Power equipment robot inspection system based on AI vision

Abstract

The invention discloses a power equipment robot inspection system based on AI vision, particularly relates to the technical field of power equipment inspection, and comprises an inspection planning module, an AI vision detection module and a work order matching module. According to the invention, the routing inspection priority indexes of the transformer are calculated, the routing inspection priority indexes integrate the historical fault data of the transformer and the transformer information, and the routing inspection route is constructed after the routing inspection priority indexes are ranked from large to small, so that high-risk equipment is ensured to be subjected to routing inspection preferentially; the problem that in the prior art, in the aspect of routing inspection route planning, general routes are mostly adopted for routing inspection, historical fault data and real-time states of equipment cannot be fully combined, and routing inspection pertinence and efficiency are not high is solved.

Description

Technical Field

[0001] This invention relates to the field of power equipment inspection technology, and more specifically, to a power equipment robot inspection system based on AI vision. Background Technology

[0002] With the continuous development of the power system, the number and complexity of power equipment are increasing, and the requirements for power equipment inspection are also getting higher and higher. Traditional power equipment inspection mainly relies on manual inspection, which has problems such as low efficiency, high labor intensity and poor safety. Especially in some high-risk or harsh environments, the difficulty and risk of manual inspection are even greater.

[0003] However, existing AI vision-based power equipment robot inspection systems still have the following shortcomings in practical applications: In terms of inspection route planning, most inspections are carried out using general routes, which cannot fully combine historical fault data and real-time status of equipment, resulting in low targeting and efficiency of inspections. Furthermore, in the fault identification stage, the correlation analysis between temperature anomalies and appearance defects in different areas of the equipment is not in-depth enough, making it difficult to accurately pinpoint the cause of the fault.

[0004] To address this, a robot inspection system for power equipment based on AI vision has been developed. Summary of the Invention

[0005] To overcome the aforementioned deficiencies of the prior art, embodiments of the present invention provide a robot inspection system for power equipment based on AI vision.

[0006] To achieve the above objectives, the present invention provides the following technical solution: A robot inspection system for power equipment based on AI vision includes the following modules: Inspection planning module: Extracts historical fault data and transformer information of each group of transformers to be inspected in the current substation area within the set time zone before the current time point, and performs a comprehensive evaluation to construct the transformer inspection route in the current substation area; AI visual inspection module: Controls the inspection robot to inspect each transformer in the area according to the inspection route. During the inspection process, it collects image information and infrared thermal images of each transformer, and after evaluation and processing, identifies the faulty transformer and simultaneously constructs fault number pairs for the faulty transformer. Work order matching module: Parses the fault number pairs of the faulty transformer, matches the fault type of the faulty transformer, and fills them into the pre-built maintenance work order template to generate the pending work order and the priority work order of the faulty transformer.

[0007] Specifically, the historical fault data includes the number of faults, fault type, and fault occurrence time; the transformer information includes the service life and the node it is currently in.

[0008] Specifically, the step of extracting historical fault data and transformer information for each group of transformers to be inspected within the current substation area within a set time zone prior to the current time point, and performing a comprehensive evaluation, specifically involves: For each group of transformers, the number of faults and fault types are extracted, and the fault types are classified and statistically analyzed. The count of each fault type in each group of transformers within the set time zone is calculated and recorded as the fault type count. Different impact weights are set for different fault types. The number of faults of each group within a set time zone is multiplied by the impact weight of the corresponding fault type, and then the sum is obtained to get the secondary fault value of the transformer within the set time zone. Each secondary fault of the transformer is matched with the corresponding fault occurrence time. The fault occurrence time closest to the current time point is extracted, and the time difference between the current time point and the fault occurrence time is calculated to obtain the time fault value of the transformer. The service life of the transformer is extracted from the transformer information as the old fault value of the transformer. Identify the nodes where each group of transformers is located within the substation area, and pre-determine a set of importance scores corresponding to different nodes; based on the nodes where the transformers are located within the substation area, convert them into corresponding importance scores, which are then used as the barrier values ​​of the transformers. The secondary barrier value, time barrier value, old barrier value, and barrier shadow value of each transformer group at the current time point are respectively marked as follows: The weighting coefficients corresponding to the preset secondary obstacle value, time obstacle value, old obstacle value, and obstacle shadow value are respectively denoted as: Using formulas Calculate the inspection priority index G for each group of transformers at the current time point.

