensuring the safe and stable operation of the power grid. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 It is a structural diagram of a power grid monitoring system based on a portable infrared detection device provided by an embodiment of the present invention.

[0058] Figure 2 It is a flow chart of a power grid monitoring method based on a portable infrared detection device provided in an embodiment of the present invention.

[0059] Figure 3 This is the Die Attach (DA) process flow of the portable infrared detection device provided by an embodiment of the present invention.

[0060] Figure 4 This is the wire bonding process flow of the portable infrared detection equipment provided by the embodiment of the present invention.

[0061] Figure 5This is the Lens Holder Attach (LHA) process flow of the portable infrared detection device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0062] It is worth noting that the methods used in the present invention are all conventional methods unless otherwise specified; the raw materials and devices used are all conventional commercially available products, and their sources are not specifically limited unless otherwise specified.

[0063] Example 1:

[0064] The embodiment of the present invention provides a power grid monitoring system based on a portable infrared detection device, such as Figure 1 As shown, it mainly includes the following steps:

[0065] Infrared detection module: Based on portable infrared detection equipment, it collects real-time temperature data of the real-time detection equipment set in the power grid and generates real-time temperature reports;

[0066] Determination module: obtains historical temperature reports and historical abnormal data of the power grid, and determines historical device report data and historical device abnormal data of each real-time device to be detected in the set of real-time devices to be detected;

[0067] Analysis module: Based on the historical device report data and historical device abnormality data of each real-time device to be detected, determine the device abnormality type data of each abnormal type of each real-time device to be detected, and determine the device type feature vector of each abnormal type of each real-time device to be detected;

[0068] Construction module: Build an anomaly diagnosis model for each real-time device to be detected based on the historical device report data, historical device anomaly data, and device type feature vectors of all anomaly types of each real-time device to be detected;

[0069] Diagnostic module: Based on real-time temperature reports and abnormal diagnosis models of all real-time devices to be detected, it calculates the predicted accuracy value of each real-time device to be detected and determines the real-time abnormal data. It issues real-time alarms based on the real-time abnormal data to achieve real-time monitoring of the power grid.

[0070] In this embodiment, the infrared detection module is the data acquisition core of the system, collecting real-time temperature data of power grid equipment based on portable infrared detection devices. These devices can measure the surface temperature of equipment without contact and generate real-time temperature reports.

[0071] In this embodiment, the function of the determination module is to extract information related to the current device to be detected from the historical data. By obtaining historical temperature reports and historical abnormality data, the system can determine the historical device report data and historical device abnormality data of each real-time device to be detected.

[0072] In this embodiment, the analysis module is the core component of the system, responsible for extracting key features from historical data. Specifically, based on the historical device report data and historical device anomaly data for each real-time device under inspection, the analysis module determines the device anomaly type data for each anomaly type and generates a device type feature vector.

[0073] In this embodiment, the construction module builds an anomaly diagnosis model for each device under inspection based on historical device report data, historical device anomaly data, and device type feature vectors. These models learn anomaly patterns in historical data and predict the presence and type of anomaly based on real-time data. By optimizing feature vectors, the models can more accurately identify anomalies, reducing false positives and missed detections.

[0074] In this embodiment, the diagnostic module is the output component of the system. It is responsible for calculating the predicted accuracy of each device under inspection based on real-time temperature reports and anomaly diagnostic models, and identifying real-time anomaly data. Based on this real-time anomaly data, the system issues real-time alerts, notifying operations and maintenance personnel for prompt response. Furthermore, the diagnostic module features dynamic optimization capabilities, adjusting the diagnostic model based on real-time data to ensure long-term stable operation of the system.

[0075] The beneficial effects of the above technical solution include: using portable infrared detection equipment to collect real-time temperature data from the power grid in real time, generating real-time temperature reports, determining device anomaly type data and device type feature vectors for each anomaly type of each real-time device to be detected based on the acquired historical temperature reports and historical anomaly data, building an anomaly diagnosis model for each real-time device to be detected, and determining real-time anomaly data and issuing real-time alarms based on the real-time temperature reports and the anomaly diagnosis models for all real-time devices to be detected, thus achieving real-time monitoring of the power grid. This system can realize intelligent real-time monitoring of the power grid, quickly discover device anomalies, accurately identify the type of device anomaly, reduce misjudgments and missed judgments, improve the accuracy and reliability of diagnosis, and achieve dynamic optimization and adaptation of the anomaly diagnosis model, ensuring that operation and maintenance personnel can respond promptly, reducing failure losses, improving power grid operation and maintenance efficiency, and ensuring the safe and stable operation of the power grid.

[0076] Example 2:

[0077] The embodiment of the present invention provides a power grid monitoring system based on a portable infrared detection device, and an infrared detection module, including the following units:

[0078] A real-time device set unit to be detected is used to obtain the real-time monitoring requirements of the power grid at the current monitoring time point, and determine a real-time device set to be detected based on the real-time monitoring requirements. The real-time device set to be detected includes multiple real-time devices to be detected.

[0079] Detection point unit: determines multiple detection points of each real-time device to be detected based on the device structure of each real-time device to be detected in the set of real-time devices to be detected;

[0080] Real-time temperature sub-data unit: Operation and maintenance personnel carry portable infrared detection equipment to collect real-time temperature sub-data of all detection points of each real-time device to be detected in the set of real-time devices to be detected. The portable infrared detection equipment is based on a multi-process integration and consists of at least a lens, components, infrared detectors, and printed circuit boards;

[0081] Real-time temperature data unit: determines the real-time temperature data of the power grid based on the real-time temperature sub-data of all the real-time devices to be detected in the real-time device set to be detected.

[0082] In this embodiment, the real-time monitoring requirements of the power grid at the current monitoring time can be automatically generated based on the grid's operating status (such as load peaks and equipment maintenance plans). For example, during high summer temperatures, main transformers and heavily loaded lines can be automatically included in the key monitoring set. When new equipment is put into operation, the frequency of its monitoring can be temporarily increased.

[0083] This embodiment pre-stores 3D models of typical power grid equipment (e.g., transformer windings, bushings, and cooling systems) to automatically identify key inspection points (e.g., electrical connection points and heat dissipation channels). For irregular equipment (e.g., GIS switchgear), a hotspot grid method is used, dividing the equipment surface into a 5cm x 5cm inspection grid. High-density inspection points (spacing ≤ 3cm) are set for heat-prone areas (e.g., circuit breaker contacts).

[0084] In this embodiment, the portable thermal imaging intelligent temperature measurement device consists of four main parts: lens, holder, sensor, and PCBA. The sensor is the core component. Traditional infrared detectors have resolutions of 1280p and 640p, and their large arrays and high power consumption make them difficult to miniaturize. However, significant technological breakthroughs have been made in the low-power design of miniaturized infrared detectors such as 384p and 256p (power consumption has been reduced to less than 100mW). This project team intends to use the 384p miniaturized infrared detector as the core component, while the remaining lens, holder, and PCBA will be based on existing, mature components from manufacturers for the overall hardware design.

[0085] In this embodiment, the portable infrared detection equipment integrates multiple processes, including a lens + holder combined process flow. This process primarily involves the lens and holder, with their primary function being to pre-tighten them together to facilitate the subsequent LHA process. Based on empirical design, the optimal torque control range is 30 kgfcm (torque unit) to 120 kgfcm. Too high a torque can affect subsequent focusing operations, while too low a torque can result in out-of-focus conditions after focusing and before dispensing.

[0086] In this embodiment, the portable infrared detection device is based on a variety of process integration, including the Die Attach (DA) process flow: In this process flow, glue is applied to the middle of the PCBA board, and the sensor is accurately glued to the PCBA according to the optical positioning points to achieve the effect of fixation. This process focuses on the technical selection of glue. The glue selected for this project has the characteristics of high strength and strong temperature adaptability, such as Figure 3 shown.

