Switch cabinet automatic inspection method and system based on inspection robot

Abstract

The invention relates to a switch cabinet automatic inspection method and system based on an inspection robot, and relates to the field of transformer substation inspection, and the method comprises the following steps: an inspection path allocation step: according to a preset path generation strategy and an inspection target, carrying out analysis to allocate an inspection path corresponding to the inspection target, and instructing the inspection robot to execute inspection operation; an unlocking environment analysis step in which a switch cabinet environment analysis strategy is configured and is used for detecting and analyzing the safety of the switch cabinet environment so as to determine whether to trigger the information acquisition and analysis step of the switch cabinet; an unlocking identification adaptation step: performing point cloud image analysis according to the switch lock image to determine and identify a key hole area and trigger a mechanical unlocking instruction to unlock the switch cabinet; and an information acquisition and analysis step: performing multi-source sensing detection according to the detection target to generate a detection instruction sequence, acquiring detection information and outputting a detection result. The method has the effects of reducing the investment of personnel auxiliary inspection and improving the inspection efficiency and the inspection accuracy of the power system.

Description

Technical Field

[0001] This application relates to the field of substation inspection technology, and in particular to an automatic inspection method for switchgear based on an inspection robot. Background Technology

[0002] The automatic inspection method for switchgear is a power equipment maintenance technology based on inspection robots. It is designed to address the problems of low efficiency, high safety risks, and strong subjectivity in anomaly identification of manual inspection of switchgear in electric field environments. It helps to improve the safety, accuracy, and intelligence level of switchgear inspection in high-voltage electric field scenarios.

[0003] In related technologies, existing switchgear inspection methods are mostly manual, supplemented by simple portable testing equipment. Staff need to be in close contact with the switchgear in a high-voltage electric field environment to check the equipment status through visual observation and manual measurement. Although inspection robots have been introduced in some scenarios, they can only monitor the surface parameters of the external environment of the cabinet or the unopened state. They cannot autonomously complete the opening operation of the switchgear and are not fully adapted to the characteristics of electromagnetic interference and safety distance requirements of the electric field environment. In practical applications, when detecting core abnormalities such as internal wiring connections, component aging, and partial discharge in the switchgear, staff still need to manually open the cabinet door and then complete the internal inspection by robot or manual means. This not only faces safety hazards such as high-voltage electric shock and electromagnetic radiation, but also the judgment of abnormalities depends on the experience of the staff and is easily affected by electric field interference and human factors.

[0004] Regarding the aforementioned technologies, the safety risks of manual unpacking during the inspection of switchgear in electric field environments are significantly increased. The manual unpacking and manual-assisted testing modes result in a cumbersome and time-consuming inspection process. Existing inspection robots lack integrated automatic unpacking and accurate anomaly detection capabilities adapted to electric field characteristics. They cannot meet the safety requirements of high-voltage environments, nor can they achieve full automation of the switchgear inspection process. This not only affects inspection efficiency but may also lead to missed inspections and misjudgments due to electromagnetic interference, resulting in the failure to promptly identify potential equipment faults and hindering the stable operation of the power system. Summary of the Invention

[0005] To reduce the investment in personnel-assisted inspections and improve the inspection efficiency and accuracy of power systems, this application provides an automatic inspection method and system for switchgear based on an inspection robot.

[0006] Firstly, this application provides an automatic inspection method for switchgear based on an inspection robot: An automated inspection method for switchgear based on an inspection robot is applied to the automated inspection robot, which is equipped with an autonomous inspection module for controlling autonomous movement, an image acquisition module for acquiring images, an infrared temperature measurement module for detecting temperature, and a robotic arm module for unlocking the switchgear, including: The inspection path allocation step involves analyzing the preset path generation strategy and inspection target to allocate the inspection path corresponding to the inspection target and triggering the inspection command to instruct the inspection robot to match and execute the inspection operation. The unlocked environment analysis step is configured with a switch cabinet environment analysis strategy to detect and analyze the security of the switch cabinet environment in order to determine whether to trigger the switch cabinet information collection and analysis step. The unlocking and identification adaptation steps involve performing point cloud image analysis based on the collected switch and lock images to determine the keyhole area and trigger a mechanical unlocking command to unlock the switch cabinet. The information acquisition and analysis steps involve multi-source sensing detection based on the detection target to determine the component detection type and elements of the switchgear, in order to generate a detection instruction sequence, collect detection information, and output the detection results.

[0007] By adopting the above technical solutions, the inspection path allocation step combines a preset path generation strategy with the inspection target to allocate paths and match robots, solving the problems of low efficiency and poor adaptability between robots and inspection targets in traditional manual path planning, and achieving precise allocation of inspection resources. The unlocking environment analysis step detects environmental safety through switchgear environment analysis strategies, avoiding blind operation in unsafe environments such as high temperature and strong electric fields, reducing equipment damage and potential safety risks. The unlocking identification and adaptation step locates keyhole gaps and triggers mechanical unlocking based on point cloud analysis of switchgear images, overcoming the limitations of low efficiency and poor adaptability to different specifications of mechanical locks in traditional manual unlocking, and improving the accuracy and automation level of mechanical lock switchgear unlocking. The information collection and analysis step determines the component detection type and elements through multi-source sensing and generates detection command sequences, solving the problems of incomplete coverage and disordered data collection in traditional single detection methods, achieving comprehensive and standardized detection of switchgear components, helping to reduce manual intervention, effectively improving the safety, automation level and detection accuracy of switchgear inspection in high-voltage electric field environments, and providing reliable protection for the stable operation of the power system.

[0008] Optionally, the path generation strategy includes: Based on the detection targets and the preset site inspection map, planning is carried out to generate different movement paths that pass through multiple detection targets. The path optimization model is used to traverse the path and eliminate overlapping paths to obtain the best movement path that passes through multiple detection targets. Based on the optimal movement path, the detection order of multiple detection targets is arranged to obtain the target sorting information when the switch cabinet is detected, and the path update strategy is triggered to update the optimal movement path in real time.

[0009] By adopting the above technical solution, the path generation strategy generates multiple sets of movement paths based on the detection target and the preset site inspection map. Then, the path optimization model eliminates overlapping paths to obtain the optimal movement path, which solves the problems of many overlapping road segments and long inspection time in traditional path planning. This significantly improves the economy of the path and the inspection efficiency. At the same time, the optimal movement path is used to arrange the detection order of multiple detection targets and trigger the path update strategy to update the path in real time, avoiding missed detections or duplicate detections caused by disordered detection of inspection targets.

[0010] Optionally, the path update strategy includes: Multiple switchgear cabinets are matched for testing status according to the testing order to determine the corresponding allowed testing status in the component testing database. The switch cabinet is marked based on the allowed detection status, and the operating status of the switching components in the switch cabinet is monitored. When the operating status of the component is inconsistent with the allowed detection status, it is marked to identify the abnormal status detection target. Targets are filtered out based on abnormal state detection, and the path generation strategy is retried to update the optimal movement path.

[0011] By adopting the above technical solution, the path update strategy constructs a dynamic adjustment mechanism of detection status matching, abnormal target marking, and path regeneration. First, the detection status of the switchgear is matched according to the detection sequence to determine the permissible detection status in the component detection database. This ensures that inspections are only conducted when the equipment is in a safe operating state, avoiding invalid inspection data or equipment damage due to unsuitable equipment status. Then, by monitoring the operating status of components, switchgear with inconsistent detection statuses is marked as abnormal detection targets. After filtering, the path generation strategy is re-triggered to update the optimal movement path, solving the problem of invalid inspections caused by the lack of updates to traditional fixed paths. This ensures that the inspection path always adapts to the actual operating status of the equipment, reducing unnecessary inspection resource consumption and guaranteeing efficient progress of inspection tasks.

