Equipment fault detection system and method for multi-agent collaboration in integrated energy system

Abstract

The present invention discloses a multi-agent collaborative equipment fault detection system and method for an integrated energy system, the system comprising: a sub-device, a parent device and a user background; the sub-device is used for equipment inspection, obtains equipment fault information, and divides the fault area where the faulty equipment is located, and obtains the spatial coordinates of the sub-device; the parent device receives the spatial coordinates, and performs a three-dimensional scan of the fault area in conjunction with the sub-device based on the spatial coordinates, and transmits the scan results to the user background; the user background generates an abnormal diagnosis report based on the equipment fault information and the scan results. The lightweight equipment of the present invention improves the comprehensiveness of inspections, enables the system to respond to equipment failures in real time, helps to quickly deal with potential problems, and reduces losses caused by failures; combines a variety of sensors to make fault location more accurate, performs three-dimensional scanning and quantitative analysis, and provides accurate data support for subsequent fault handling.

Description

Technical Field

[0001] The present invention belongs to the technical field of integrated energy system operation and maintenance, and specifically relates to an equipment fault detection system and method for integrated energy system multi-agent collaboration. Background Art

[0002] The operation and maintenance of integrated energy systems is crucial to improving the resilience of power grids and achieving sustainable development. Traditional methods have limitations in terms of inspection efficiency and fault location accuracy. Fixed monitoring or manual inspections are usually used, but this method cannot achieve full coverage in large-scale systems and is prone to missed detections and misjudgments. The present invention uses lightweight equipment to achieve efficient inspections, reduce labor costs, and reduce energy consumption. At the same time, it introduces a multi-agent collaboration and interaction framework to effectively improve the comprehensiveness of equipment fault detection. The present invention provides a comprehensive fault analysis basis for operation and maintenance personnel, and provides a more advanced and reliable solution for the sustainable operation of integrated energy systems. Summary of the invention

[0003] The present invention aims to solve the deficiencies of the prior art and proposes a multi-agent collaborative equipment fault detection system and method for an integrated energy system, thereby achieving efficient, reliable and sustainable operation and maintenance of the integrated energy system through a multi-agent collaborative and interactive mode.

[0004] To achieve the above object, the present invention provides the following solutions:

[0005] The equipment fault detection system of multi-agent collaboration in the integrated energy system includes: sub-equipment, parent equipment and user background;

[0006] The sub-device is used for equipment inspection, obtaining equipment fault information, segmenting the fault area where the faulty equipment is located, and obtaining the spatial coordinates of the sub-device;

[0007] The parent device receives the spatial coordinates, performs a three-dimensional scan of the fault area in conjunction with the child device based on the spatial coordinates, and transmits the scan result to the user background;

[0008] The user backend generates an abnormality diagnosis report based on the device fault information and the scanning result.

[0009] Preferably, the sub-device uses a first drone device, including: a first drone body, an infrared thermal imaging camera, a natural light camera, an acoustic sensor, an electromagnetic field sensor, a first wireless communication device, a first micro-computing unit, a high-precision GPS and an inertial navigation device;

[0010] The first drone body is used to carry the infrared thermal imaging camera, the natural light camera, the acoustic sensor, the electromagnetic field sensor, the first wireless communication device, the first micro-computing unit, the high-precision GPS and the inertial navigation device; the infrared thermal imaging camera is used to detect thermal failures of the equipment;

[0011] The natural light camera is used to detect appearance damage and deformation failure of the equipment;

[0012] The acoustic sensor is used to detect mechanical vibration faults of the equipment;

[0013] The electromagnetic field sensor is used to detect magnetic field faults of the equipment;

[0014] The high-precision GPS and the inertial navigation device are used to obtain the spatial coordinates of the first drone device in real time;

[0015] The first wireless communication device is used to transmit the fault area and the spatial coordinates to the mother device, and transmit the device fault information to the user background;

[0016] The first micro-computing unit divides the integrated energy system space into a number of voxel spaces of preset volumes, and each of the voxel spaces sets the normal working characteristics Si[x, y, z, cx, t, cc, zd, p1, p2, ..., p n ], where S i represents the eigenvector of the i-th voxel, x, y and z represent the spatial coordinates of the drone, cx represents the direction of the drone, t represents the operating temperature of the device, cc represents the magnetic field state of the device when it is working normally, zd represents the vibration state value of the device, p1, p2…p n Image features for proper operation of the device.