[0087] In this embodiment, the portable infrared detection device is based on multiple process integration, including the Wire Bonding process flow: This process flow realizes the electrical conduction between the sensor and the PCBA board (the first and second solder joints in the figure below). This project uses gold wire as the medium. The diameter of the gold wire used is between 20 and 25 μm. The process indicators include: temperature, pressure, time, etc. Figure 4 shown.

[0088] In this embodiment, the portable infrared detection device is integrated based on multiple processes, and also includes the Lens Holder Attach (LHA) process flow: After the above three processes, the lens has been locked on the holder, and the sensor has achieved electrical conduction with the PCBA. In this process flow, the above two achievements are bonded and coupled. Figure 5 As shown in the figure, on the lens bonding machine, the principle is to rely on the positioning marks on the PCBA board and the holder, and the CCD cameras in three directions to complete the coincidence correction of the sensor optical center and the lens optical center. The accuracy of the thermal imaging module in this step can reach within 25um.

[0089] In this embodiment, the portable infrared detection equipment is based on the integration of multiple processes and also includes: Back-end production process flow: The back-end production process needs to be transferred to a Class 1000 dust-free workshop, in which there are two key steps: focusing and final inspection of the output image. Compared with the visible light module, the focusing process of the thermal imaging module is very different, because the imaging principle of the thermal imaging module is infrared light, and a step difference in the target object's heat is required to act as a target object, that is, a black body is required. The solution is to use a black body + metal four-bar target solution, that is, a piece of metal sheet at room temperature is blocked in front of the high-temperature / low-temperature black body, and multiple groups of four-bar target patterns are left on the metal sheet. Then, the focusing distance is set according to the focal length of the lens, and the automatic focusing function is realized in conjunction with the customized AF program for thermal imaging.

[0090] The beneficial effects of the above technical solution are: based on the real-time collection of real-time temperature data of the real-time detection equipment set in the power grid by portable infrared detection equipment, data support can be provided for generating real-time temperature reports, on-demand monitoring can be realized to avoid monitoring blindness and improve monitoring efficiency.

[0091] Example 3:

[0092] The embodiment of the present invention provides a power grid monitoring system based on a portable infrared detection device, an infrared detection module, and further includes the following units:

[0093] Real-time temperature sub-report unit: The portable infrared detection device generates a real-time temperature sub-report for each real-time device to be detected based on the template library and the real-time temperature sub-data of each real-time device to be detected collected in real time. The template library includes multiple preset professional report templates. The real-time temperature sub-report includes at least the real-time device label to be detected, the real-time thermal image of the device to be detected, and the current monitoring time point;

[0094] Real-time temperature reporting unit: determines the real-time temperature report of the power grid based on the real-time temperature sub-reports of all the real-time devices to be detected in the real-time device set to be detected.

[0095] In this embodiment, the template library presets a variety of professional report templates, each template contains: fixed format areas (such as report headers, device information columns); dynamically generated areas (such as thermal imaging images, temperature statistics tables); dynamic template updates: new templates can be pushed through the cloud; and users are supported to customize template fields (such as adding corporate logos, specific analysis items).

[0096] In this embodiment, the real-time thermal imaging image of the device to be detected may be coded using pseudo-color (eg, blue-green-yellow-red indicates increasing temperature).

[0097] The beneficial effects of the above technical solution are: generating real-time temperature reports, which can provide high-quality data support for determining real-time abnormal data, and improve monitoring efficiency and detection accuracy.

[0098] Example 4:

[0099] An embodiment of the present invention provides a power grid monitoring system based on a portable infrared detection device, wherein the determination module includes the following units:

[0100] Historical temperature reporting unit: obtains historical temperature reports of multiple historical monitoring time points before the current monitoring time point. The historical temperature report includes a historical temperature sub-report for each historical detection device in the historical detection device set at the historical monitoring time point. The historical temperature sub-report includes at least a historical detection device label, a historical detection device thermal image, and a historical monitoring time point.

[0101] Historical anomaly data unit: obtains historical anomaly data of multiple historical monitoring time points before the current monitoring time point. The historical anomaly data includes historical anomaly sub-data of each historical detection device in the historical detection device set at the historical monitoring time point. The historical anomaly sub-data includes the historical detection device label, the historical monitoring time point, multiple historical anomaly areas, the anomaly type of each historical anomaly area, the area of ​​the historical anomaly area and the historical anomaly level.

[0102] In this embodiment, temperature reports for power grid equipment are obtained at multiple historical monitoring time points. These include: Historical Monitoring Equipment Tags: used to uniquely identify each device; Historical Monitoring Equipment Thermal Images: images generated using thermal imaging technology that visually display the temperature distribution on the device surface. Thermal images clearly display hot spots on the device, helping operators quickly locate overheating areas, which is difficult to achieve with traditional temperature measurement methods; and Historical Monitoring Time Points: records the specific time when data was collected.

[0103] In this embodiment, abnormal data of power grid equipment at multiple historical monitoring time points are obtained. These data record the abnormal conditions that occurred during the past operation of the equipment and are an important basis for diagnosing and preventing faults. The historical abnormal data contains the following key information: historical detection equipment label: consistent with the equipment label in the historical temperature reporting unit, used to identify the specific equipment; historical monitoring time point: records the specific time when the abnormality occurred; multiple historical abnormal areas: clearly indicate the specific location where the abnormality occurred on the equipment. In complex equipment, different areas may have abnormalities due to different reasons; abnormality type of each historical abnormal area: classify abnormalities, such as overheating, short circuit, insulation aging, etc.; historical abnormality level: classify the severity of the abnormality, such as minor, medium, and severe. The abnormality level helps operation and maintenance personnel quickly assess the urgency of the fault, give priority to serious abnormalities, and reasonably arrange maintenance resources.

[0104] The beneficial effects of the above technical solution are: obtaining historical temperature reports and historical abnormal data of the power grid can provide data support for determining historical equipment report data and historical equipment abnormal data, improve the comprehensiveness and reliability of monitoring, reduce operation and maintenance costs, and ensure stable operation of the power grid.

[0105] Example 5:

[0106] An embodiment of the present invention provides a power grid monitoring system based on a portable infrared detection device, wherein the determination module further includes the following units:

[0107] Historical device report data unit: Based on the real-time device tag of the real-time temperature sub-report of each real-time device to be detected in the real-time temperature report, extract the historical temperature sub-report corresponding to the historical detection device tag that is consistent with the real-time device tag in the historical temperature report of all historical monitoring time points, and obtain the historical device report data of each real-time device to be detected in the real-time device to be detected set;

[0108] Historical device abnormality data unit: Based on the real-time device tag of the real-time temperature sub-report of each real-time device to be detected in the real-time temperature report, the historical abnormality sub-data corresponding to the historical detection device tag that is consistent with the real-time device tag in the historical abnormal data of all historical monitoring time points are extracted to obtain the historical device abnormality data of each real-time device to be detected in the real-time device to be detected set.

[0109] In this embodiment, the core function of the historical device report data unit is to extract historical temperature data related to the current real-time device under inspection. This process is based on the device tags in the real-time temperature report and is achieved through the following steps: the system first extracts the tag information of each device under inspection from the real-time temperature report. The system then traverses the historical temperature reports of all historical monitoring time points, searching for a historical device tag that matches the tag of the real-time device under inspection. If a match is found, the system extracts the corresponding historical temperature sub-report. Through this matching and extraction process, the system ultimately determines the historical device report data for each real-time device under inspection.