[0012] Optionally, the information collection and analysis step is further configured with a component abnormal change analysis strategy, including: Feature analysis is performed on the detection images from multi-source sensing to determine the three-dimensional contour coordinates of the switching elements in the switch cabinet. Based on the three-dimensional contour coordinates of the switching element, at least one set of coordinates is selected to construct the absolute coordinate information of the element. The absolute coordinate information of the element is used to reflect the current contour features of the element. An absolute position coordinate library is generated by associating and storing the absolute coordinate information of the switch components. When the switch cabinet is inspected for the second time, the absolute coordinate position information of the switch components is compared to determine the component outline deviation information and generate a switch position abnormality prompt.

[0013] By adopting the above technical solution, the component abnormal change analysis strategy analyzes the three-dimensional contour coordinates of the switching components based on multi-source sensor detection images. This overcomes the limitations of traditional two-dimensional images, which can only judge surface features and cannot accurately reflect spatial positions, thus improving the accuracy of component contour capture. At the same time, it constructs the absolute coordinate information of the components through three-dimensional contour coordinates, and generates an absolute position coordinate library by association and storage, providing a benchmark reference for component position comparison. When the switch cabinet is inspected for the second time, the contour deviation is determined by absolute coordinate comparison, generating a switch position abnormality prompt. This can identify the position abnormality of the components caused by vibration, aging or external force, and provide early warning of potential faults.

[0014] Optionally, the switchgear environment analysis strategy includes: Collect environmental safety parameters inside the switchgear where the target is located, including ambient temperature and ambient voltage field strength; Based on a comparative analysis of preset allowable thresholds and environmental safety parameters, calculations are performed to determine the safety parameter difference when the environmental safety parameters exceed the preset allowable thresholds. Based on the comparison between the safety parameter difference and the preset effective measurement difference range, the information collection and analysis step is triggered when the safety parameter difference is less than the preset effective measurement difference range, and the switchgear safety prompt is triggered otherwise.

[0015] By adopting the above technical solution, the core safety risk points of switchgear inspection under high-voltage electric field environment are covered by collecting the ambient temperature and voltage field strength inside the switchgear. This avoids the problem of traditional manual inspection neglecting environmental safety and blindly carrying out tests. The collected environmental safety parameters are then compared with preset allowable thresholds to calculate the safety parameter difference and clarify the quantitative basis for whether the environment is safe or not. Finally, by comparing the safety parameter difference with the preset effective measurement difference range, the information collection and analysis step is triggered only when the difference is less than the range, avoiding equipment damage or data distortion caused by testing in unsafe environments such as high temperature and strong electric field.

[0016] Optionally, environmental disturbance compensation strategies may also be included, including: Based on the detection target, the switchgear environment is analyzed to determine environmental interference information, including electromagnetic interference information and electric field interference information. Interference compensation data is calculated based on a preset interference compensation model and environmental interference information. The interference compensation data includes robotic arm control signals and multi-source sensor detection data. Based on interference compensation data, the inspection robot monitors inspection signals, and corrects the robotic arm control signals and multi-source sensor detection data when the unlocking and identification steps are triggered.

[0017] By adopting the above technical solutions, the environmental interference compensation strategy specifically addresses electromagnetic and electric field interference issues in high-voltage electric field environments, improves the accuracy of inspection data and motion control, and compensates for interference in signal transmission and acquisition to prevent electromagnetic and electric field interference. This ensures that the robotic arm's unlocking action is not disturbed and that the detection data accurately reflects the actual state of the components, thereby enhancing the adaptability of the inspection robot in complex electric field environments and reducing the impact of environmental interference on inspection accuracy.

[0018] Optionally, the interference compensation model includes a signal filtering correction sub-model and an electromagnetic interference correction sub-model, which are calculated using the following formulas: The signal filtering correction sub-model is as follows: ; The electromagnetic interference correction sub-model is as follows: ; in, This is the filtered control signal for the robotic arm. The raw control signals for the robotic arm. To detect the angular frequency of the control signal, The preset filter resistor value, The preset filter capacitor value, For the compensated multi-source sensor detection data, These are actual measured data. The preset electric field compensation coefficient is used. Given the current electric field strength, This is the set calibration electric field strength reference value.

[0019] By adopting the above technical solutions, the signal filtering correction sub-model of the interference compensation model organically combines the original control signal of the robotic arm, the signal angular frequency, and the preset filter resistor and capacitor values ​​through formulas, thereby achieving precise filtering of the control signal and solving the problems of traditional filtering relying on experience and unstable filtering effect. The electromagnetic interference correction sub-model integrates measured data, electric field compensation coefficient, and current and calibrated electric field strength through formulas, thereby achieving precise electromagnetic compensation of multi-source sensor detection data and avoiding the limitations of traditional electromagnetic compensation that lacks quantitative standards and has large compensation errors.

[0020] Optionally, the information collection and analysis step is further configured with a component anomaly verification strategy, including: Matching is performed based on the component detection type to determine the corresponding operating state test current and operating state test voltage for the detection component type in the preset attachment verification database; The detection elements of the switchgear are tested and adjusted based on the operating state test current and operating state test voltage, and the test temperature of the detection elements is used to generate temperature change curves. The curvature is calculated based on the temperature change curve to determine the test temperature change rate. When the test temperature change rate exceeds the preset standard change rate, the detection element is marked as an abnormal element, and abnormal element feedback information is generated.

[0021] By adopting the above technical solution, the component anomaly verification strategy first matches the preset working state test current and voltage according to the component detection type, solving the problem of traditional test parameters not matching the component type and resulting in invalid test results, and ensuring that the test conditions fit the actual working requirements of the component; then, based on the test current and voltage, the component is adjusted and a temperature change curve is generated, realizing the recording of the dynamic temperature change process of the component, breaking through the limitation of traditional methods that only detect static temperature and cannot reflect the temperature trend under the working state of the component; finally, by calculating the temperature change rate and comparing it with the standard change rate, abnormal components are marked, which can promptly detect the potential risks of abnormal temperature changes of components, avoiding the problem of traditional methods that rely solely on temperature values ​​and miss hidden anomalies such as sudden temperature rises and falls.

[0022] Optionally, the component anomaly verification strategy further includes an anomaly level evaluation sub-strategy, including: Multiple anomaly levels are obtained through analysis of multi-source sensing detection by the detection elements, and a comprehensive anomaly level is calculated based on the anomaly level evaluation model, which uses the following formula: ; in, To determine the overall anomaly level, to These are preset weighting coefficients for circuit abnormalities, component aging, partial discharge abnormalities, and temperature abnormalities. The line connection is at an abnormal level. The component is considered to be at an abnormal aging level. It is classified as a partial discharge abnormality. This indicates an abnormal temperature level.

[0023] By adopting the above technical solution, the anomaly level assessment sub-strategy first obtains multiple anomaly levels such as circuit anomaly, component aging, partial discharge anomaly, and temperature anomaly through multi-source sensor detection, which solves the problem that the traditional single-dimensional assessment of anomalies cannot fully reflect the severity of component failures. Then, by combining the formula with preset weight coefficients, the comprehensive anomaly level is calculated, realizing the quantitative integration of the degree of anomaly, and avoiding the limitations of the traditional subjective judgment of anomaly level and lack of unified standards.