[0017] Preferably, the sub-device detects the real-time working feature Y in the current voxel space in real time i [x′, y′, z′, cx′, t′, cc′, zd′, p1′, p2′,...p n ′], and the device works normally with characteristic S i Perform Euclidean distance calculation, and when the calculated value exceeds a preset threshold value α, determine that the device is faulty, and record the faulty area and the device fault information.

[0018] Preferably, the mother device uses a second drone device, including: a second drone body, a laser radar, a second wireless communication device, and a second micro-computing unit;

[0019] The second drone body is used to carry the laser radar, the second wireless communication device and the second micro computing unit;

[0020] The second wireless communication device is used to receive the fault area and the spatial coordinates;

[0021] The second micro-computing unit is used to calculate the optimal route based on the spatial coordinates;

[0022] The laser radar is used to perform a three-dimensional scan on the fault area to obtain a three-dimensional scan image;

[0023] The second micro-computing unit is further used to complete the point cloud of the three-dimensional scanned image using a point cloud game completion method to obtain the scan result;

[0024] The second wireless communication device is further used to transmit the scanning result to the user background.

[0025] The present invention also provides a device fault detection method for multi-agent collaboration in an integrated energy system, the detection method being applied to any of the above-mentioned detection systems, comprising the following steps:

[0026] Inspect the equipment, obtain equipment fault information, segment the fault area where the faulty equipment is located, and obtain the spatial coordinates of the sub-equipment;

[0027] Performing a three-dimensional scan of the fault area based on the spatial coordinates;

[0028] An abnormality diagnosis report is generated based on the device fault information and the scanning result.

[0029] Preferably, the sub-device is used to inspect the equipment to obtain thermal faults, appearance damage and deformation faults, mechanical vibration faults and magnetic field faults of the equipment, and the spatial coordinates of the sub-device are obtained at the same time;

[0030] The sub-device uses the first drone equipment, including: a first drone body, an infrared thermal imaging camera, a natural light camera, an acoustic sensor, an electromagnetic field sensor, a first wireless communication device, a first micro-computing unit, a high-precision GPS and an inertial navigation device.

[0031] Preferably, the method for obtaining the fault area includes:

[0032] The integrated energy system space is divided into a number of voxel spaces of preset volumes, and each of the voxel spaces sets the normal working characteristics S of the equipment. i [x,y,z,cx,t,cc,zd,p1,p2,...,p n ], where S irepresents the eigenvector of the i-th voxel, x, y and z represent the spatial coordinates of the drone, cx represents the direction of the drone, t represents the operating temperature of the device, cc represents the magnetic field state of the device when it is working normally, zd represents the vibration state value of the device, p1, p2…p n Image features for proper operation of the device;

[0033] Real-time detection of the real-time working feature Y in the current voxel space i [x',y',z',cx',t',cc',zd',p1',p2',...,p n '], and the device works properly with characteristic S i Perform Euclidean distance calculation, and when the calculated value exceeds a preset threshold value α, determine that the device is faulty, and record the faulty area and the device fault information.

[0034] Preferably, the mother device is used to perform a three-dimensional scan on the device, an optimal route is calculated based on the spatial coordinates, the fault area is three-dimensionally scanned to obtain a three-dimensional scanned image, and the three-dimensional scanned image is completed with a point cloud game completion method to obtain the scan result;

[0035] The mother device uses a second drone device, including: a second drone body, a laser radar, a second wireless communication device and a second micro computing unit.