[0110] In this embodiment, the function of the historical device anomaly data unit is to extract historical anomaly data related to the current real-time device to be detected. This process is also based on the device tags in the real-time temperature report and is achieved through the following steps: the system extracts the tag information of each device to be detected from the real-time temperature report as the key basis for data retrieval. The system traverses the historical anomaly data of all historical monitoring time points, searching for historical detection device tags that are consistent with the real-time device tags to be detected. Once a match is successful, the system extracts the corresponding historical anomaly sub-data. Through the above matching and extraction process, the system ultimately determines the historical device anomaly data for each real-time device to be detected. This data includes detailed information such as the type of anomaly, abnormal area, and abnormality level that occurred during the device's past operation.

[0111] The beneficial effects of the above technical solution are: determining the historical device report data and historical device abnormal data of each real-time device to be detected in the real-time device set to be detected can achieve personalized extraction, improve the pertinence and accuracy of detection, and enhance detection precision.

[0112] Example 6:

[0113] The embodiment of the present invention provides a power grid monitoring system based on a portable infrared detection device, and an analysis module, including the following units:

[0114] Historical imaging image vector unit: performs feature extraction on the historical detection device thermal imaging image of each historical temperature sub-report in the historical device report data of each real-time device to be detected, and determines the historical imaging image vector of each historical temperature sub-report;

[0115] Historical anomaly vector unit: performs feature extraction on each historical anomaly region of each historical anomaly sub-data in the historical anomaly data of each real-time device to be detected, and determines the historical anomaly vector of each historical anomaly region of each historical anomaly sub-data in the historical anomaly data of each real-time device to be detected based on the anomaly type and anomaly level of each historical anomaly region;

[0116] A historical sub-thermal imaging unit maps and intercepts the historical detection device thermal imaging image in the historical temperature sub-report corresponding to the historical device report data of each real-time device to be detected based on each historical abnormal area of ​​each historical abnormal sub-data in the historical device abnormal data of each real-time device to be detected, and determines a historical sub-thermal imaging image for each historical abnormal area of ​​each historical abnormal sub-data in the historical device abnormal data of each real-time device to be detected;

[0117] A historical sub-imaging image vector unit is configured to extract features from a historical sub-thermal imaging image of each historical abnormal region of each historical abnormal sub-data in the historical device abnormal data of each real-time device to be detected, and determine a historical sub-imaging image vector of a historical sub-thermal imaging image of each historical abnormal region of each historical abnormal sub-data in the historical device abnormal data of each real-time device to be detected;

[0118] Equipment abnormality type data unit: Based on the abnormality types of all historical abnormal areas of all historical abnormal sub-data in the historical equipment abnormality data of each real-time equipment to be detected, the historical abnormality vectors and historical sub-imaging map vectors of each historical abnormal area of ​​each historical abnormal sub-data in the historical equipment abnormality data of each real-time equipment to be detected are extracted to determine the equipment abnormality type data of each abnormal type of each real-time equipment to be detected. The equipment abnormality type data includes the historical abnormality vectors and historical sub-imaging map vectors of multiple historical abnormal areas of multiple historical abnormal sub-data.

[0119] In this embodiment, the historical image vector unit primarily extracts features from the historical thermal images of each device under inspection within the historical device report data. The specific steps are as follows: The system analyzes the historical thermal images of each device in each historical temperature sub-report and extracts key features. The extracted features are then converted into vector form.

[0120] In this embodiment, the main function of the historical anomaly vector unit is to extract features from the historical anomaly sub-data in the historical device anomaly data of each real-time device to be detected. The specific steps are as follows: the system obtains each historical anomaly sub-data, including information such as the abnormal area, abnormality type, and abnormality level. Combined with the abnormality type and abnormality level, feature extraction is performed on each historical abnormal area. For example, for overheating anomalies, features may include the temperature range and area of ​​the abnormal area; for insulation aging anomalies, features may include texture changes in the abnormal area. The extracted features are converted into vector form. These vectors not only contain the type and level information of the anomaly, but also further refine the detailed information of the anomaly through feature extraction.

[0121] In this embodiment, the primary function of the historical sub-thermal image unit is to extract sub-thermal images related to the abnormal region from the historical thermal image. The specific steps are as follows: Based on the abnormal region information in the historical abnormal sub-data, the system locates the corresponding region in the historical thermal image. The region corresponding to the abnormal region in the thermal image is mapped and clipped to generate historical sub-thermal images. These sub-thermal images focus on the abnormal region, more clearly displaying the temperature distribution in the abnormal region. Through mapping and clipping, the system generates a historical sub-thermal image for each historical abnormal region.

[0122] In this embodiment, the main function of the historical sub-image vector unit is to extract features from historical sub-thermal images and generate historical sub-image vectors. The specific steps are as follows: the system obtains each historical sub-thermal image, which has been generated by the historical sub-thermal image unit and focuses on the abnormal area. The sub-thermal image is analyzed to extract key features. Because these images focus on the abnormal area, feature extraction can more accurately reflect the abnormal situation. The extracted features are converted into vector form. This vectorized data further refines the characteristic information of the abnormal area.

[0123] In this embodiment, the device anomaly type data unit primarily integrates historical anomaly vectors and historical sub-image vectors to determine device anomaly type data for each anomaly type for each device under real-time detection. The specific steps are as follows: Based on the anomaly type, the system integrates the historical anomaly vectors and historical sub-image vectors for all historical anomaly sub-data. This generates device anomaly type data for each anomaly type. This data includes historical anomaly vectors and historical sub-image vectors for multiple historical anomaly regions across multiple historical anomaly sub-data.

[0124] In this embodiment, the historical imaging image vector may include the global average temperature, global temperature variance, global maximum temperature, global temperature gradient distribution, and global temperature peak value of the historical detection device thermal imaging image of the corresponding historical temperature sub-report in the historical device report data.

[0125] In this embodiment, the historical anomaly vector may include an anomaly type, anomaly level, and area complexity of the historical anomaly region of the historical anomaly sub-data in the corresponding historical device anomaly data.

[0126] In this embodiment, the historical sub-image vector may include the local average temperature, local temperature variance, local maximum temperature, local temperature gradient distribution, and local temperature peak value of the historical abnormal region of the historical abnormal sub-data in the corresponding historical device abnormal data.

[0127] The beneficial effects of the above technical solution are: based on the historical device report data and historical device abnormality data of each real-time device to be detected, the device abnormality type data of each abnormal type of each real-time device to be detected is determined, and refined vital signs extraction can be performed to improve monitoring accuracy, and multi-dimensional data fusion can be achieved to enhance the comprehensiveness of monitoring.

[0128] Example 7:

[0129] The embodiment of the present invention provides a power grid monitoring system based on a portable infrared detection device, and an analysis module, including the following units:

[0130] Correlation covariance matrix unit: calculates the correlation covariance matrix of each abnormal type of each real-time device to be detected based on the device abnormality type data of each abnormal type of each real-time device to be detected and the historical imaging map vectors of all historical temperature sub-reports in the historical device report data of each real-time device to be detected;

[0131] A time-region feature vector unit is configured to calculate a first weight and a region feature vector for each historical abnormality sub-data of each abnormality type of each real-time device to be detected based on the historical device abnormality data of each real-time device to be detected, the device abnormality type data of each abnormality type of each real-time device to be detected, and the historical imaging image vectors of all historical temperature sub-reports in the historical device report data of each real-time device to be detected, and calculate a time-region feature vector for each abnormality type of each real-time device to be detected;

[0132] Projection matrix unit: performs singular value decomposition on the associated covariance matrix of each abnormality type of each real-time device to be detected, extracts the first specified number of principal component directions of the left singular vector matrix after the singular value decomposition, and determines the projection matrix of each abnormality type of each real-time device to be detected;

[0133] First eigenvector unit: performs matrix vectorization on the projection matrix of each abnormality type of each real-time device to be detected, and determines the first eigenvector of each abnormality type of each real-time device to be detected;

[0134] Device type feature vector unit: determines a device type feature vector for each abnormal type of each real-time device to be detected based on the time region feature vector of each abnormal type of each real-time device to be detected and the first feature vector.