[0024] Secondly, this application provides an automatic inspection method and system for switchgear based on an inspection robot, which adopts the following technical solution: An automated inspection system for switchgear based on an inspection robot includes: The inspection path allocation module analyzes the preset path generation strategy and inspection target to allocate the inspection path corresponding to the inspection target and triggers the inspection command to instruct the inspection robot to match and execute the inspection operation. The unlocked environment analysis module is configured with a switch cabinet environment analysis strategy to detect and analyze the safety of the switch cabinet environment in order to determine whether to trigger the switch cabinet information collection and analysis steps. The unlocking and recognition adaptation module performs point cloud image analysis based on the collected switch and lock images to determine the keyhole area and trigger a mechanical unlocking command to unlock the switch cabinet; The information acquisition and analysis module determines the component detection type and elements of the switchgear based on the detection target through multi-source sensing detection, in order to generate a detection instruction sequence, collect detection information, and output the detection results.

[0025] By adopting the above technical solutions, the path allocation module ensures that the inspection path is optimized and accurately matched with the robot, solving the problem of disorder in traditional path planning; the unlocking environment analysis module ensures the safety of the inspection environment and avoids safety risks; the unlocking identification and adaptation module enables precise unlocking of switchgear, improving operational accuracy; and the information collection and analysis module completes multi-source information collection and analysis, ensuring comprehensive and accurate detection. This realizes full-process automation of switchgear inspection, solving the problems of low efficiency, large errors, and poor safety of traditional manual inspection, and improving the overall automation, safety, and detection accuracy of inspection operations.

[0026] In summary, this application includes at least one of the following beneficial technical effects: 1. The inspection path allocation step, unlocking environment analysis step, unlocking identification and adaptation step, and information collection and analysis step work together to realize the automatic inspection robot from allocation to switch cabinet unlocking and detection of components. This eliminates the need for staff intervention to assist in the inspection, saving manpower and material resources. At the same time, it can effectively carry out comprehensive detection from multiple sources, monitor switch cabinets for anomalies, and enable the power system to operate stably and detect faults in a timely manner. 2. Component Anomaly Analysis Strategy: Based on the analysis of multi-source sensor images, the three-dimensional contour coordinates of switch components are analyzed. This overcomes the limitations of traditional two-dimensional images, which can only judge surface features and cannot accurately reflect spatial positions, thus improving the accuracy of component contour capture. At the same time, the absolute coordinate information of components is constructed through three-dimensional contour coordinates, and the absolute position coordinate library is generated by association and storage. This provides a benchmark reference for component position comparison. When the switch cabinet is inspected for the second time, the contour deviation is determined by comparison of absolute coordinates, and an abnormal switch position prompt is generated. This can identify the positional abnormalities of components caused by vibration, aging, or external forces, and provide early warning of potential faults. 3. Environmental interference compensation strategy: Specifically addresses electromagnetic and electric field interference issues in high-voltage electric field environments, improves the accuracy of inspection data and motion control, and performs interference compensation and correction for signal transmission and acquisition to prevent interference with electromagnetic and electric field interference. This ensures that the robotic arm's unlocking action is not disturbed and that the detection data accurately reflects the actual state of the components. It helps improve the adaptability of the inspection robot in complex electric field environments and reduces the impact of environmental interference on inspection accuracy. Attached Figure Description

[0027] Figure 1 This is a flowchart of steps S100 to S400 in this application.

[0028] Figure 2 This is a flowchart of steps S101 to S102 in this application.

[0029] Figure 3 This is a flowchart of steps S1021 to S1023 in this application.

[0030] Figure 4 This is a flowchart of steps S401 to S403 in this application.

[0031] Figure 5 This is a flowchart of steps S201 to S203 in this application.

[0032] Figure 6 This is a flowchart of steps S204 to S206 in this application.

[0033] Figure 7 This is a flowchart of steps S404 to S406 in this application. Detailed Implementation

[0034] To make the purpose, technical solution, and advantages of this application clearer, the following description is provided in conjunction with the appendix. Figure 1-7 The present application will be further described in detail below with reference to embodiments. It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the scope of the application.

[0035] The embodiments of the present invention will now be described in further detail with reference to the accompanying drawings.

[0036] This application discloses an automatic inspection method for switchgear based on an inspection robot. The inspection path allocation step combines a preset path generation strategy with the inspection target to allocate paths and match them with the robot, solving the problems of low efficiency and poor adaptability between the robot and the inspection target in traditional manual path planning. This achieves precise allocation of inspection resources. The unlocking environment analysis step detects environmental safety through a switchgear environment analysis strategy, avoiding blind operation in unsafe environments such as high temperature and strong electric field, reducing equipment damage and potential safety risks. The unlocking identification and adaptation step locates the keyhole gap and triggers mechanical unlocking based on point cloud analysis of the switch lock image, overcoming the limitations of low efficiency and poor adaptability to different specifications of mechanical locks in traditional manual unlocking, improving the accuracy and automation level of mechanical lock switchgear unlocking. The information collection and analysis step determines the component detection type and elements through multi-source sensing and generates a detection command sequence, solving the problems of incomplete coverage and disordered data collection in traditional single detection methods, achieving comprehensive and standardized detection of switchgear components.

[0037] Reference Figure 1 The method flow of the automatic inspection method for switchgear based on inspection robots includes the following steps: Step S100: Inspection path allocation step, analyze according to the preset path generation strategy and inspection target, allocate the inspection path corresponding to the inspection target, and trigger the inspection command to instruct the inspection robot to match and execute the inspection operation. The execution of the path generation strategy relies on a pre-set site inspection map, which must mark the spatial coordinates, passage widths, and fixed obstacle locations of all switchgear within the inspection area. In practice, the site inspection map is first imported into the path planning system of the inspection robot. For example, in the inspection area of ​​a substation, there are 10 switchgears, numbered 1 to 10. The coordinates of switchgear 1 are (10.0m, 5.0m), switchgear 2 is (15.0m, 5.0m), and switchgear 3 is (15.0m, 10.0m). The passage width is uniformly 2.0m, and the obstacle is a fire hydrant located at (12.0m, 7.5m). All of the above information must be accurately marked on the map.

[0038] After determining the inspection targets, for example, if the inspection targets are switch cabinets 1 to 8, the path planning system first generates three different initial movement paths based on the inspection targets and the site inspection map. The first path is for cabinets 1-3-2-4-6-5-7-8, the second path is for cabinets 1-2-4-3-5-6-8-7, and the third path is for cabinets 2-1-3-5-4-7-6-8. Then, a preset path optimization model, such as a genetic algorithm, is used to traverse the initial paths, eliminating overlapping sections. Finally, the optimal movement path passing through all inspection targets is obtained: cabinets 1-2-3-4-5-6-7-8. This path's total length is 8.5m shorter than the initial longest path, effectively reducing the inspection robot's ineffective movement distance.

[0039] The matching process for inspection robots needs to consider the robot's current load status, assigned tasks, and inspection range coverage. For example, if there are two automatic inspection robots, Robot A and Robot B, in the inspection area, Robot A currently has no tasks and its load capacity can cover the inspection of up to 5 switch cabinets, while Robot B currently only has remaining tasks to handle cabinets 9 and 10, then after the inspection command is triggered, the system will match the inspection tasks of cabinets 1 to 4 to Robot A and the inspection tasks of cabinets 5 to 8 to Robot B. After the command is sent, the robot must respond and start moving to the first inspection target location within 5 minutes.

[0040] If a target cannot be inspected temporarily during the inspection process, such as cabinet 3 being powered off due to temporary maintenance, the system will re-trigger the path generation strategy and update the optimal movement path. The updated path will be adjusted to cabinets 1-2-4-5-6-7-8. At the same time, cabinet 3 will be marked as pending inspection. Once it is restored to the inspection-allowed state, the inspection task will be reassigned.