[0009] Specifically, the construction of the transformer inspection route for the current substation area is as follows: Based on the inspection priority index G, the transformer groups are sorted from largest to smallest, and the sorting results are used as the inspection order for each group of transformers. The location of each group of transformers is extracted, and an inspection route is constructed according to the inspection order.

[0010] Specifically, the process for identifying the faulty transformer is as follows: Identify the pre-divided oil conservator area, winding area, and heat dissipation area in the transformer; extract the thermal image information of each divided area of ​​the transformer from the acquired image information, identify the temperature data of each distribution point in each divided area, and then calculate the average value to obtain the average temperature value corresponding to each divided area; preset the temperature threshold corresponding to the average temperature value of each divided area. The average temperature of each zone is compared with the temperature threshold. If the average temperature of a certain zone is higher than the temperature threshold, it is marked as a faulty transformer, and the first group number of that zone is used... The symbol represents the number of each region, where c represents the number of each region, and F = 1, 2 or 3, representing the oil conservator region, winding region and heat dissipation region respectively.

[0011] Specifically, the fault number pair for constructing the faulty transformer is as follows: If the area above the temperature threshold is the oil conservator area or the winding area, then the shell damage features in the oil conservator area and the bushing damage features in the winding area are identified from the image information. The number of pixels of the shell damage features in the oil conservator area and the number of pixels of the bushing damage features in the winding area are counted respectively. Based on the resolution of the image, they are converted into actual areas to obtain the shell damage area of ​​the oil conservator area and the bushing damage area of ​​the winding area. If the area above the temperature threshold is a heat dissipation area, then the dust coverage features on the surface of the heat sink within the heat dissipation area are identified from the image information; the number of pixels with dust coverage features within the heat dissipation area is counted, and converted into actual area based on the image resolution to obtain the dust coverage area of ​​the heat dissipation area. After determining the division area based on the number 'c', the calculated shell damage area, sleeve damage area, and dust coverage area are compared with the corresponding preset shell damage threshold area and sleeve damage threshold area. If the shell damage area, sleeve damage area, or dust coverage area corresponding to the division area that exceeds the temperature threshold is higher than the preset shell damage threshold area, sleeve damage threshold area, or dust coverage threshold area, then the subgroup number of this division area is used... In other words, the subgroup number of the partitioned area is used to indicate the subgroup number. This indicates that the fault number pairs for the faulty transformer are constructed, specifically represented as ( , )or( , ).

[0012] Specifically, the process of generating a work order for a faulty transformer is as follows: Parse the fault number pairs of the faulty transformer. If the result of the fault number pair shows ( , ), then extract The corresponding division areas determine the specific fault types, which include casing abnormalities, bushing abnormalities, and heat sink abnormalities, and correspond to the oil conservator area, winding area, and heat dissipation area, respectively. After determining the fault type, the location of the transformer, the fault type, and the fault number are filled into a pre-built maintenance work order template to generate a pending work order for the faulty transformer and stored in the robot's database for maintenance personnel to process.

[0013] Specifically, the process for generating a fault transformer optimization work order is as follows: If the fault number corresponds to the result ( , Then, the average temperature and temperature threshold corresponding to the faulty transformer are extracted, and the ratio is calculated with the average temperature as the numerator and the temperature threshold as the denominator to obtain the excess ratio of the area corresponding to the faulty transformer. Extract the intervals of each group of excess ratios corresponding to the divided area of ​​the faulty transformer from the pre-built fault database; each interval of excess ratios corresponds to a fault prediction cause; match the excess ratios of the divided area corresponding to the faulty transformer with the intervals of each group of excess ratios to determine the fault prediction cause of the faulty transformer; after determining the fault prediction cause, fill the transformer's location, fault prediction cause, and fault number into the pre-built maintenance work order template to generate the optimal processing work order for the faulty transformer.