[0135] In this embodiment, the correlation covariance matrix unit calculates the correlation covariance matrix of each abnormality type of each real-time device to be detected based on the device abnormality type data of each abnormality type of each real-time device to be detected and the historical imaging image vectors of all historical temperature sub-reports in the historical device report data of each real-time device to be detected. The calculation formula of the correlation covariance matrix can be:

[0136] ;

[0137] ;

[0138] in, represents the correlation covariance matrix of the jth abnormality type of the i-th real-time device to be detected, ijN2 represents the number of historical abnormal sub-data in the device abnormality type data of the jth abnormality type of the i-th real-time device to be detected, represents the number of historical abnormal regions in the t-th historical abnormal sub-data of the device abnormality type data of the j-th abnormality type of the i-th real-time device to be detected, The historical sub-image image vector representing the historical sub-thermal image of the k-th historical abnormal area of ​​the t-th historical abnormal sub-data in the device abnormality type data of the j-th abnormality type of the i-th real-time device to be detected, The historical imaging image vector representing the thermal imaging image of the historical detection device of the historical temperature sub-report in the historical device report data corresponding to the t-th historical abnormality sub-data of the device abnormality type data of the j-th abnormality type of the i-th real-time device to be detected, The average historical sub-image vector of the historical sub-thermal images of all historical abnormal areas of the t-th historical abnormal sub-data in the device abnormality type data of the j-th abnormality type of the i-th real-time device to be detected, The average historical image vector of the historical detection device thermal images of all historical temperature sub-reports in the historical device report data corresponding to the t-th historical abnormality sub-data of the device abnormality type data of the j-th abnormality type of the i-th real-time device to be detected, A historical sub-imaging image vector representing a historical sub-thermal image of the kth historical abnormality region of the tth historical abnormality sub-data in the device abnormality type data of the jth abnormality type of the i-th real-time device to be detected, and an average historical sub-imaging image vector representing historical sub-thermal image of all historical abnormality regions of the tth historical abnormality sub-data in the device abnormality type data of the jth abnormality type of the i-th real-time device to be detected, based on a first indicator function of the j-th abnormality type, Indicates the jth exception type, Indicates the abnormality type of the kth historical abnormality area of ​​the tth historical abnormality sub-data in the device abnormality type data of the jth abnormality type of the i-th real-time device to be detected.

[0139] In this embodiment, Represents the number of all historical abnormal regions in all historical abnormal sub-data of the device abnormality type data of the jth abnormality type of the i-th real-time device to be detected.

[0140] In this embodiment, the correlation covariance matrix Reflects the association between the historical sub-imaging image vector of the historical sub-thermal imaging image of all historical abnormal areas of all historical abnormal sub-data in the device abnormality type data of the jth abnormality type of the i-th real-time device to be detected, and the historical imaging image vector of the historical detection device thermal imaging image of all historical temperature sub-reports in the historical device report data corresponding to all historical abnormal sub-data of the jth abnormality type of the i-th real-time device to be detected.

[0141] In this embodiment, The matrix dimension is based on the historical sub-image vector The vector dimension and historical imaging vector The vector dimension determines, for example, the historical sub-image vector The vector dimension is D1, the historical imaging vector The vector dimension is D2, then The matrix dimensions are D1×D2.

[0142] In this embodiment, Represents the historical imaging vector Peace and history imaging vector The transposed vector of the difference.

[0143] In this embodiment, the time-region feature vector unit calculates a first weight and a region feature vector for each historical abnormality sub-data of each abnormality type of each real-time device to be detected based on the historical device abnormality data of each real-time device to be detected, the device abnormality type data of each abnormality type of each real-time device to be detected, and the historical imaging image vectors of all historical temperature sub-reports in the historical device report data of each real-time device to be detected, and calculates a time-region feature vector for each abnormality type of each real-time device to be detected. The calculation formula of the time-region feature vector may be:

[0144] ;

[0145] ;

[0146] ;

[0147] ;

[0148] ;

[0149] ;

[0150] in, The deviation value of the historical sub-image image vector of the historical sub-thermal image of the k-th historical abnormal area of ​​the t-th historical abnormal sub-data in the device abnormality type data of the j-th abnormality type of the i-th real-time device to be detected based on the historical image image vector of the historical detection device thermal image corresponding to the t-th historical abnormal sub-data, The first weight of the t-th historical abnormal sub-data of the device abnormality type data of the j-th abnormality type of the i-th real-time device to be detected, Indicates the anomaly type of the bth historical anomaly area of ​​the ath historical anomaly sub-data in the historical device anomaly data of the i-th real-time device to be detected, represents the number of historical abnormal regions of the ath historical abnormal sub-data in the historical device abnormal data of the i-th real-time device to be detected, represents the time decay factor, Indicates the current monitoring time point. represents the historical monitoring time point of the ath historical abnormality sub-data in the historical device abnormality data of the i-th real-time device to be detected; represents the second weight of the kth historical abnormal region of the tth historical abnormal sub-data in the device abnormality type data of the jth abnormal type of the i-th real-time device to be detected, The area of ​​the historical abnormal region of the kth historical abnormal region of the tth historical abnormal sub-data in the device abnormality type data of the jth abnormal type of the i-th real-time device to be detected is represented. The area of ​​the historical detection device thermal image of the historical temperature sub-report in the historical device report data corresponding to the t-th historical abnormality sub-data of the device abnormality type data of the j-th abnormality type of the i-th real-time device to be detected, The regional feature vector representing all historical abnormal regions of the t-th historical abnormal sub-data in the device abnormality type data of the j-th abnormal type of the i-th real-time device to be detected, The second weight of the kth historical abnormal area of ​​the tth historical abnormal sub-data in the device abnormality type data of the jth abnormality type of the i-th real-time device to be detected is based on the second indicator function of the jth abnormality type, A historical anomaly vector representing the kth historical anomaly region of the tth historical anomaly sub-data in the device anomaly type data of the jth anomaly type of the i-th real-time device to be detected; Represents the time domain feature vector of all historical abnormal sub-data in the device abnormality type data of the jth abnormality type of the i-th real-time device to be detected.

[0151] In this embodiment, Measure the historical sub-image vector and historical imaging vectors The distance between The larger the value is, the greater the difference between the historical sub-thermal image of the kth historical abnormal area of ​​the tth historical abnormal sub-data in the device abnormality type data of the jth abnormal type of the i-th real-time device to be detected and the historical imaging image vector of the historical detection device thermal image of the historical temperature sub-report in the historical device report data corresponding to the tth historical abnormal sub-data of the device abnormality type data of the jth abnormal type of the i-th real-time device to be detected, which indicates that the degree of abnormality in the historical abnormal area is higher.

[0152] In this embodiment, Represents historical anomaly vector , historical sub-image vector and deviation values The stitching vector.

[0153] In this embodiment, Represents the regional feature vector and historical imaging vectors The stitching vector.

[0154] In this embodiment, the main function of the projection matrix unit is to extract key features through singular value decomposition and generate a projection matrix. The specific steps are as follows: singular value decomposition is performed on the associated covariance matrix of each abnormal type of each real-time device to be detected. Singular value decomposition is a commonly used dimensionality reduction technique that can decompose a matrix into singular values, left singular vectors, and right singular vectors. The first specified number of principal component directions are extracted from the left singular vector matrix after singular value decomposition. The principal component direction reflects the main change trend of the data and can effectively reduce the dimension and extract key features. The extracted principal component directions are combined into a projection matrix. The projection matrix is ​​used to project the original data into a low-dimensional space.