[0041] Step S200: Unlock the environmental analysis step. A switch cabinet environmental analysis strategy is configured to detect and analyze the safety of the switch cabinet environment in order to determine whether to trigger the switch cabinet information collection and analysis step. Environmental safety parameters are collected using sensors deployed at specific locations within the switchgear. The core parameters to be collected are the ambient temperature and voltage field strength inside the switchgear. In practice, a contact temperature sensor is installed at the center of the top of each switchgear, and an electric field strength sensor is installed 10cm from the handle on the outside of the cabinet door. The sensors are set to collect data once every 2 seconds, and the average value is taken after three consecutive collections as the final environmental safety parameters to avoid data deviations caused by momentary interference.

[0042] The preset allowable thresholds need to be determined in conjunction with the equipment specifications and safety standards of the switchgear. For example, the allowable threshold for ambient temperature of a certain type of high-voltage switchgear is set to not exceed 40℃, and the allowable threshold for ambient voltage field strength is set to not exceed 5V / m. The preset effective measurement difference range is set to the corresponding temperature parameter from 0℃ to 3℃ and the corresponding voltage field strength parameter from 0V / m to 1V / m. This range needs to be determined based on the measurement accuracy of the sensor and the tolerance margin of the equipment.

[0043] The comparative analysis process of environmental safety parameters is as follows: When the sensor collects an ambient temperature of 38℃ and an ambient voltage field strength of 4.5V / m, these values ​​are compared with the allowable thresholds. Since neither exceeds the allowable threshold, environmental safety is directly determined, triggering the information acquisition and analysis steps of the switchgear. When the collected ambient temperature is 42℃ and the ambient voltage field strength is 5.8V / m, the safety parameter difference is first calculated: the temperature difference is 42℃ - 40℃ = 2℃, and the voltage field strength difference is 5.8V / m - 5V / m = 0.8V. / m, and then compare the difference with the effective measurement difference range. If the temperature difference is 2℃ less than 3℃ and the voltage field strength difference is 0.8V / m less than 1V / m, it is still determined that information can be collected, and the information collection and analysis step is triggered. When the collected ambient temperature value is 44℃ and the ambient voltage field strength value is 6.2V / m, the temperature difference is 4℃ and the voltage field strength difference is 1.2V / m, both of which exceed the effective measurement difference range. It is determined that there is a safety risk in the environment, and the switch cabinet safety prompt is triggered. The prompt content includes the type of parameter that exceeds the standard and the specific difference value.

[0044] The installation and calibration of sensors must be completed before the inspection operation. For example, temperature sensors must be calibrated in a standard constant temperature chamber before installation to ensure that the measurement error does not exceed ±0.5℃ in the range of 20℃ to 50℃; electric field strength sensors must be calibrated in a standard electric field generator before installation to ensure that the measurement error does not exceed ±0.2V / m in the range of 0V / m to 10V / m.

[0045] Step S300: Unlocking and identification adaptation step, performing point cloud image analysis based on the collected switch lock images to determine the keyhole area and triggering a mechanical unlocking command to unlock the switch cabinet; The acquisition of lock opening and closing images is accomplished by the image acquisition module and LiDAR configured on the inspection robot. The image acquisition module uses a high-definition industrial camera with a resolution of 1920×1080 and a frame rate of 30 frames / second. The scanning accuracy of the LiDAR is set to ±0.1mm. Both are integrated into the end of the robot's robotic arm, with the installation position 15cm away from the unlocking key head of the robotic arm, ensuring that the acquired images and point cloud data can accurately correspond to the lock position.

[0046] In actual operation, the inspection robot moves to the switch cabinet, first adjusting the robotic arm's posture so that the industrial camera and LiDAR are perpendicularly aligned with the switch cabinet's lock surface, 30cm away. Then, image acquisition and laser scanning are initiated, continuously acquiring five frames of lock images and corresponding point cloud data. The frame with the highest clarity is selected as the analysis object. Taking a cross-shaped lock of a certain model of switch cabinet as an example, the distance from the lock center to the cabinet door edge is 8cm, the lock diameter is 2cm, and the keyhole is a cross-shaped groove with a width of 5mm and a depth of 8mm.

[0047] The point cloud image analysis process is as follows: First, the acquired lock image is processed by grayscale conversion and edge detection to extract the lock's contour features. Then, combined with the point cloud data generated by the LiDAR, a three-dimensional model of the lock is constructed. The spatial coordinates of the keyhole area are determined through model analysis. For example, a local coordinate system is established with the lock center as the origin. The four slots of the cross-shaped hole correspond to the positive x-axis, negative x-axis, positive y-axis, and negative y-axis of the coordinate system, respectively. The center coordinates of each slot are (2.5mm, 0mm), (-2.5mm, 0mm), (0mm, 2.5mm), and (0mm, -2.5mm), respectively, and the slot depth coordinates are (0mm, 0mm, -8mm).

[0048] After identifying the keyhole area, the system generates a mechanical unlocking command. This command includes the robotic arm's movement path and unlocking parameters: the robotic arm first adjusts its position according to the spatial coordinates of the keyhole area, aligning the unlocking key head with the center of the lock, with a movement accuracy controlled within ±0.5mm. Then, it drives the key head to insert into the keyhole to a depth of 8mm, and after insertion, rotates it 90 degrees clockwise at a speed of 10 degrees / second. After reaching the desired position, it holds for 2 seconds. Once the lock is confirmed unlocked, the key head is removed, completing the unlocking operation. If the initial point cloud image analysis fails to accurately identify the keyhole area, for example, due to contour extraction deviation caused by dirt on the lock surface, the system will trigger a re-acquisition process. Step S400: Information acquisition and analysis step. Based on the detection target, multi-source sensing detection is performed to determine the component detection type and elements of the switchgear, so as to generate a detection instruction sequence, collect detection information, and output the detection results.

[0049] Multi-source sensing detection relies on the image acquisition module, infrared temperature measurement module, and current sensor configured in the inspection robot. The detection type and elements for different components need to be determined according to the component composition of the switchgear. For example, a switchgear contains three core components: circuit breakers, disconnectors, and grounding switches. The detection type for circuit breakers is electrical performance and temperature detection, and the detection elements are operating current, terminal temperature, and appearance integrity. The detection type for disconnectors is mechanical condition and temperature detection, and the detection elements are opening and closing status, contact temperature, and appearance integrity. The detection type for grounding switches is mechanical condition detection, and the detection elements are grounding status and appearance integrity.

[0050] The generation of the detection command sequence needs to be based on the importance of the components and the detection logic. For example, the circuit breaker with the greatest impact on power supply safety should be detected first, followed by the disconnecting switch, and finally the grounding switch. The specific command sequence is as follows: 1. The infrared temperature measurement module detects the temperature of the circuit breaker terminals; 2. The current sensor detects the circuit breaker's operating current; 3. The image acquisition module captures the appearance of the circuit breaker; 4. The infrared temperature measurement module detects the temperature of the disconnecting switch contacts; 5. The image acquisition module identifies the open / closed status of the disconnecting switch; 6. The image acquisition module captures the appearance of the disconnecting switch; 7. The image acquisition module identifies the grounding status of the grounding switch; 8. The image acquisition module captures the appearance of the grounding switch.