[0014] The technical effects and advantages of this invention are as follows: (1) By calculating the inspection priority index of the transformer, and the inspection priority index integrates the historical fault data and transformer information of the transformer, the inspection route is constructed after sorting the inspection priority index from large to small, so as to ensure that high-risk equipment is inspected first. This solves the problem that the existing technology mostly uses general routes for inspection in terms of inspection route planning, which cannot fully combine the historical fault data and real-time status of the equipment, resulting in low targeting and efficiency of inspection. (2) By collecting image information and infrared thermal images, combined with sub-region division, including oil conservator area, winding area and heat dissipation area, we can analyze appearance defects and temperature anomalies, determine the fault number of the transformer, improve the handling effect of subsequent maintenance personnel, and solve the problem that the existing technology does not have a deep enough correlation analysis between temperature anomalies and appearance defects in different areas of the equipment in the fault identification stage, and it is difficult to accurately locate the cause of the fault. (3) By generating work orders to be processed and work orders to be optimized based on the fault number, the former marks the fault type, and the latter matches the fault prediction cause by the excess ratio and optimizes the allocation in combination with the maintenance resource library, providing data support for subsequent processing and improving processing efficiency. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of a robot inspection system for power equipment based on AI vision, according to the present invention. Detailed Implementation

[0016] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0017] Example like Figure 1 As shown, a robot inspection system module for power equipment based on AI vision includes an inspection planning module, an AI vision detection module, and a work order matching module. The inspection planning module is used to preset the inspection interval for a substation area; for example, inspecting the substation once a week. Upon reaching the set inspection interval, the robot is activated, identifying the number and location of transformers to be inspected within the current substation area. Each group of transformers to be inspected within the current substation area is numbered, denoted by j (j=1,2,...,k), where k is the total number of transformers to be inspected. Historical fault data and transformer information for each group of transformers to be inspected within the set time zone prior to the current time point are extracted and comprehensively evaluated to construct the transformer inspection route for the current substation area. Historical fault data includes the number of faults, fault type, and fault occurrence time; transformer information includes service life and location. Specifically: For each group of transformers, the number of faults and fault types are extracted, and the faults are classified and statistically analyzed according to the fault type. The count of each fault type in each group of transformers within the set time zone is calculated and recorded as the fault type count. Fault types include, but are not limited to, winding short circuit, oil tanker leakage fault, and casing corrosion. Set the influence weights corresponding to different fault types, multiply the number of faults of each group of transformers in the set time zone by the influence weight of the corresponding fault type, and then sum them to obtain the fault value of the transformer in the set time zone. Using the equipment number as an index, all fault records of the same transformer are aggregated together. The aggregated records are then categorized and statistically analyzed according to fault type, calculating the frequency of each fault type. Taking a certain group of transformers as an example, statistics show that: winding overheating faults occurred twice, and oil tank leakage faults occurred once. The severity and impact range of different fault types vary significantly. For example, "winding short circuit" is a high-risk fault, which may lead to equipment scrapping or even grid collapse, while "minor casing corrosion" is a low-risk defect. By assigning "impact weights" to fault types, equipment requiring urgent inspection can be identified. Match each fault of the transformer with the corresponding fault occurrence time, extract the fault occurrence time closest to the current time point, and calculate the time difference between the current time point and the fault occurrence time to obtain the fault value of the transformer. The shorter the time fault value, the closer the time of the transformer's most recent fault is to the current time, and the higher the degree of potential fault. Extract the service life of the transformer from the transformer information and use it as the transformer's aging value; Service life is directly related to the aging of equipment insulation and the wear of components. For example, transformers that are more than 20 years