[0155] In this embodiment, the primary function of the first eigenvector unit is to perform matrix vectorization on the projection matrix to generate the first eigenvector. The specific steps are as follows: Matrix vectorization is performed on the projection matrix for each anomaly type of each real-time device to be detected. Matrix vectorization is the process of expanding a matrix into a one-dimensional vector. Through matrix vectorization, the first eigenvector for each anomaly type is generated. The first eigenvector is a compact representation of the projection matrix, which can further simplify the data.

[0156] In this embodiment, the device type feature vector unit primarily integrates the time and region feature vectors and the first feature vector to generate a device type feature vector. The specific steps are as follows: The system combines the time and region feature vectors and the first feature vector for each anomaly type. Through this concatenation, a device type feature vector is generated for each anomaly type. The device type feature vector is the final feature vector that comprehensively considers time, region, and principal component features.

[0157] The beneficial effects of the above technical solution are: determining the device type feature vector of each abnormal type of each real-time device to be detected can achieve accurate abnormal correlation analysis and multi-dimensional feature integration, improve the comprehensiveness and accuracy of monitoring, and improve monitoring accuracy.

[0158] Example 8:

[0159] The embodiment of the present invention provides a power grid monitoring system based on a portable infrared detection device, and the building blocks include the following units:

[0160] Equipment anomaly diagnosis model unit: uses the historical equipment report data of each real-time device to be detected as the input of the equipment anomaly diagnosis model of each real-time device to be detected, and uses the historical equipment anomaly data of each real-time device to be detected as the output of the equipment anomaly diagnosis model of each real-time device to be detected;

[0161] Construction unit: Based on the device type feature vectors of all abnormal types of each real-time device to be detected, a device abnormality diagnosis model for each real-time device to be detected is constructed.

[0162] In this embodiment, historical device anomaly data for each device under real-time monitoring serves as the output of the device anomaly diagnosis model. This historical device anomaly data records anomalies that occurred during past operation, including detailed information such as anomaly type, anomaly area, and anomaly level. This data serves as the target for model learning and is used to train the model to identify anomaly patterns.

[0163] This example provides a clear data foundation for building a device anomaly diagnosis model by clarifying the input and output data. The input data provides background information about the device's operation, while the output data provides labels for anomalies, enabling the model to learn the mapping between device operating status and anomalies.

[0164] In this embodiment, the construction unit's primary function is to construct a device anomaly diagnosis model for each device under real-time detection based on the device type feature vector. The specific steps are as follows: The system constructs a device anomaly diagnosis model based on the device type feature vector for each anomaly type for each device under real-time detection. The model construction process learns the mapping relationship between input feature vectors and output anomaly data, enabling the model to predict the presence and type of anomaly based on the input device operating data.

[0165] The beneficial effects of the above technical solution are: based on the historical device report data, historical device abnormality data and device type feature vectors of all abnormal types of each real-time device to be detected, an abnormality diagnosis model for each real-time device to be detected is constructed, which can improve the accuracy and generalization ability of the abnormality diagnosis model and enhance the ability to identify different abnormality types.

[0166] Example 9:

[0167] The embodiment of the present invention provides a power grid monitoring system based on a portable infrared detection device, and the building blocks include the following units:

[0168] Real-time abnormality sub-data unit: inputs the real-time temperature sub-report of each real-time device to be detected in the real-time temperature report into the device abnormality diagnosis model of the corresponding real-time device to be detected, and determines the real-time abnormality sub-data of each real-time device to be detected based on the output result of the abnormality diagnosis model. The real-time abnormality sub-data includes the real-time device label to be detected, the current monitoring time point, multiple real-time abnormal areas, the first abnormality type of each real-time abnormal area, the size of the real-time abnormal area, and the real-time abnormality level;

[0169] A real-time sub-thermal imaging unit maps and intercepts the real-time thermal imaging image of each device to be detected in the real-time temperature sub-report of each device to be detected in the real-time temperature report based on each real-time abnormal area in the real-time abnormal sub-data of each device to be detected, and determines a real-time sub-thermal imaging image of each real-time abnormal area in the real-time abnormal sub-data of each device to be detected;

[0170] A real-time sub-imaging image vector unit is configured to extract features from a real-time sub-thermal imaging image of each real-time abnormal area in the real-time abnormal sub-data of each real-time device to be detected, and determine a real-time sub-imaging image vector of each real-time abnormal area in the real-time abnormal sub-data of each real-time device to be detected;

[0171] The sub-imaging map vector consistency value unit calculates the real-time sub-imaging map vector of each real-time abnormal area in the real-time abnormal sub-data of each real-time device to be detected and the sub-imaging map vector consistency value of the device abnormality type data of each abnormal type of each real-time device to be detected based on all device abnormality type data, real-time abnormal sub-data, and real-time sub-imaging map vectors of all real-time abnormal areas in the real-time abnormal sub-data of each real-time device to be detected;

[0172] A second abnormality type unit is configured to select, from the real-time sub-imaging image vector of each real-time abnormal area in the real-time abnormal sub-data of each real-time device to be detected and the sub-imaging image vector consistency values ​​of the device abnormality type data of each abnormal type of each real-time device to be detected, the abnormality type corresponding to the largest sub-imaging image vector consistency value as the second abnormality type of each real-time abnormal area in the real-time abnormal sub-data of each real-time device to be detected;

[0173] Prediction accuracy value unit: compares the first abnormality type and the second abnormality type of each real-time abnormal area in the real-time abnormality sub-data of each real-time device to be detected, and calculates the prediction accuracy value of each real-time device to be detected based on the comparison results of all real-time abnormal areas in the real-time abnormality sub-data of each real-time device to be detected;

[0174] Optimization unit: optimizes the device anomaly diagnosis model of each real-time device to be detected whose prediction accuracy is lower than a preset threshold based on the real-time anomaly sub-data;

[0175] Alarm unit: determines real-time abnormal data based on the real-time abnormal sub-data of all real-time devices to be detected whose prediction accuracy values ​​are higher than a preset threshold, and issues a real-time alarm based on the real-time abnormal sub-data of all real-time devices to be detected in the real-time abnormal data.

[0176] In this embodiment, the core function of the real-time anomaly sub-data unit is to input the data in the real-time temperature report into the device anomaly diagnosis model and output the real-time anomaly sub-data. The specific steps are as follows: input the real-time temperature sub-report of each real-time device to be detected in the real-time temperature report into the corresponding device anomaly diagnosis model. Based on the output result of the device anomaly diagnosis model, the real-time anomaly sub-data of each real-time device to be detected is determined. The real-time anomaly sub-data includes the following detailed information: real-time device label to be detected: used to uniquely identify the device; current monitoring time point: records the specific time of data collection; multiple real-time anomaly areas: identify the specific location where the anomaly occurs on the device; the first anomaly type of each real-time anomaly area; real-time anomaly area size: the range size of the anomaly area; real-time anomaly level: to grade the severity of the anomaly.

[0177] In this embodiment, the main function of the real-time sub-thermal imaging unit is to extract corresponding sub-thermal images from the real-time thermal image based on the real-time abnormal area. The specific steps are as follows: Based on the real-time abnormal area information in the real-time abnormal sub-data, the corresponding area is located in the real-time thermal image. The real-time thermal image is mapped and clipped to extract the sub-thermal images corresponding to the abnormal area. These sub-thermal images focus on the abnormal area, more clearly displaying the temperature distribution in the abnormal area. Through the mapping and clipping operations, a real-time sub-thermal image is generated for each real-time abnormal area.