[0051] In actual operation, the sensor's operating parameters and methods need to be adjusted according to the component's position: When the infrared temperature measurement module detects the circuit breaker terminals, it needs to be moved to a distance of 15cm from the terminals, with the measurement accuracy set to ±0.5℃ and the measurement range set to -20℃ to 150℃; the current sensor needs to be connected in series in the circuit breaker's power supply circuit, with the measurement range set to 0A to 500A and the measurement accuracy set to ±1A; when the image acquisition module photographs the component's appearance, the angle needs to be adjusted to the front of the component to ensure a clear view of the entire component, with a resolution of 1920×1080.

[0052] The specific process of collecting and processing detection information is as follows: three sets of data are collected for each detection element, and the average value is taken as the final detection value. For example, the three collected values ​​of the circuit breaker terminal temperature are 35℃, 36℃, and 35℃, and the average value is 35.3℃. If a set of data exceeds the normal range, for example, if the current collection value is 550A, it exceeds the 500A measurement range, then the set of data is discarded, and the average value is calculated using the remaining two sets of data.

[0053] After data collection, the detected values ​​are compared with preset acceptable thresholds. For example, the acceptable threshold for circuit breaker operating current is 0A to 400A, and the acceptable threshold for terminal temperature is 0℃ to 40℃. If the detected value is within the threshold range, an acceptable result is output; if the detected value exceeds the threshold, for example, the circuit breaker terminal temperature is 45℃, an abnormal result is output, and the abnormal parameter type and specific value are marked.

[0054] The execution of the detection command sequence must be carried out after the switch cabinet door is opened. The door opening is completed by the previous unlocking and identification adaptation steps to ensure that the sensor can directly detect the internal components. If a certain detection command fails to be executed, for example, if the current sensor does not collect data, the system will pause the command and prioritize the execution of subsequent commands. After all commands have been executed, the execution of the failed command will be retried, with a maximum of 2 retries.

[0055] Reference Figure 2 The path generation strategies include: Step S101: Based on the detection targets and the preset site inspection map, plan to generate different movement paths that pass through multiple detection targets, and use the preset path optimization model to traverse the path to eliminate overlapping paths and obtain the best movement path that passes through multiple detection targets. The pre-set site inspection map must include the spatial location information of all inspection targets within the inspection area, the distribution of fixed obstacles, and the parameters of passable passages. Taking the inspection area of ​​a 10kV substation switchgear as an example, the inspection targets in this area are 8 switchgears, numbered 1 to 8, with spatial coordinates of (4.0m, 2.0m), (7.0m, 2.0m), (7.0m, 5.0m), (10.0m, 5.0m), (10.0m, 8.0m), (13.0m, 8.0m), (13.0m, 11.0m), and (16.0m, 11.0m). The fixed obstacles in the area are 2 distribution cabinets with coordinates of (6.0m, 4.0m) and (12.0m, 7.0m), respectively. The width of the passable passage is uniformly 2.5m. All of the above information must be fully marked in the site inspection map and imported into the path planning system of the inspection robot.

[0056] After the detection targets are determined, the path planning system first generates three initial movement paths based on the principle of covering all detection targets and avoiding obstacles. Each initial path must be verified by an obstacle avoidance algorithm to ensure that the safe distance between the robot and obstacles during movement is not less than 0.5m. For example, in the first initial path, when the robot moves from cabinet 2 (7.0m, 2.0m) to cabinet 3 (7.0m, 5.0m), the horizontal distance between the path along the positive y-axis and the distribution cabinet (6.0m, 4.0m) is 1.0m, which meets the safe distance requirement. The preset path optimization model adopts the ant colony algorithm, which simulates the path selection mechanism of ants foraging and analyzes the three initial paths. During the traversal, the algorithm uses the shortest total path length as its objective function, calculating the total length of each initial path: the first initial path has a total length of 24.0m, the second initial path has a total length of 28.5m, and the third initial path has a total length of 31.2m. Simultaneously, it identifies and eliminates overlapping road segments. For example, both the second and third initial paths contain a segment from cabinet 2 to cabinet 3; the algorithm adjusts the weight of this segment during optimization to avoid redundant calculations. Finally, through iterative calculation, the algorithm determines the first initial path as the optimal movement path passing all detected targets, with a total length 4.5m and 7.2m shorter than the other two initial paths, respectively, thus minimizing the inspection robot's movement time.

[0057] In practice, the path optimization model is set to iterate 50 times. After each iteration, the path length deviation is calculated. When the path length deviation is less than 0.1m for 5 consecutive iterations, the iteration stops and the optimal movement path is output. If the number of targets to be inspected increases or the site environment changes, such as the addition of switch cabinet No. 9 (19.0m, 11.0m), the updated site inspection map needs to be re-imported, and the above initial path generation and optimization process needs to be repeated to ensure that the optimal movement path always adapts to the current inspection requirements.

[0058] This step addresses the problems of overlapping road segments, long total length, and low efficiency inherent in traditional manual path planning. By collaborating with the path planning system and optimization model, it generates an optimal movement path that balances full coverage and high efficiency, laying the foundation for subsequent detection order arrangement and real-time path updates. The target sorting information in subsequent step S102 must be based on the optimal movement path output in this step, and the path planning and optimization logic of this step must be invoked again when the path update strategy is triggered.

[0059] Step S102: Based on the optimal movement path, the detection order of multiple detection targets is arranged to obtain the target sorting information when the switch cabinet is detected, and the path update strategy is triggered to update the optimal movement path in real time.

[0060] The generation of target sorting information must completely follow the optimal movement path order determined in step S101. This sorting information must be synchronized to the task scheduling module of the inspection robot. The robot will move to each detection target position in this order. After completing the inspection of each detection target, the robot will automatically trigger the movement command of the next detection target.

[0061] The path update strategy is triggered by the inspection robot's status monitoring module. This module communicates in real time with the status monitoring units of each switchgear to obtain the switchgear's permitted inspection status, including whether inspection is permitted or prohibited. The criteria for determining the permitted inspection status are whether the switchgear is in normal operation and whether there are no temporary maintenance tasks. For example, when switchgear No. 2 sends a prohibited inspection signal due to line maintenance, the status monitoring module receives the signal and immediately triggers the path update strategy.

[0062] After the new optimal movement path is generated, the system automatically updates the task scheduling module of the inspection robot. The robot performs inspection work according to the new target sorting information and movement path. At the same time, the switch cabinets that are prohibited from being inspected are marked as pending re-inspection. The status monitoring module continuously monitors their inspection status. For example, when the inspection signal of cabinet No. 2 is received, the path update strategy is re-triggered to insert cabinet No. 2 into a reasonable position in the current target sorting information. After cabinet No. 3 is inspected, the latest optimal movement path containing cabinet No. 2 is generated.

[0063] This step addresses the problem that traditional fixed inspection sequences cannot adapt to changes in switchgear status. It ensures the orderly nature of inspection operations through target sequencing information and achieves real-time adjustment of optimal movement paths through path update strategies. This prevents inspection delays or resource waste caused by a switchgear being prohibited from inspection, thus guaranteeing the continuity and efficiency of inspection operations. Reference Figure 3 The path update strategy includes: Step S1021: Match the detection status of multiple switch cabinets according to the detection order to determine the corresponding allowed detection status in the component detection database; The inspection sequence is determined by a preceding path generation strategy. This sequence information is synchronized to the inspection robot's status matching module, which connects to the component inspection database via a switchgear number association query. The component inspection database pre-stores basic information and permitted inspection statuses for all switchgear, categorized as permitted and prohibited. The status matching module extracts the switchgear number sequentially according to the inspection order, sends a query request to the database, and the database retrieves the corresponding permitted inspection status and returns it to the module within one second. In actual operation, the component inspection database needs to synchronize data with the substation operation and maintenance system in real time to ensure accurate status matching. If the database feedback times out, the module defaults to the permitted inspection status of the switchgear being pending confirmation and temporarily excludes it from the current inspection sequence.