[0036] Compared with the prior art, the present invention has the following beneficial effects:

[0037] (1) Efficient inspection and real-time response: Lightweight equipment improves the comprehensiveness of inspections and enables the system to respond to equipment failures in real time, helping to quickly deal with potential problems and reduce losses caused by failures;

[0038] (2) Accurate fault location and quantitative analysis: The multi-agent collaborative and interactive framework, combined with a variety of sensors, makes fault location more accurate, performs three-dimensional scanning and quantitative analysis, and provides precise data support for subsequent fault handling;

[0039] (3) Reduce labor costs and energy consumption: The sub-equipment can conduct autonomous inspection and positioning, reducing the reliance on human resources. At the same time, the lightweight design can reduce energy consumption and improve the sustainability of the system.

[0040] (4) Comprehensive diagnostic reports: The generated diagnostic reports provide operation and maintenance personnel with a comprehensive basis for fault analysis, helping to formulate maintenance measures more accurately. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the technical solution of the present invention, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0042] Figure 1 A schematic diagram of the system structure of an embodiment of the present invention;

[0043] Figure 2 A schematic diagram of a method flow chart of an embodiment of the present invention;

[0044] Figure 3 Schematic diagram of the game completion principle of an embodiment of the present invention. DETAILED DESCRIPTION

[0045] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0046] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0047] The present invention sends information to the parent device after the child device locates the device fault. The parent device is in a dormant state after not receiving the fault information. When receiving the fault information of the child device, it automatically powers on and flies to the fault area to jointly detect and quantify the fault information with the child device. The child device and the parent device are both intelligent entities and can process the target individually or jointly.

[0048] Embodiment 1

[0049] In this embodiment, if Figure 1 As shown, the equipment fault detection system for multi-agent collaboration in the integrated energy system includes: sub-equipment, parent equipment and user background.

[0050] The sub-device is used for equipment inspection, obtaining equipment fault information, segmenting the fault area where the faulty equipment is located, and obtaining the spatial coordinates of the sub-device.

[0051] The sub-device can use a lightweight intelligent drone, including: a lightweight drone body, an infrared thermal imaging camera, a natural light camera, an acoustic sensor, an electromagnetic field sensor, a first wireless communication device, a first micro-computing unit, a high-precision GPS and an inertial navigation device. The first drone body can use a lightweight drone to carry an infrared thermal imaging camera, a natural light camera, an acoustic sensor, an electromagnetic field sensor, a first wireless communication device, a first micro-computing unit, a high-precision GPS and an inertial navigation device; the infrared thermal imaging camera is used to detect thermal failures of the equipment. The inspection of the sub-device is set by the administrator through the flight route, inspection time, and cycle. During non-inspection time, it is dormant and charged in the base. When the inspection time is set, the lightweight drone flies and inspects according to the set route. The infrared thermal imaging camera is used to detect thermal failures of the equipment; the natural light camera is used to detect appearance damage and deformation failures of the equipment; the acoustic sensor is used to detect mechanical vibration failures of the equipment; the electromagnetic field sensor is used to detect magnetic field failures of the equipment; the high-precision GPS and inertial navigation device are used to obtain the spatial coordinates of the drone in real time; the first wireless communication device is used to transmit the fault area and spatial coordinates to the parent device, and transmit the equipment failure information to the user background. The first micro-computing unit divides the integrated energy system space into a number of voxel spaces of preset volumes, and each voxel space sets the normal working characteristics S of the equipment i [x,y,z,cx,t,cc,zd,p1,p2,...,p n ], where S i represents the eigenvector of the i-th voxel, x, y and z represent the spatial coordinates of the drone, cx represents the direction of the drone, t represents the operating temperature of the device, cc represents the magnetic field state of the device when it is working normally, zd represents the vibration state value of the device, p1, p2…p n Image features for normal operation of the device; real-time detection of real-time working features Y in the current voxel space i [x',y',z',cx',t',cc',zd',p1',p2',...,p n '], and the device works normally with characteristic S i Perform Euclidean distance calculation. When the calculated value exceeds the preset threshold α, the device is judged to be faulty and the fault area and device fault information are recorded.