old have a high risk of insulation oil deterioration, so they should be inspected with a higher priority. For newly commissioned equipment, the inspection priority can be appropriately reduced to achieve differentiated inspection. Identify the nodes where each group of transformers is located within the substation area, and pre-set a set of importance scores for different nodes; the importance score range is initially set to 1-5, with higher node importance corresponding to higher importance scores; based on the nodes where the transformers are located within the substation area, convert them into corresponding importance scores, which are then used as the transformer's barrier value. Transformers located at critical nodes in the power grid, such as the main transformers of hub substations, undertake the core function of regional power collection and distribution. Once a fault occurs, it may lead to a large-scale power outage, affecting the normal operation of many users and important facilities (such as hospitals, transportation hubs, data centers, etc.). Transformers located at the end or secondary nodes have a relatively smaller impact range when they fail. Therefore, the more important the node, the more serious the consequences of a transformer failure, and the higher the priority of inspection should be. The secondary barrier value, time barrier value, old barrier value, and barrier shadow value of each transformer group at the current time point are respectively marked as follows: The weighting coefficients corresponding to the preset secondary obstacle value, time obstacle value, old obstacle value, and obstacle shadow value are respectively denoted as: Using formulas Calculate the inspection priority index G for each group of transformers at the current time point; By integrating secondary fault values ​​(number of faults × type weight), time fault values ​​(time difference of the most recent fault), old fault values ​​(service years), and fault shadow values ​​(node ​​importance), the historical fault severity, recent fault risks, equipment aging degree, and power grid impact range of transformers are transformed into quantifiable values, avoiding the one-sidedness of evaluation by a single indicator. For example, although a transformer has few faults, its secondary fault value may be significantly higher than that of equipment with frequent low-risk faults due to the high impact weight of high-risk fault types such as "winding short circuit". Therefore, it will be prioritized for inspection. According to the size of the inspection priority index G, the transformers in each group are sorted from largest to smallest. The sorting result is used as the inspection order of each group of transformers. The location of each group of transformers is extracted and the inspection route is constructed according to the inspection order. Based on the inspection priority index G, high-risk transformers are prioritized for inspection. For example, by arranging transformer locations from largest to smallest G value, a straight or circular route of "high priority → medium priority → low priority" can be planned to reduce unnecessary movement distance. If there are interstitial transformers in the route (such as the path between two transformers containing a third group of transformers), they are inserted for inspection through dynamic scheduling to avoid repeatedly passing through the same area and further reduce the inspection time. The AI ​​vision detection module is used to determine the inspection route and control the inspection robot to inspect each transformer in the area according to the inspection route. During the inspection process, according to the preset collection position of each transformer, after reaching the collection position, the module uses the onboard high-definition visible light camera and infrared thermal imager to collect image information and infrared thermal images of each transformer, and after evaluation and processing, identifies the faulty transformer and simultaneously constructs the fault number pair of the faulty transformer. Data collection process: The robot moves to a preset position around a transformer and takes multi-angle pictures of various parts of the transformer, including the oil tank, bushings, oil conservator, radiator, tap changer, etc., using a high-definition visible light camera. At the same time, the camera is equipped with image stabilization and autofocus to ensure clear images with a resolution of 4K. After completing image acquisition, the robot switches to an infrared thermal imager to acquire infrared images of the transformer. The infrared thermal imager can detect the temperature distribution on the surface of the transformer. The robot takes pictures of the entire transformer and key parts (such as windings, joints, and tap changer contacts) at a distance of 3-5 meters from the transformer, according to the preset scanning path, to obtain infrared thermal images. By default, image information and infrared thermal images are preprocessed, including but not limited to: Image denoising: Gaussian filtering and median filtering are used to remove noise from visible light images, and non-uniformity correction (NUC) is used to eliminate fixed pattern noise in infrared images.