[0064] Step S1022: Mark the switch cabinet based on the allowed detection status and monitor the component operating status of the switch components in the switch cabinet. When the component operating status is inconsistent with the allowed detection status, mark it to identify the abnormal status detection target. The detection status marking is completed collaboratively by the robot's display panel and the task scheduling module: switch cabinets that are allowed to be detected are marked as pending detection and a detection task is added; those that are prohibited from detection are marked as temporarily removed and not yet detected; and those awaiting confirmation are temporarily marked as pending. The operating status of components relies on pre-installed status sensors within the switch cabinet, such as circuit breaker and disconnector status sensors, which transmit real-time data to the robot via wireless communication. If the sensor data is inconsistent with the preset operating status of the allowed detection status for three consecutive times, or if the sensor malfunctions and fails to transmit data, the robot immediately marks the switch cabinet as a target for abnormal status detection.

[0065] Step S1023: Filter out targets based on state anomaly detection and re-trigger the path generation strategy to update the best movement path.

[0066] First, a list of abnormal targets is compiled and removed from the current detection sequence to obtain a temporary detection sequence. Then, the previous path generation strategy is re-triggered. The path planning system generates an initial movement path based on the temporary detection sequence and the site inspection map. After traversal analysis by the path optimization model, a new optimal movement path is determined. After the new optimal movement path is generated, the system automatically updates the robot's task scheduling and path navigation modules. The robot terminates the execution of the original path and moves according to the new path. At the same time, the removed abnormal detection targets are stored in the abnormal target ledger. Once the abnormality is resolved, the path update strategy is re-triggered to include them in the detection sequence.

[0067] Reference Figure 4 The information collection and analysis steps are also equipped with a component abnormal change analysis strategy, including: Step S401: Perform feature analysis based on the detection images during multi-source sensing to determine the three-dimensional contour coordinates of the switching elements in the switch cabinet; Multi-source sensing detection relies on the robot's high-definition industrial camera and LiDAR. During data acquisition, the robotic arm's posture is adjusted to ensure the entire component is included in the acquisition range. Feature analysis consists of two steps: first, the 2D image is preprocessed by converting it to grayscale and performing edge detection; then, the LiDAR point cloud data is aligned with the 2D contour to construct a 3D point cloud model of the component, extracting contour feature points and calculating 3D coordinates. The 3D coordinate system is set with a fixed reference point inside the switch cabinet as the origin. The 3D contour coordinates of the same component are acquired three times consecutively, and the average value is taken as the final result to avoid errors from a single acquisition.

[0068] Step S402: Select at least one set of coordinates based on the three-dimensional contour coordinates of the switching element to construct the absolute coordinate information of the element. The absolute coordinate information of the element is used to reflect the current contour features of the element. Key coordinate groups are selected based on the principles of covering the main functional parts of the components, ensuring high coordinate stability, and reflecting positional changes, to construct the absolute coordinate information of the components. The absolute coordinate information must be associated with the unique identifier of the component, forming an identifier-coordinate correspondence. If the acquisition of a certain set of coordinates fails, the robotic arm's posture is readjusted; after three consecutive failures, a backup feature point is selected to ensure that the absolute coordinate information contains valid coordinates.

[0069] Step S403: Generate an absolute position coordinate library based on the absolute coordinate information of the switch element position and the associated storage of the element. When the switch cabinet is inspected for the second time, the absolute coordinate position information of the switch element is compared to determine the component outline deviation information and generate a switch position abnormality prompt.

[0070] The absolute position coordinate library is deployed in the robot's local storage module and stored in a structured format, containing component identifiers, absolute coordinates, acquisition time, and switch cabinet number. Secondary inspections are performed within a preset period after the initial inspection. First, the initial absolute coordinate information is retrieved, then secondary coordinate information is acquired, and the deviation value is calculated using a coordinate comparison algorithm.

[0071] If the deviation exceeds a preset threshold, the system generates an abnormal switch position warning, including the component identifier, the location of the deviation, and the deviation value. If the deviation is within the threshold range, the component position is determined to be normal, and the acquisition time of that component in the coordinate library is updated.

[0072] Reference Figure 5 The switchgear environment analysis strategy includes: Step S201: Collect environmental safety parameters inside the switchgear where the target is located, including ambient temperature and ambient voltage field strength; The collection of environmental safety parameters relies on sensors deployed in specific locations within the switchgear: ambient temperature values ​​are collected through contact temperature sensors, which are installed at the top center of the switchgear to ensure they are not directly affected by localized heating of components; ambient voltage field strength is collected through electric field strength sensors, which are installed on the outside of the cabinet door near the handle area to prevent collision damage when the cabinet door is opened.

[0073] During the data collection process, the sensor acquires data every 2 seconds at a preset frequency. After three consecutive collections, the average value is taken as the final environmental safety parameter to eliminate single-data deviations caused by transient interference. The collected parameters are transmitted in real time to the environmental analysis module of the inspection robot. The transmission uses wireless communication to ensure that the data transmission delay does not exceed 0.5 seconds, avoiding parameter lag from affecting subsequent analysis.

[0074] Step S202: Based on the preset allowable threshold and environmental safety parameters, a comparative analysis is performed. When the environmental safety parameters exceed the preset allowable threshold, a calculation is performed to determine the safety parameter difference. The preset allowable thresholds are determined based on the switchgear equipment specifications and power safety standards to ensure compliance with the safety boundary requirements for equipment operation. During comparative analysis, the environmental analysis module first compares the collected ambient temperature and ambient voltage field strength values ​​with the corresponding allowable thresholds one by one: if the parameters do not exceed the allowable thresholds, the environment is directly determined to be safe, and there is no need to calculate the safety parameter difference; if the parameters exceed the allowable thresholds, the difference calculation is initiated.

[0075] The difference in safety parameters is calculated by subtracting the allowable threshold from the measured value: for ambient temperature, the difference is the difference between the measured temperature value and the allowable temperature threshold; for ambient voltage field strength, the difference is the difference between the measured field strength value and the allowable field strength threshold. The difference result must be retained to one decimal place to ensure that the calculation accuracy meets the analysis requirements.

[0076] Step S203: Based on the comparison between the safety parameter difference and the preset effective measurement difference range, if the safety parameter difference is less than the preset effective measurement difference range, the information collection and analysis step is triggered; otherwise, the switchgear safety prompt is triggered.

[0077] The preset effective measurement difference range is set based on the sensor measurement accuracy and equipment tolerance margin, and is used to distinguish between the acceptable range and the risk range of parameter exceedance. During comparison, the environmental analysis module compares the safety parameter difference calculated in step S202 with the effective measurement difference range: if the difference is less than the upper limit of the range, it is determined that the degree of parameter exceedance is within the acceptable range, the environment still meets the detection conditions, and the information acquisition and analysis step of the switchgear is immediately triggered; if the difference is greater than or equal to the upper limit of the range, it is determined that there is a safety risk in the environment, and subsequent detection cannot be carried out, and the switchgear safety prompt is triggered.

[0078] Safety alerts for switchgear are triggered by the audible and visual alarm devices of the inspection robot. Simultaneously, the types of parameters exceeding the standard, measured values, allowable thresholds, and differences are uploaded to the substation operation and maintenance system. This provides data support for operation and maintenance personnel to identify risks and avoids safety hazards caused by direct operation.