[0052] In this embodiment, the sub-device has autonomous obstacle avoidance and memory functions. For equipment areas where multiple faults occur, the sub-device will autonomously stay for a long time for detection based on the fault history in the voxel. The stay time for faults is T. k =ln(ke)·T0, where k is the number of faults within the voxel, T0 is the dwell time without historical faults, and e is a natural constant.

[0053] The mother device receives the spatial coordinates, performs a three-dimensional scan of the fault area in conjunction with the child device based on the spatial coordinates, and transmits the scan results to the user background. In this embodiment, when the child device detects that the device is working abnormally, it wirelessly transmits the status information to the mother device, and the mother device flies to the target area according to the optimal route. The mother device uses a second drone device, including: a second drone body, a laser radar, a second wireless communication device, and a second micro-computing unit; the second drone body can use a heavyweight drone to carry the laser radar, the second wireless communication device, and the second micro-computing unit; the second wireless communication device is used to receive the fault area and spatial coordinates; the second micro-computing unit is used to calculate the optimal route based on the spatial coordinates; the laser radar is used to perform a three-dimensional scan of the fault area to obtain a three-dimensional scanned image; the second micro-computing unit is also used to use a point cloud game completion method to complete the three-dimensional scanned image to obtain a scan result; the second wireless communication device is also used to transmit the scan result to the user background.

[0054] In this embodiment, a point cloud game completion method is proposed by referring to the Go game mode, including the following steps:

[0055] The scanned point cloud data is transformed into C[w,v,3] according to the spatial state, where w is the width of the point cloud, v is the length of the point cloud, and 3 represents the point cloud coordinates;

[0056] Perform feature alignment on the equipment fault image P collected by the sub-equipment and the point cloud data C;

[0057] Assume that there is no missing in the shape point cloud C, and all the points on it are white pieces. When some of the points are black pieces, since the black pieces are surrounded by white pieces, all the "qi" of the black pieces are occupied and eaten, thus there are gaps;

[0058] The image P is all white chess pieces. Find the corresponding positions of the black chess pieces in the point cloud C. Connect all the "qi" values ​​around the original black chess pieces on P. Use the least squares method to fit the straight line l1. Then substitute the black chess pieces' position values ​​on P into l1 to obtain the x-axis coordinates.

[0059] Fit the "qi" values ​​around the black chess piece's empty spot in point cloud C with a straight line l2, and substitute the x-axis coordinate obtained by P into l2 to obtain the current missing position value.

[0060] The user backend generates an abnormal diagnosis report based on device fault information and scanning results.

[0061] In this embodiment, the abnormal diagnosis report includes the working state value S of the device in the current voxel space when it is working normally. i [x,y,z,cx,t,cc,zd,p1,p2,...,p n ], working state value Y at fault i[x',y',z',cx',t',cc',zd',p1',p2',...,p n '], fault analysis through Euclidean distance calculation, RGB visible light image, and complete 3D status information of faulty equipment. Operation and maintenance personnel analyze and formulate maintenance measures based on equipment fault information and 3D reconstruction status.

[0062] Embodiment 2

[0063] In this embodiment, if Figure 2 As shown, the equipment fault detection method of multi-agent collaboration in the integrated energy system includes the following steps:

[0064] S1. Inspect the equipment to obtain equipment fault information, segment the fault area where the faulty equipment is located, and obtain the spatial coordinates of the sub-equipment.

[0065] The sub-device is used to inspect the equipment, obtain the equipment's thermal faults, appearance damage and deformation faults, mechanical vibration faults and magnetic field faults, and obtain the spatial coordinates of the lightweight intelligent drone; the administrator sets the flight route, detection time and cycle of the lightweight intelligent drone. During non-inspection time, the lightweight intelligent drone is dormant and charged on the base. When the inspection time is set, the lightweight drone flies and inspects according to the set route. The sub-device uses the first drone equipment, including: the first drone body, infrared thermal imaging camera, natural light camera, acoustic sensor, electromagnetic field sensor, first wireless communication device, first micro-computing unit, high-precision GPS and inertial navigation device.