[0018] Brightness / contrast normalization: Adjusts the illumination uniformity of visible light images to avoid missed defects caused by shadows; performs temperature grayscale mapping on infrared images (e.g., cool tones represent low temperatures and warm tones represent high temperatures) to enhance the visibility of areas with abnormal temperatures.

[0019] Image registration: Visible light and infrared images are aligned by feature point matching (such as SIFT, ORB) to ensure that pixels at the same physical location correspond one-to-one, laying the foundation for subsequent fusion analysis; Specifically: Identify the pre-divided oil conservator area, winding area, and heat dissipation area in the transformer; extract the thermal image information of each divided area of ​​the transformer from the acquired image information, identify the temperature data of each distribution point in each divided area, and then calculate the average value to obtain the average temperature value corresponding to each divided area; preset the temperature threshold corresponding to the average temperature value of each divided area. The temperature threshold corresponding to the average temperature of each divided region varies according to the temperature change of the transformer's environment. The temperature range corresponding to each group of ambient temperature can be set, and the temperature threshold of each group of temperature range can be set to correspond to the temperature of each divided region of the transformer. Thus, the temperature threshold of each divided region can be determined during the analysis process. The average temperature of each zone is compared with the temperature threshold. If the average temperature of a certain zone is higher than the temperature threshold, it is marked as a faulty transformer, and the first group number of that zone is used... The symbol represents the number of each region; where c represents the sequence number of each region, and F=1,2 or3, representing the oil conservator region, winding region and heat dissipation region respectively. If the area above the temperature threshold is classified as either the oil conservator area or the winding area, then the shell damage features in the oil conservator area and the bushing damage features in the winding area are identified from the image information. The damage detection model can be trained using YOLOv8 or Faster R-CNN, and the damage location, including cracks and holes, can be directly output from the input RGB image. The number of pixels of shell damage features in the oil conservator area and the number of pixels of bushing damage features in the winding area are counted respectively, and converted into actual areas based on the image resolution to obtain the shell damage area in the oil conservator area and the bushing damage area in the winding area. If the area above the temperature threshold is a heat dissipation area, then the dust coverage features on the surface of the heat sink within the heat dissipation area are identified from the image information; the dust coverage features are determined by the image grayscale value; the number of pixels with dust coverage features within the heat dissipation area is counted, and converted into actual area based on the image resolution to obtain the dust coverage area of ​​the heat dissipation area. After determining the division area based on the number 'c', the calculated shell damage area, sleeve damage area, and dust coverage area are compared with the corresponding preset shell damage threshold area and sleeve damage threshold area. If the shell damage area, sleeve damage area, or dust coverage area corresponding to the division area that exceeds the temperature threshold is higher than the preset shell damage threshold area, sleeve damage threshold area, or dust coverage threshold area, then the subgroup number of this division area is used... In other words, the subgroup number of the partitioned area is used to indicate the subgroup number. This indicates that the fault number pairs for the faulty transformer are constructed, specifically represented as ( , )or( , ); Additional explanation, This indicates the secondary group number, used to remind subsequent maintenance personnel that the transformer's zone temperature is abnormal, but it is not caused by damage or excessive dust, and the specific cause of the temperature abnormality is uncertain. The work order matching module is used to parse the fault number pairs of the faulty transformer, match the fault type of the faulty transformer, and fill them into the pre-built maintenance work order template to generate the pending work order and the priority work order of the faulty transformer. Specifically: Parse the fault number pairs of the faulty transformer. If the result of the fault number pair shows ( , ), then extract The corresponding division areas determine the specific fault types, which include casing abnormalities, bushing abnormalities, and heat sink abnormalities, and correspond to the oil conservator area, winding area, and heat dissipation area, respectively. Based on different regions, the system automatically calls up the corresponding maintenance work order template, which contains fixed fields such as work order number, transformer location name, and fault number; the integrated information is automatically filled into the work order template. After determining the fault type, the location of the transformer, the fault type, and the fault number are filled into the pre-built maintenance work order template to generate a pending work order for the faulty transformer and stored in the robot's database for maintenance personnel to process. If the fault number corresponds to the result ( , Then, the average temperature and temperature threshold corresponding to the faulty transformer are extracted, and the ratio is calculated with the average temperature as the numerator and the temperature threshold as the denominator to obtain the excess ratio of the area corresponding to the faulty transformer. Extract the intervals of each group of excess ratios in the divided regions corresponding to the faulty transformer from the pre-built fault database; each interval of the excess ratio corresponds to a fault prediction cause; match the excess ratios of the divided regions corresponding to the faulty transformer with the intervals of each group of excess ratios to determine the fault prediction cause of the faulty transformer. Additional explanation: When the fault number triggers the display, the corresponding area (oil conservator / winding / heat dissipation) of the faulty transformer is calculated. For example, if the average temperature of a certain winding area is 85℃ and the threshold is 70℃, then the overload ratio = 85 / 70≈1.21. The overload ratio intervals corresponding to each group of winding areas and the corresponding estimated fault causes are extracted from the fault database. [1.0,1.2)={Slight overload of winding}, [1.2,1.5)={Poor bushing contact}, [1.5,+∞)={Short circuit of winding}. After determining the predicted cause of the fault, the location of the transformer, the predicted cause of the fault, and the fault number are filled into a pre-built maintenance work order template to generate a best-case work order for the faulty transformer; at the same time, the maintenance resource library is searched, and the best-case work order is sent to the optimization index. The highest-ranking maintenance personnel; Maintenance resource database construction: Establish a maintenance resource database containing various information about maintenance personnel, including skills and qualifications, work experience, and current work status (available, busy, task progress). The system retrieves the work status of each maintenance worker in the maintenance resource database at the current time, including whether they are working, on vacation, or idle. After filtering the idle maintenance workers, a circle is drawn with the location of the faulty transformer as the center and a set radius. Each idle maintenance worker within the circle is marked. The historical maintenance count of each idle maintenance worker is then obtained and compared with a set reference value. Idle maintenance workers whose counts exceed the reference value are selected and marked as standby personnel. The distance between the standby personnel and the location of the faulty transformer is obtained and recorded as v1; the historical number of maintenance operations performed by the standby personnel is recorded as v2; at the same time, the time consumed by the standby personnel for each maintenance operation is extracted, and the average time consumed by each operation is taken as v3. According to the formula Calculate the optimal selection index for standby personnel. ;in These are the weighting coefficients for distance, number of historical maintenance operations, and average time consumption, respectively. The system combines information from the maintenance resource database for matching. In the matching process, in addition to considering distance factors, it also comprehensively evaluates factors such as the work efficiency of maintenance personnel, sorts the matching results, recommends the best personnel, and then updates the work status of maintenance personnel in real time to avoid duplicate assignment. The above formulas are all dimensionless calculations. Dimensionless calculations can be performed using various methods such as standardization, which will not be elaborated here. The formulas are derived from software simulations based on a large amount of collected data, and the preset parameters in the formulas can be set by those skilled in the art according to the actual situation.