[0178] In this embodiment, the real-time sub-image vector unit primarily extracts features from real-time sub-thermal images and generates real-time sub-image vectors. The specific steps are as follows: Receive a real-time sub-thermal image for each abnormal region; Analyze the real-time sub-thermal image to extract key features; and Convert the extracted features into vector form.

[0179] In this embodiment, the sub-imaging image vector consistency value unit calculates the real-time sub-imaging image vector of each real-time abnormal area in the real-time abnormal sub-data of each real-time device to be detected, and the sub-imaging image vector consistency value of the device abnormality type data of each abnormal type of each real-time device to be detected based on all device abnormality type data, real-time abnormal sub-data, and real-time sub-imaging image vectors of all real-time abnormal areas in the real-time abnormal sub-data of each real-time device to be detected. The calculation formula of the sub-imaging image vector consistency value can be expressed as:

[0180] ;

[0181] ;

[0182] ;

[0183] ;

[0184] in, The sub-imaging image vector consistency value representing the real-time sub-imaging image vector of the p-th real-time abnormal area in the real-time abnormal sub-data of the i-th real-time device to be detected and the historical sub-imaging image vectors of all historical abnormal areas in all historical abnormal sub-data of the device abnormality type data of the j-th abnormal type of the i-th real-time device to be detected, represents the real-time sub-imaging image vector of the p-th real-time abnormal area in the real-time abnormal sub-data of the i-th real-time device to be detected, The historical sub-image vector representing the historical sub-thermal image of the k-th historical abnormal area of ​​the t-th historical abnormal sub-data in the device abnormality type data of the j-th abnormality type of the i-th real-time device to be detected, based on the third indicator function of the j-th abnormality type, A historical sub-image image vector representing a historical sub-thermal image of the kth historical abnormal area of ​​the tth historical abnormal sub-data in the device abnormality type data of the jth abnormality type of the i-th real-time device to be detected, based on the fourth indicator function of the jth abnormality type, The similarity value between the real-time sub-image vector of the p-th real-time abnormal area in the real-time abnormal sub-data of the i-th real-time device to be detected and the historical sub-image vector of the k-th historical abnormal area in the t-th historical abnormal sub-data of the device abnormality type data of the j-th abnormal type of the i-th real-time device to be detected, It represents the probability of occurrence of the jth abnormality type among all the device abnormality type data of the i-th real-time device to be detected.

[0185] In this embodiment, The number of abnormality types of all historical abnormality areas of all historical abnormality sub-data in the device abnormality type data of the jth abnormality type of the i-th real-time device to be detected is equal to the number of the jth abnormality type.

[0186] In this embodiment, Represents the number of all historical abnormal regions of all historical abnormal sub-data in the device abnormality type data of all abnormal types of the i-th real-time device to be detected.

[0187] In this embodiment, Represents the similarity value between the real-time sub-imaging map vector of the p-th real-time abnormal area in the real-time abnormal sub-data of the i-th real-time device to be detected and the historical sub-imaging map vectors of all historical abnormal areas in all historical abnormal sub-data of the device abnormality type data of the j-th abnormal type of the i-th real-time device to be detected.

[0188] In this embodiment, Represents the average similarity value between the real-time sub-imaging map vector of the p-th real-time abnormal area in the real-time abnormal sub-data of the i-th real-time device to be detected and the historical sub-imaging map vectors of all historical abnormal areas in all historical abnormal sub-data of the device abnormality type data of the j-th abnormal type of the i-th real-time device to be detected.

[0189] In this embodiment, the prediction accuracy unit calculates the prediction accuracy of each real-time device to be detected based on the comparison results of all real-time abnormal areas in the real-time abnormal sub-data of each real-time device to be detected. The calculation formula of the prediction accuracy can be:

[0190] ;

[0191] ;

[0192] ;

[0193] ;

[0194] in, represents the predicted accuracy value of the i-th real-time device to be detected, represents the first abnormality type of the pth real-time abnormal area in the real-time abnormality sub-data of the i-th real-time device to be detected, represents the second abnormality type of the pth real-time abnormal area in the real-time abnormality sub-data of the i-th real-time device to be detected, The comparison result of the pth real-time abnormal area in the real-time abnormal sub-data of the i-th real-time device to be detected, iN4 represents the number of real-time abnormal areas in the real-time abnormal sub-data of the i-th real-time device to be detected, represents the probability of occurrence of the gth abnormality level of the jth abnormality type in all device abnormality type data of the i-th real-time device to be detected, The historical abnormality level of the kth historical abnormality area of ​​the tth historical abnormality sub-data in the device abnormality type data of the jth abnormality type of the i-th real-time device to be detected, represents the gth abnormality level of the jth abnormality type, The real-time abnormality level of the p-th real-time abnormal area in the real-time abnormality sub-data of the i-th real-time device to be detected, jN5 represents the number of abnormality levels of the j-th abnormality type, The abnormality level indicator function represents the p-th real-time abnormal area in the real-time abnormality sub-data of the i-th real-time device to be detected.

[0195] In this embodiment, the optimization unit's primary function is to dynamically optimize the device anomaly diagnosis model based on the predicted accuracy. The specific steps are as follows: Identify devices whose predicted accuracy falls below a preset threshold. Optimize the anomaly diagnosis model for these devices based on real-time anomaly sub-data. This optimization process may include adjusting model parameters, updating the model structure, or retraining the model. Optimization result feedback: The optimized model is reapplied to real-time monitoring to ensure it remains optimal.

[0196] In this embodiment, the alarm unit's primary function is to issue real-time alarms based on high-accuracy real-time anomaly data. The specific steps are as follows: High-accuracy screening: Filters devices with predicted accuracy values ​​exceeding a preset threshold. Based on the real-time anomaly sub-data from these devices, the real-time anomaly data is determined. A real-time alarm is issued based on the real-time anomaly data, notifying operations and maintenance personnel to respond promptly.

[0197] The beneficial effects of this technical solution include: Based on real-time temperature reports and anomaly diagnosis models for all real-time devices to be detected, the predicted accuracy value for each device to be detected is calculated, and real-time anomaly data is determined. Real-time alerts are issued based on this real-time anomaly data, enabling real-time monitoring of the power grid. This allows for accurate real-time anomaly identification, improving the real-time and accuracy of power grid monitoring. Double verification of anomaly types reduces misjudgments, enhancing diagnostic reliability, improving the adaptive capabilities of monitoring, reducing failure losses, and ensuring stable power grid operation.

[0198] Example 10:

[0199] The embodiment of the present invention provides a power grid monitoring method based on a portable infrared detection device, such as Figure 2 As shown, it mainly includes the following steps:

[0200] Step 1: Using a portable infrared detection device, real-time temperature data of a set of devices to be detected in the power grid is collected in real time, and a real-time temperature report is generated;

[0201] Step 2: Obtain historical temperature reports and historical abnormal data of the power grid, and determine historical device report data and historical device abnormal data of each real-time device to be detected in the set of real-time devices to be detected;

[0202] Step 3: Based on the historical device report data and historical device abnormality data of each real-time device to be detected, determine the device abnormality type data of each abnormality type of each real-time device to be detected, and determine the device type feature vector of each abnormality type of each real-time device to be detected;

[0203] Step 4: Build an anomaly diagnosis model for each real-time device to be detected based on the historical device report data, historical device anomaly data, and device type feature vectors of all anomaly types of each real-time device to be detected;

[0204] Step 4: Based on the real-time temperature report and the abnormal diagnosis model of all real-time devices to be detected, determine the real-time abnormal data, issue a real-time alarm based on the real-time abnormal data, and realize real-time monitoring of the power grid.