[0066] The method for obtaining the fault area includes: dividing the integrated energy system space into a number of voxel spaces with preset volumes, and setting the normal working characteristics S of the equipment in each voxel space i [x,y,z,cx,t,cc,zd,p1,p2,...,p n ], where S i represents the eigenvector of the i-th voxel, x, y and z represent the spatial coordinates of the drone, cx represents the direction of the drone, t represents the operating temperature of the device, cc represents the magnetic field state of the device when it is working normally, zd represents the vibration state value of the device, p1, p2…p n Image features for normal operation of the device; real-time detection of real-time working features Y in the current voxel space i [x',y',z',cx',t',cc',zd',p1',p2',...,p n '], and the device works normally with characteristic S i Perform Euclidean distance calculation. When the calculated value exceeds the preset threshold α, the device is judged to be faulty and the fault area and device fault information are recorded.

[0067] In this embodiment, the lightweight intelligent drone has autonomous obstacle avoidance and memory functions. For equipment areas where multiple failures occur, the lightweight intelligent drone will autonomously stay for a long time for detection based on the failure history in the voxel. The stay time for failures is T. k =ln(ke)·T0, where k is the number of faults within the voxel, T0 is the dwell time without historical faults, and e is a natural constant.

[0068] S2. Perform a three-dimensional scan of the fault area based on the spatial coordinates.

[0069] The mother device is used to perform a three-dimensional scan of the device. In this embodiment, when the child device detects that the device is working abnormally, it wirelessly transmits the status information to the mother device, calculates the optimal route based on the spatial coordinates, and the mother device flies to the target area along the optimal route and performs a three-dimensional scan of the fault area to obtain a three-dimensional scanned image. The three-dimensional scanned image is completed with point cloud using a point cloud completion method to obtain a scan result. The mother device uses a second drone device, including: a second drone body, a laser radar, a second wireless communication device and a second micro-computing unit.

[0070] In this embodiment, a point cloud game completion method is proposed by referring to the Go game mode, including the following steps:

[0071] The scanned point cloud data is transformed into C[w,v,3] according to the spatial state, where w is the width of the point cloud, v is the length of the point cloud, and 3 represents the point cloud coordinates;

[0072] Perform feature alignment on the equipment fault image P collected by the sub-equipment and the point cloud data C;

[0073] Assume that there is no missing in the shape point cloud C, and all the points on it are white pieces. When some of the points are black pieces, since the black pieces are surrounded by white pieces, all the "qi" of the black pieces are occupied and eaten, thus there are gaps;

[0074] The image P is all white chess pieces. Find the corresponding positions of the black chess pieces in the point cloud C. Connect all the "qi" values ​​around the original black chess pieces on P. Use the least squares method to fit the straight line l1. Then substitute the black chess pieces' position values ​​on P into l1 to obtain the x-axis coordinates.

[0075] Fit the "qi" values ​​around the black chess piece's empty spot in point cloud C with a straight line l2, and substitute the x-axis coordinate obtained by P into l2 to obtain the current missing position value.

[0076] S3. Generate an abnormal diagnosis report based on the equipment fault information and scanning results.

[0077] In this embodiment, the abnormal diagnosis report includes the working state value S of the device in the current voxel space when it is working normally. i[x,y,z,cx,t,cc,zd,p1,p2,...,p n ], working state value Y at fault i [x′, y′, z′, cx′, t′, cc′, zd′, p1′, p2′,...′ p′ n ], fault analysis through Euclidean distance calculation, RGB visible light image, and complete 3D status information of faulty equipment. Operation and maintenance personnel analyze and formulate maintenance measures based on equipment fault information and 3D reconstruction status.