[0079] Reference Figure 6 It is also equipped with environmental interference compensation strategies, including: Step S204: Perform switchgear environment analysis based on the detection target to determine environmental interference information, including electromagnetic interference information and electric field interference information; The determination of environmental interference information relies on the environmental interference detection module of the inspection robot. This module is linked with the data interface of the preceding switchgear environmental analysis step. It combines the location of the switchgear where the detection target is located, the operating status of surrounding equipment, and historical interference data to comprehensively analyze and obtain environmental interference information.

[0080] Among them, electromagnetic interference information mainly includes the frequency, intensity, and duration of interference signals, which are collected by the electromagnetic induction sensor built into the module. The sensor is deployed at the joints of the robot arm and around the multi-source sensor to ensure close proximity to the area affected by interference. Electric field interference information mainly includes the real-time fluctuation value of electric field intensity, which is extracted by continuous monitoring data from the electric field sensor. The acquisition frequency is consistent with the acquisition frequency of environmental safety parameters to avoid interference analysis deviations caused by asynchronous data acquisition.

[0081] The collected electromagnetic and electric field interference data need to be transmitted to the interference analysis unit in real time. Data compression is used during transmission to ensure that the transmission delay does not exceed 0.3 seconds. At the same time, abnormal data is initially screened, and stable interference data collected three times consecutively is retained as the final environmental interference information to provide reliable input for subsequent compensation calculations.

[0082] Step S205: Based on the preset interference compensation model and environmental interference information, interference compensation data is calculated to obtain interference compensation data, which includes robotic arm control signals and multi-source sensor detection data. The preset interference compensation model is pre-stored in the inspection robot's algorithm processing module. The model establishes corresponding data compensation logic for the different characteristics of electromagnetic interference and electric field interference. During calculation, the algorithm processing module first inputs the environmental interference information determined in step S204 into the model according to the interference type. The model combines preset equipment parameters, such as the reference value of the robotic arm control signal and the detection accuracy threshold of the multi-source sensor, and outputs interference compensation data through built-in calculation logic.

[0083] The compensation data for the robotic arm control signals must match the type of control commands given by the robotic arm, such as compensation values ​​for position control and force control. The compensation data for multi-source sensor detection data must cover the detection dimensions of various sensors, including infrared thermography and image acquisition. The numerical precision of the compensation data must be retained to two decimal places to ensure consistency with the precision level of the original signals and detection data, avoiding correction errors caused by precision mismatch. After calculation, the compensation data is temporarily stored in the robot's temporary data cache, awaiting subsequent correction triggers.

[0084] Step S206: Monitor the inspection signal of the inspection robot based on the interference compensation data, and correct the robotic arm control signal and multi-source sensor detection data when the unlocking and identification step is triggered.

[0085] Inspection signal monitoring is performed by the robot's signal monitoring unit. The monitored objects include the real-time control signal waveform of the robotic arm and the real-time detection data change trends of multiple sources of sensors. The monitoring frequency is set to 10 times per second to ensure timely capture of signal fluctuations affected by interference. When the system triggers the unlocking and identification step, the signal monitoring unit immediately retrieves the interference compensation data from the temporary data buffer and initiates the correction process.

[0086] During the correction operation, the output parameters of the control command are adjusted according to the compensation data for the robotic arm control signal. For example, the position deviation compensation value of the robotic arm movement and the fine adjustment amount of the action force are corrected to ensure that the robotic arm can still accurately align with the keyhole area in the interference environment. For multi-source sensor detection data, the deviation of infrared temperature measurement value and the pixel offset of image acquisition are corrected according to the compensation data. For example, the problem of high temperature measurement value caused by electromagnetic interference and the problem of blurred image edges caused by electric field interference are corrected.

[0087] After the correction is completed, the signal monitoring unit verifies the corrected signal and data in real time. If the fluctuation amplitude of the robotic arm control signal is less than the preset threshold and the deviation of the sensor detection data is within the allowable range, the correction is deemed effective and the unlocking and identification steps are executed normally. If the correction still does not meet the requirements, the interference compensation data is retrieved again for a second correction, and the process is repeated up to 2 times to ensure that the interference effect is effectively offset.

[0088] The interference compensation model includes a signal filtering correction sub-model and an electromagnetic interference correction sub-model, which are calculated using the following formulas: The signal filtering correction sub-model is as follows: ; The electromagnetic interference correction sub-model is as follows: ; in, This is the filtered control signal for the robotic arm. The raw control signals for the robotic arm. To detect the angular frequency of the control signal, The preset filter resistor value, The preset filter capacitor value, For the compensated multi-source sensor detection data, These are actual measured data. The preset electric field compensation coefficient is used. Given the current electric field strength, The set calibration electric field strength reference value, The imaginary unit is used only to characterize the phase relationship of AC signals.

[0089] Reference Figure 7 The information collection and analysis steps are also configured with a component anomaly verification strategy, including: Step S404: Match the components according to their detection types to determine the operating state test current and operating state test voltage corresponding to the detection component types in the preset attachment verification database; The component testing type must be determined based on the results of previous multi-source sensor testing, mainly covering core component types of switchgear such as circuit breakers, disconnect switches, and grounding switches. Each type corresponds to a unique testing identifier. A pre-set accessory verification database is deployed in the local storage unit of the inspection robot. The database stores the corresponding operating state test current and test voltage reference values ​​according to component type. The data is set according to the component's manufacturer's specifications and power equipment testing standards to ensure that it meets the parameter range for normal component operation.

[0090] The matching process is executed by the robot's parameter matching module: the module first extracts the type identifier of the currently detected component, then performs a precise search in the attached verification database using the type identifier, and retrieves the corresponding operating state test current and test voltage data for that type of component. After the search is completed, the module verifies the validity of the data, confirming that the data format and numerical range meet the test requirements. If the search result is empty or the data is invalid, the database is reloaded to ensure that accurate test parameters are obtained, providing a basis for subsequent test adjustments.

[0091] Step S405: Based on the operating state test current and operating state test voltage, test and adjust the detection elements of the switchgear, and generate a temperature change curve by detecting the test temperature of the elements; The testing and adjustment are completed using the electrical testing module of the inspection robot. This module is connected to the terminal block of the detection element through wires. The current and voltage are tested according to the working state determined in step S404. The input electrical parameters of the element are gradually adjusted: first, the current is adjusted to the test current value and kept stable, and then the voltage is adjusted to the test voltage value. During the adjustment process, the rate of change of current and voltage is controlled within the preset range to avoid damage to the element caused by sudden changes in parameters.

[0092] The component's test temperature is detected using an infrared thermography module. The module is aimed at the component's critical heat-generating areas and acquires temperature data twice per second. The acquisition duration is set according to the component's test cycle, typically 5-10 minutes. The acquired temperature data is transmitted in real-time to the data processing unit, which constructs a temperature change curve with time on the horizontal axis and temperature on the vertical axis. The curve continuously records the temperature value at each acquisition moment to ensure a complete reflection of the dynamic temperature change trend of the component during the test.

[0093] Step S406: Calculate the curvature based on the temperature change curve to determine the test temperature change rate. When the test temperature change rate exceeds the preset standard change rate, mark the detection element as an abnormal element and generate abnormal element feedback information.

[0094] Curvature calculation is performed by the robot's curve analysis module. The module selects multiple consecutive data points on the temperature change curve, calculates the curvature value of each curve segment using a curve fitting algorithm, and then converts the curvature value to obtain the test temperature change rate. The conversion process must ensure numerical accuracy, retaining one decimal place. The preset standard change rate is set based on the component material, heat dissipation performance, and safe operation requirements; the standard change rate varies for different types of components.