[0205] The beneficial effects of the above technical solution include: using portable infrared detection equipment to collect real-time temperature data from the power grid in real time, generating real-time temperature reports, determining device anomaly type data and device type feature vectors for each anomaly type of each real-time device to be detected based on the acquired historical temperature reports and historical anomaly data, building an anomaly diagnosis model for each real-time device to be detected, and determining real-time anomaly data and issuing real-time alarms based on the real-time temperature reports and the anomaly diagnosis models for all real-time devices to be detected, thus achieving real-time monitoring of the power grid. This system can realize intelligent real-time monitoring of the power grid, quickly discover device anomalies, accurately identify the type of device anomaly, reduce misjudgments and missed judgments, improve the accuracy and reliability of diagnosis, and achieve dynamic optimization and adaptation of the anomaly diagnosis model, ensuring that operation and maintenance personnel can respond promptly, reducing failure losses, improving power grid operation and maintenance efficiency, and ensuring the safe and stable operation of the power grid.

[0206] Finally, it should be noted that the above content is only used to illustrate the technical solution of the present invention, rather than to limit the scope of protection of the present invention. Simple modifications or equivalent substitutions of the technical solution of the present invention by ordinary technicians in this field do not deviate from the essence and scope of the technical solution of the present invention.

Claims

1. A power grid monitoring system based on portable infrared detection equipment, characterized by: Includes the following modules: Infrared detection module: Based on portable infrared detection equipment, it collects real-time temperature data of the real-time detection equipment set in the power grid and generates real-time temperature reports; Determination module: obtains historical temperature reports and historical abnormal data of the power grid, and determines historical device report data and historical device abnormal data of each real-time device to be detected in the set of real-time devices to be detected; Analysis module: Based on the historical device report data and historical device abnormality data of each real-time device to be detected, determine the device abnormality type data of each abnormal type of each real-time device to be detected, and determine the device type feature vector of each abnormal type of each real-time device to be detected; Construction module: Build an anomaly diagnosis model for each real-time device to be detected based on the historical device report data, historical device anomaly data, and device type feature vectors of all anomaly types of each real-time device to be detected; Diagnostic module: Based on real-time temperature reports and abnormal diagnosis models of all real-time devices to be detected, it calculates the predicted accuracy value of each real-time device to be detected and determines real-time abnormal data. It issues real-time alarms based on real-time abnormal data to achieve real-time monitoring of the power grid; The diagnostic module includes: Real-time abnormality sub-data unit: inputs the real-time temperature sub-report of each real-time device to be detected in the real-time temperature report into the device abnormality diagnosis model of the corresponding real-time device to be detected, and determines the real-time abnormality sub-data of each real-time device to be detected based on the output result of the abnormality diagnosis model; A real-time sub-thermal imaging unit maps and intercepts the real-time thermal imaging image of each device to be detected in the real-time temperature sub-report of each device to be detected in the real-time temperature report based on each real-time abnormal area in the real-time abnormal sub-data of each device to be detected, and determines a real-time sub-thermal imaging image of each real-time abnormal area in the real-time abnormal sub-data of each device to be detected; A real-time sub-imaging image vector unit is configured to extract features from a real-time sub-thermal imaging image of each real-time abnormal area in the real-time abnormal sub-data of each real-time device to be detected, and determine a real-time sub-imaging image vector of each real-time abnormal area in the real-time abnormal sub-data of each real-time device to be detected; Sub-imaging map vector consistency value unit: Based on all device abnormality type data, real-time abnormality sub-data and real-time abnormality areas in the real-time abnormality sub-data of each real-time device to be detected, calculate the real-time sub-imaging map vector of each real-time abnormality area in the real-time abnormality sub-data of each real-time device to be detected, and the sub-imaging map vector consistency value of the device abnormality type data of each abnormal type of each real-time device to be detected.

2. The power grid monitoring system based on portable infrared detection equipment according to claim 1, characterized in that: Infrared detection module, including the following units: A real-time device set unit to be detected is used to obtain the real-time monitoring requirements of the power grid at the current monitoring time point, and determine a real-time device set to be detected based on the real-time monitoring requirements. The real-time device set to be detected includes multiple real-time devices to be detected. Detection point unit: determines multiple detection points of each real-time device to be detected based on the device structure of each real-time device to be detected in the set of real-time devices to be detected; Real-time temperature sub-data unit: Operation and maintenance personnel carry portable infrared detection equipment to collect real-time temperature sub-data of all detection points of each real-time device to be detected in the set of real-time devices to be detected. The portable infrared detection equipment is based on a multi-process integration and consists of at least a lens, components, infrared detectors, and printed circuit boards; Real-time temperature data unit: determines the real-time temperature data of the power grid based on the real-time temperature sub-data of all the real-time devices to be detected in the real-time device set to be detected.

3. The power grid monitoring system based on portable infrared detection equipment according to claim 2, characterized in that: The infrared detection module also includes the following units: Real-time temperature sub-report unit: The portable infrared detection device generates a real-time temperature sub-report for each real-time device to be detected based on the template library and the real-time temperature sub-data of each real-time device to be detected collected in real time. The template library includes multiple preset professional report templates. The real-time temperature sub-report includes at least the real-time device label to be detected, the real-time thermal image of the device to be detected, and the current monitoring time point; Real-time temperature reporting unit: determines the real-time temperature report of the power grid based on the real-time temperature sub-reports of all the real-time devices to be detected in the real-time device set to be detected.

4. The power grid monitoring system based on portable infrared detection equipment according to claim 1, characterized in that: Determine the module, including the following units: Historical temperature reporting unit: obtains historical temperature reports of multiple historical monitoring time points before the current monitoring time point. The historical temperature report includes a historical temperature sub-report for each historical detection device in the historical detection device set at the historical monitoring time point. The historical temperature sub-report includes at least a historical detection device label, a historical detection device thermal image, and a historical monitoring time point. Historical anomaly data unit: obtains historical anomaly data of multiple historical monitoring time points before the current monitoring time point. The historical anomaly data includes historical anomaly sub-data of each historical detection device in the historical detection device set at the historical monitoring time point. The historical anomaly sub-data includes the historical detection device label, the historical monitoring time point, multiple historical anomaly areas, the anomaly type of each historical anomaly area, the area of ​​the historical anomaly area and the historical anomaly level.

5. The power grid monitoring system based on portable infrared detection equipment according to claim 4 is characterized in that: The determination module also includes the following units: Historical device report data unit: Based on the real-time device tag of the real-time temperature sub-report of each real-time device to be detected in the real-time temperature report, extract the historical temperature sub-report corresponding to the historical detection device tag that is consistent with the real-time device tag in the historical temperature report of all historical monitoring time points, and obtain the historical device report data of each real-time device to be detected in the real-time device to be detected set; Historical device abnormality data unit: Based on the real-time device tag of the real-time temperature sub-report of each real-time device to be detected in the real-time temperature report, the historical abnormality sub-data corresponding to the historical detection device tag that is consistent with the real-time device tag in the historical abnormal data of all historical monitoring time points are extracted to obtain the historical device abnormality data of each real-time device to be detected in the real-time device to be detected set.