[0078] Embodiment 3

[0079] In this embodiment, taking actual detection as an example:

[0080] The integrated energy system space is divided into voxels, each space voxel is 1m 3 , after sorting, set the normal working value of the equipment in each spatial voxel. For example, the normal working characteristic of the equipment in voxel number 21 is S i [52,18,37,35°,60,5,78,34,36,…,78]. When the sub-device drone flies to the voxel coordinate, the device status in the area is detected, including the thermal state, magnetic field state, mechanical vibration state, and the device appearance damage and deformation fault characteristics, and the state feature Y is obtained. i [51,16,38,36°,90,6,,85,35,38,…,80], calculated by Euclidean distance, is about 31.08. If the fault threshold is 25, it is determined that the device detected by this voxel is faulty. At the same time, according to the element operation weight, it can be preliminarily determined that the problem is high temperature.

[0081] In the above example, the child device sends the status information [52, 18, 37, 35°] to the parent device, which then flies to the coordinate area and performs a 3D scan of the target device. After scanning, the point cloud data is aligned with the image captured by the child device, and the image is used to complete the point cloud across modalities. Figure 3As shown, first, let the point cloud be all white chess pieces; then let the actual position be black chess pieces. At this time, the black chess pieces have no "qi" and are captured, resulting in a vacancy; then find the corresponding position of the vacancy on the image chessboard P, take out the "qi" value of the corresponding position of the vacancy for linear fitting, and set eight qi values, from left to right, from top to bottom, as [52, 54, 62, 87, 85, 90, 88, 92], and the fitting line is approximately: y≈4.2143x+47.2143, and at the same time Substitute the value 80 of the vacant position into the straight line, and get x of approximately 7.78; next, perform linear fitting on the qi of the vacant point C on the point cloud chessboard. Suppose there are eight qi values, from left to right and from top to bottom, [12, 17, 27, 24, 29, 27, 32, 33]. The fitting straight line is approximately: y≈2.3571x+14.25. Substitute the x value calculated at the vacant position corresponding to the image chessboard into the equation, and get the calculated point cloud missing value of 32.58.

[0082] In the above example, the child and parent multi-agent devices will complete the equipment 3D reconstruction point cloud data and the normal working characteristics of the equipment S i [52,18,37,35°,60,5,78,,34,36,…,78], abnormal working state feature Y i [51,16,38,36°,90,6,,85,35,38,…,80], and the analysis report is uploaded to the user end for operation and maintenance personnel to review and formulate maintenance measures.

[0083] The embodiments described above are only descriptions of the preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Without departing from the design spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by ordinary technicians in this field should all fall within the protection scope determined by the claims of the present invention.

Claims

1. An equipment fault detection system for multi-agent collaboration in an integrated energy system, characterized in that: include: Sub-device, parent device and user background; The sub-device is used for equipment inspection, obtaining equipment fault information, segmenting the fault area where the faulty equipment is located, and obtaining the spatial coordinates of the sub-device; The parent device receives the spatial coordinates, performs a three-dimensional scan of the fault area in conjunction with the child device based on the spatial coordinates, and transmits the scan result to the user background; The user backend generates an abnormality diagnosis report based on the device fault information and the scanning result; The mother device uses a second drone device, including: a second drone body, a laser radar, a second wireless communication device, and a second micro-computing unit; The second drone body is used to carry the laser radar, the second wireless communication device and the second micro computing unit; The second wireless communication device is used to receive the fault area and the spatial coordinates; The second micro-computing unit is used to calculate the optimal route based on the spatial coordinates; The laser radar is used to perform a three-dimensional scan on the fault area to obtain a three-dimensional scan image; The second micro-computing unit is further used to complete the point cloud of the three-dimensional scanned image using a point cloud game completion method to obtain the scan result; The second wireless communication device is also used to transmit the scanning result to the user background; The sub-device uses a first drone device, including: a first drone body, an infrared thermal imaging camera, a natural light camera, an acoustic sensor, an electromagnetic field sensor, a first wireless communication device, a first micro-computing unit, a high-precision GPS, and an inertial navigation device; The first drone body is used to carry the infrared thermal imaging camera, the natural light camera, the acoustic sensor, the electromagnetic field sensor, the first wireless communication device, the first micro-computing unit, the high-precision GPS and the inertial navigation device; the infrared thermal imaging camera is used to detect thermal failures of the equipment; The natural light camera is used to detect appearance damage and deformation failure of the equipment; The acoustic sensor is used to detect mechanical vibration faults of the equipment; The electromagnetic field sensor is used to detect magnetic field faults of the equipment; The high-precision GPS and the inertial navigation device are used to obtain the spatial coordinates of the first drone device in real time; The first wireless communication device is used to transmit the fault area and the spatial coordinates to the mother device, and transmit the device fault information to the user background; The first micro-computing unit divides the integrated energy system space into a number of voxel spaces of preset volumes, each of which sets the normal working characteristics of the device. , where S i represents the eigenvector of the i-th voxel, x, y and z represent the spatial coordinates of the drone, cx represents the direction of the drone, t represents the operating temperature of the device, cc represents the magnetic field state of the device when it is working normally, zd represents the vibration state value of the device, p1, p2…p n Image features for proper operation of the device.