[0020] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, ATA hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. The semiconductor medium can be a solid-state ATA hard disk.

[0021] It should be understood that in the various embodiments of this application, the order of the above-mentioned processes does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0022] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0023] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0024] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0025] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0026] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable ATA hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0027] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A robot inspection system for power equipment based on AI vision, characterized in that, Includes the following modules: Inspection planning module: Extracts historical fault data and transformer information of each group of transformers to be inspected in the current substation area within the set time zone before the current time point, and performs a comprehensive evaluation to construct the transformer inspection route in the current substation area; AI visual inspection module: Controls the inspection robot to inspect each transformer in the area according to the inspection route. During the inspection process, it collects image information and infrared thermal images of each transformer, and after evaluation and processing, identifies the faulty transformer and simultaneously constructs fault number pairs for the faulty transformer. Work order matching module: Parses the fault number pairs of the faulty transformer, matches the fault type of the faulty transformer, and fills them into the pre-built maintenance work order template to generate the pending work order and the priority work order of the faulty transformer.

2. The AI ​​vision-based robot inspection system for power equipment according to claim 1, characterized in that, The historical fault data includes the number of faults, fault types, and fault occurrence times; transformer information includes service life and current node.

3. The AI ​​vision-based robot inspection system for power equipment according to claim 2, characterized in that, The process involves extracting historical fault data and transformer information for each group of transformers requiring inspection within the current substation area, prior to the current time point, and performing a comprehensive evaluation. Specifically: For each group of transformers, the number of faults and fault types are extracted, and the fault types are classified and statistically analyzed. The count of each fault type in each group of transformers within the set time zone is calculated and recorded as the fault type count. Set the influence weights corresponding to different fault types, multiply the number of each fault type of the transformer in the set time zone by the influence weight of the corresponding fault type, and then sum them to obtain the secondary fault value of the transformer in the set time zone; match each secondary fault of the transformer with the corresponding fault occurrence time, extract the fault occurrence time closest to the current time point, and calculate the time difference between the current time point to obtain the time fault value of the transformer. Extract the service life of the transformer from the transformer information and use it as the transformer's aging value; Identify the nodes of each group of transformers within the substation area, and pre-set a set of importance scores for each node; Based on the node where the transformer is located within the substation area, the corresponding importance score is converted and used as the barrier value of the transformer. The secondary barrier value, time barrier value, old barrier value, and barrier shadow value of each transformer group at the current time point are respectively marked as follows: ; The weighting coefficients corresponding to the preset secondary obstacle value, time obstacle value, old obstacle value, and obstacle shadow value are respectively denoted as: Using formulas Calculate the inspection priority index G for each group of transformers at the current time point.