6. The power grid monitoring system based on portable infrared detection equipment according to claim 5, characterized in that: The analysis module includes the following units: Historical imaging image vector unit: performs feature extraction on the historical detection device thermal imaging image of each historical temperature sub-report in the historical device report data of each real-time device to be detected, and determines the historical imaging image vector of each historical temperature sub-report; Historical anomaly vector unit: performs feature extraction on each historical anomaly region of each historical anomaly sub-data in the historical anomaly data of each real-time device to be detected, and determines the historical anomaly vector of each historical anomaly region of each historical anomaly sub-data in the historical anomaly data of each real-time device to be detected based on the anomaly type and anomaly level of each historical anomaly region; A historical sub-thermal imaging unit maps and intercepts the historical detection device thermal imaging image in the historical temperature sub-report corresponding to the historical device report data of each real-time device to be detected based on each historical abnormal area of ​​each historical abnormal sub-data in the historical device abnormal data of each real-time device to be detected, and determines a historical sub-thermal imaging image for each historical abnormal area of ​​each historical abnormal sub-data in the historical device abnormal data of each real-time device to be detected; A historical sub-imaging image vector unit is configured to extract features from a historical sub-thermal imaging image of each historical abnormal region of each historical abnormal sub-data in the historical device abnormal data of each real-time device to be detected, and determine a historical sub-imaging image vector of a historical sub-thermal imaging image of each historical abnormal region of each historical abnormal sub-data in the historical device abnormal data of each real-time device to be detected; Equipment abnormality type data unit: Based on the abnormality types of all historical abnormal areas of all historical abnormal sub-data in the historical equipment abnormality data of each real-time equipment to be detected, the historical abnormality vectors and historical sub-imaging map vectors of each historical abnormal area of ​​each historical abnormal sub-data in the historical equipment abnormality data of each real-time equipment to be detected are extracted to determine the equipment abnormality type data of each abnormal type of each real-time equipment to be detected. The equipment abnormality type data includes the historical abnormality vectors and historical sub-imaging map vectors of multiple historical abnormal areas of multiple historical abnormal sub-data.

7. The power grid monitoring system based on portable infrared detection equipment according to claim 6, characterized in that: The analysis module includes the following units: Correlation covariance matrix unit: calculates the correlation covariance matrix of each abnormal type of each real-time device to be detected based on the device abnormality type data of each abnormal type of each real-time device to be detected and the historical imaging map vectors of all historical temperature sub-reports in the historical device report data of each real-time device to be detected; A time-region feature vector unit is configured to calculate a first weight and a region feature vector for each historical abnormality sub-data of each abnormality type of each real-time device to be detected based on the historical device abnormality data of each real-time device to be detected, the device abnormality type data of each abnormality type of each real-time device to be detected, and the historical imaging image vectors of all historical temperature sub-reports in the historical device report data of each real-time device to be detected, and calculate a time-region feature vector for each abnormality type of each real-time device to be detected; Projection matrix unit: performs singular value decomposition on the associated covariance matrix of each abnormality type of each real-time device to be detected, extracts the first specified number of principal component directions of the left singular vector matrix after the singular value decomposition, and determines the projection matrix of each abnormality type of each real-time device to be detected; First eigenvector unit: performs matrix vectorization on the projection matrix of each abnormality type of each real-time device to be detected, and determines the first eigenvector of each abnormality type of each real-time device to be detected; Device type feature vector unit: determines a device type feature vector for each abnormal type of each real-time device to be detected based on the time region feature vector of each abnormal type of each real-time device to be detected and the first feature vector.

8. The power grid monitoring system based on portable infrared detection equipment according to claim 1, characterized in that: Building modules include the following units: Equipment anomaly diagnosis model unit: uses the historical equipment report data of each real-time device to be detected as the input of the equipment anomaly diagnosis model of each real-time device to be detected, and uses the historical equipment anomaly data of each real-time device to be detected as the output of the equipment anomaly diagnosis model of each real-time device to be detected; Construction unit: Based on the device type feature vectors of all abnormal types of each real-time device to be detected, a device abnormality diagnosis model for each real-time device to be detected is constructed.

9. The power grid monitoring system based on portable infrared detection equipment according to claim 3, characterized in that: Diagnostic module, including the following units: The real-time anomaly sub-data includes the real-time device tag to be detected, the current monitoring time point, multiple real-time anomaly areas, the first anomaly type of each real-time anomaly area, the size of the real-time anomaly area, and the real-time anomaly level; A second abnormality type unit is configured to select, from the real-time sub-imaging image vector of each real-time abnormal area in the real-time abnormal sub-data of each real-time device to be detected and the sub-imaging image vector consistency values ​​of the device abnormality type data of each abnormal type of each real-time device to be detected, the abnormality type corresponding to the largest sub-imaging image vector consistency value as the second abnormality type of each real-time abnormal area in the real-time abnormal sub-data of each real-time device to be detected; Prediction accuracy value unit: compares the first abnormality type and the second abnormality type of each real-time abnormal area in the real-time abnormality sub-data of each real-time device to be detected, and calculates the prediction accuracy value of each real-time device to be detected based on the comparison results of all real-time abnormal areas in the real-time abnormality sub-data of each real-time device to be detected; Optimization unit: optimizes the device anomaly diagnosis model of each real-time device to be detected whose prediction accuracy is lower than a preset threshold based on the real-time anomaly sub-data; Alarm unit: determines real-time abnormal data based on the real-time abnormal sub-data of all real-time devices to be detected whose prediction accuracy values ​​are higher than a preset threshold, and issues a real-time alarm based on the real-time abnormal sub-data of all real-time devices to be detected in the real-time abnormal data.

10. A power grid monitoring method based on a portable infrared detection device, characterized in that: The steps include: Step 1: Using a portable infrared detection device, real-time temperature data of a set of devices to be detected in the power grid is collected in real time, and a real-time temperature report is generated; Step 2: Obtain historical temperature reports and historical abnormal data of the power grid, and determine historical device report data and historical device abnormal data of each real-time device to be detected in the set of real-time devices to be detected; Step 3: Based on the historical device report data and historical device abnormality data of each real-time device to be detected, determine the device abnormality type data of each abnormality type of each real-time device to be detected, and determine the device type feature vector of each abnormality type of each real-time device to be detected; Step 4: Build an anomaly diagnosis model for each real-time device to be detected based on the historical device report data, historical device anomaly data, and device type feature vectors of all anomaly types of each real-time device to be detected; Step 5: Based on the real-time temperature report and the abnormal diagnosis model of all real-time devices to be detected, the predicted accuracy value of each real-time device to be detected is calculated and the real-time abnormal data is determined. A real-time alarm is issued based on the real-time abnormal data to achieve real-time monitoring of the power grid; Among them, based on the real-time temperature report and the abnormal diagnosis model of all real-time devices to be detected, the predicted accuracy value of each real-time device to be detected is calculated, including: Real-time abnormality sub-data unit: inputs the real-time temperature sub-report of each real-time device to be detected in the real-time temperature report into the device abnormality diagnosis model of the corresponding real-time device to be detected, and determines the real-time abnormality sub-data of each real-time device to be detected based on the output result of the abnormality diagnosis model; A real-time sub-thermal imaging unit maps and intercepts the real-time thermal imaging image of each device to be detected in the real-time temperature sub-report of each device to be detected in the real-time temperature report based on each real-time abnormal area in the real-time abnormal sub-data of each device to be detected, and determines a real-time sub-thermal imaging image of each real-time abnormal area in the real-time abnormal sub-data of each device to be detected; A real-time sub-imaging image vector unit is configured to extract features from a real-time sub-thermal imaging image of each real-time abnormal area in the real-time abnormal sub-data of each real-time device to be detected, and determine a real-time sub-imaging image vector of each real-time abnormal area in the real-time abnormal sub-data of each real-time device to be detected; Sub-imaging map vector consistency value unit: Based on all device abnormality type data, real-time abnormality sub-data and real-time abnormality areas in the real-time abnormality sub-data of each real-time device to be detected, calculate the real-time sub-imaging map vector of each real-time abnormality area in the real-time abnormality sub-data of each real-time device to be detected, and the sub-imaging map vector consistency value of the device abnormality type data of each abnormal type of each real-time device to be detected.

Images