2. According to claim 1, the equipment fault detection system for multi-agent collaboration in the integrated energy system is characterized in that: The sub-device detects the real-time working characteristics in the current voxel space in real time and the device works properly with the characteristic S i Perform Euclidean distance calculation, and when the calculated value exceeds a preset threshold value α, determine that the device is faulty, and record the faulty area and the device fault information.

3. A multi-agent collaborative equipment fault detection method for an integrated energy system, the detection method being applied to the detection system according to any one of claims 1 to 2, characterized in that: The steps include: Inspect the equipment, obtain equipment fault information, segment the fault area where the faulty equipment is located, and obtain the spatial coordinates of the sub-equipment; Performing a three-dimensional scan of the fault area based on the spatial coordinates; An abnormality diagnosis report is generated based on the device fault information and the scanning result.

4. The device fault detection method for multi-agent collaboration in an integrated energy system according to claim 3 is characterized in that: Using the sub-device to inspect the equipment, obtain the thermal fault, appearance damage and deformation fault, mechanical vibration fault and magnetic field fault of the equipment, and obtain the spatial coordinates of the sub-device; The sub-device uses the first drone device, including: a first drone body, an infrared thermal imaging camera, a natural light camera, an acoustic sensor, an electromagnetic field sensor, a first wireless communication device, a first micro-computing unit, a high-precision GPS and an inertial navigation device.

5. The device fault detection method for multi-agent collaboration in an integrated energy system according to claim 4 is characterized in that: The method for obtaining the fault area includes: The integrated energy system space is divided into a number of voxel spaces of preset volumes, each of which sets the normal working characteristics of the equipment. , where S i represents the eigenvector of the i-th voxel, x, y and z represent the spatial coordinates of the drone, cx represents the direction of the drone, t represents the operating temperature of the device, cc represents the magnetic field state of the device when it is working normally, zd represents the vibration state value of the device, p1, p2…p n Image features for proper operation of the device; Real-time detection of real-time working features in the current voxel space and the device works properly with the characteristic S i Perform Euclidean distance calculation, and when the calculated value exceeds a preset threshold value α, determine that the device is faulty, and record the faulty area and the device fault information.

6. The device fault detection method for multi-agent collaboration in an integrated energy system according to claim 3 is characterized in that: Using the mother device to perform a three-dimensional scan on the device, calculating the optimal route based on the spatial coordinates, performing a three-dimensional scan on the fault area to obtain a three-dimensional scanned image, and performing point cloud completion on the three-dimensional scanned image using a point cloud game completion method to obtain the scan result; The mother device uses a second drone device, including: a second drone body, a laser radar, a second wireless communication device and a second micro computing unit.

Images