4. The AI ​​vision-based robot inspection system for power equipment according to claim 3, characterized in that, The specific steps for constructing the transformer inspection route in the current substation area are as follows: Based on the inspection priority index G, the transformer groups are sorted from largest to smallest, and the sorting results are used as the inspection order for each group of transformers. The location of each group of transformers is extracted, and an inspection route is constructed according to the inspection order.

5. The AI ​​vision-based robot inspection system for power equipment according to claim 4, characterized in that, The specific process for identifying faulty transformers is as follows: Identify the pre-divided oil conservator area, winding area, and heat dissipation area in the transformer; extract the thermal image information of each divided area of ​​the transformer from the acquired image information, identify the temperature data of each distribution point in each divided area, and then calculate the average value to obtain the average temperature value corresponding to each divided area; preset the temperature threshold corresponding to the average temperature value of each divided area. The average temperature of each zone is compared with the temperature threshold. If the average temperature of a certain zone is higher than the temperature threshold, it is marked as a faulty transformer, and the first group number of that zone is used... The symbol represents the number of each region, where c represents the number of each region, and F = 1, 2 or 3, representing the oil conservator region, winding region and heat dissipation region respectively.

6. The AI ​​vision-based robot inspection system for power equipment according to claim 5, characterized in that, The fault number pairs for constructing the faulty transformer are specifically as follows: If the area above the temperature threshold is the oil conservator area or the winding area, then the shell damage features in the oil conservator area and the bushing damage features in the winding area are identified from the image information. The number of pixels of the shell damage features in the oil conservator area and the number of pixels of the bushing damage features in the winding area are counted respectively. Based on the resolution of the image, they are converted into actual areas to obtain the shell damage area of ​​the oil conservator area and the bushing damage area of ​​the winding area. If the area defined as being above the temperature threshold is a heat dissipation area, then the dust coverage features on the surface of the heat sink within the heat dissipation area are identified from the image information. The number of pixels with dust coverage features within the heat dissipation area is counted, and then converted into actual area based on the image resolution to obtain the dust coverage area of ​​the heat dissipation area. After determining the division area based on the number 'c', the calculated shell damage area, sleeve damage area, and dust coverage area are compared with the corresponding preset shell damage threshold area and sleeve damage threshold area. If the shell damage area, sleeve damage area, or dust coverage area corresponding to the division area that exceeds the temperature threshold is higher than the preset shell damage threshold area, sleeve damage threshold area, or dust coverage threshold area, then the subgroup number of this division area is used... In other words, the subgroup number of the partitioned area is used to indicate the subgroup number. This indicates that the fault number pairs for the faulty transformer are constructed, specifically represented as ( , )or( , ).

7. The AI ​​vision-based robot inspection system for power equipment according to claim 6, characterized in that, The pending work order for generating the faulty transformer: Parse the fault number pairs of the faulty transformer. If the result of the fault number pair shows ( , ), then extract The corresponding division areas determine the specific fault types, which include casing abnormalities, bushing abnormalities, and heat sink abnormalities, and correspond to the oil conservator area, winding area, and heat dissipation area, respectively. After determining the fault type, the location of the transformer, the fault type, and the fault number are filled into a pre-built maintenance work order template to generate a pending work order for the faulty transformer and stored in the robot's database for maintenance personnel to process.

8. The AI ​​vision-based robot inspection system for power equipment according to claim 7, characterized in that, The optimized processing work order for the generated faulty transformer: If the fault number corresponds to the result ( , Then, the average temperature and temperature threshold corresponding to the faulty transformer are extracted, and the ratio is calculated with the average temperature as the numerator and the temperature threshold as the denominator to obtain the excess ratio of the area corresponding to the faulty transformer. Extract the intervals of each group of excess ratios corresponding to the divided regions of the faulty transformer from the pre-built fault database; each interval of the excess ratio corresponds to a fault prediction cause. The excess ratio of the corresponding region of the faulty transformer is matched with the interval of each group of excess ratios to determine the estimated cause of the faulty transformer. After determining the estimated cause of the fault, the location of the transformer, the estimated cause of the fault, and the fault number are filled into the pre-built maintenance work order template to generate the optimal handling work order for the faulty transformer.

Images