[0095] The curve analysis module compares the calculated test temperature change rate with the standard change rate of the corresponding component: if the change rate does not exceed the standard change rate, the component temperature change is considered normal; if the change rate exceeds the standard change rate, the tested component is immediately marked as an abnormal component. Abnormal component feedback information is constructed by the information generation module, including component identification, measured test temperature change rate, standard change rate threshold, and abnormality judgment criteria. This feedback information is synchronously uploaded to the substation operation and maintenance system via the robot's communication module, and simultaneously displayed locally on the robot's display panel, facilitating timely awareness of component abnormalities by maintenance personnel and enabling subsequent troubleshooting and maintenance work.

[0096] The component anomaly verification strategy also includes an anomaly level evaluation sub-strategy, including: Multiple anomaly levels are obtained through analysis of multi-source sensing detection by the detection elements, and a comprehensive anomaly level is calculated based on the anomaly level evaluation model, which uses the following formula: ; in, To determine the overall anomaly level, to These are preset weighting coefficients for circuit abnormalities, component aging, partial discharge abnormalities, and temperature abnormalities. The line connection is at an abnormal level. The component is considered to be at an abnormal aging level. It is classified as a partial discharge abnormality. This indicates an abnormal temperature level.

[0097] Based on the same inventive concept, embodiments of the present invention provide an automatic inspection method for switchgear based on an inspection robot, including: The inspection path allocation module analyzes the preset path generation strategy and inspection target to allocate the inspection path corresponding to the inspection target and triggers the inspection command to instruct the inspection robot to match and execute the inspection operation. The unlocked environment analysis module is configured with a switch cabinet environment analysis strategy to detect and analyze the safety of the switch cabinet environment in order to determine whether to trigger the switch cabinet information collection and analysis steps. The unlocking and recognition adaptation module performs point cloud image analysis based on the collected switch and lock images to determine the keyhole area and trigger a mechanical unlocking command to unlock the switch cabinet; The information acquisition and analysis module determines the component detection type and elements of the switchgear based on the detection target through multi-source sensing detection, in order to generate a detection instruction sequence, collect detection information, and output the detection results.

[0098] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional modules is used as an example. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. The specific working process of the system, device, and unit described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0099] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional modules is used as an example. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. The specific working process of the system, device, and unit described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0100] The above are all preferred embodiments of this application and are not intended to limit the scope of protection of this application. Any feature disclosed in this specification (including the abstract and drawings) may be replaced by other equivalent or similar features unless specifically stated otherwise. That is, unless specifically stated otherwise, each feature is only one example of a series of equivalent or similar features.

Claims

1. An automatic inspection method for switchgear based on an inspection robot, applied to an automatic inspection robot, wherein the automatic inspection robot is equipped with an autonomous inspection module for controlling autonomous movement, an image acquisition module for acquiring images, an infrared temperature measurement module for detecting temperature, and a robotic arm module for unlocking the switchgear, characterized in that, include: The inspection path allocation step involves analyzing the preset path generation strategy and inspection target to allocate the inspection path corresponding to the inspection target and triggering the inspection command to instruct the inspection robot to match and execute the inspection operation. The unlocked environment analysis step is configured with a switch cabinet environment analysis strategy to detect and analyze the security of the switch cabinet environment in order to determine whether to trigger the switch cabinet information collection and analysis step. The unlocking and identification adaptation steps involve performing point cloud image analysis based on the collected switch and lock images to determine the keyhole area and trigger a mechanical unlocking command to unlock the switch cabinet. The information acquisition and analysis steps involve multi-source sensing detection based on the detection target to determine the component detection type and elements of the switchgear, in order to generate a detection instruction sequence, collect detection information, and output the detection results.

2. The automatic inspection method for switchgear based on an inspection robot according to claim 1, characterized in that, The path generation strategy includes: Based on the detection targets and the preset site inspection map, planning is carried out to generate different movement paths that pass through multiple detection targets. The path optimization model is used to traverse the path and eliminate overlapping paths to obtain the best movement path that passes through multiple detection targets. Based on the optimal movement path, the detection order of multiple detection targets is arranged to obtain the target sorting information when the switch cabinet is detected, and the path update strategy is triggered to update the optimal movement path in real time.

3. The automatic inspection method for switchgear based on an inspection robot according to claim 2, characterized in that, The path update strategy includes: The detection status of multiple switchgear cabinets is matched according to the detection order to determine the corresponding allowed detection status in the component detection database. The switch cabinet is marked based on the allowed detection status, and the operating status of the switching components in the switch cabinet is monitored. When the operating status of the component is inconsistent with the allowed detection status, it is marked to identify the abnormal status detection target. Targets are filtered out based on abnormal state detection, and the path generation strategy is retried to update the optimal movement path.

4. The automatic inspection method for switchgear based on an inspection robot according to claim 1, characterized in that, The information collection and analysis steps are also configured with a component abnormal change analysis strategy, including: Feature analysis is performed on the detection images from multi-source sensing to determine the three-dimensional contour coordinates of the switching elements in the switch cabinet. Based on the three-dimensional contour coordinates of the switching element, at least one set of coordinates is selected to construct the absolute coordinate information of the element. The absolute coordinate information of the element is used to reflect the current contour features of the element. An absolute position coordinate library is generated by associating and storing the absolute coordinate information of the switch components. When the switch cabinet is inspected for the second time, the absolute coordinate position information of the switch components is compared to determine the component outline deviation information and generate a switch position abnormality prompt.

5. The automatic inspection method for switchgear based on an inspection robot according to claim 1, characterized in that, The aforementioned switchgear environment analysis strategy includes: Collect environmental safety parameters inside the switchgear where the target is located, including ambient temperature and ambient voltage field strength; Based on a comparative analysis of preset allowable thresholds and environmental safety parameters, calculations are performed to determine the safety parameter difference when the environmental safety parameters exceed the preset allowable thresholds. Based on the comparison between the safety parameter difference and the preset effective measurement difference range, the information collection and analysis step is triggered when the safety parameter difference is less than the preset effective measurement difference range, and the switchgear safety prompt is triggered otherwise.

6. A method for automatic inspection of switchgear based on an inspection robot according to claim 1 or 5, characterized in that, It also includes environmental disturbance compensation strategies, including: Based on the detection target, the switchgear environment is analyzed to determine the environmental interference information, which includes electromagnetic interference information and electric field interference information. Interference compensation data is calculated based on a preset interference compensation model and environmental interference information. The interference compensation data includes robotic arm control signals and multi-source sensor detection data. Based on interference compensation data, the inspection robot monitors inspection signals, and corrects the robotic arm control signals and multi-source sensor detection data when the unlocking and identification steps are triggered.

7. The automatic inspection method for switchgear based on an inspection robot according to claim 6, characterized in that, The interference compensation model includes a signal filtering correction sub-model and an electromagnetic interference correction sub-model, which are calculated using the following formulas: The signal filtering correction sub-model is as follows: ； The electromagnetic interference correction sub-model is as follows: ； in, This is the filtered control signal for the robotic arm. The raw control signals for the robotic arm. To detect the angular frequency of the control signal, The preset filter resistor value, The preset filter capacitor value, For the compensated multi-source sensor detection data, These are actual measured data. The preset electric field compensation coefficient is used. Given the current electric field strength, This is the set calibration electric field strength reference value.

8. The automatic inspection method for switchgear based on an inspection robot according to claim 1, characterized in that, The information collection and analysis steps also include a component anomaly verification strategy, including: Matching is performed based on the component detection type to determine the corresponding operating state test current and operating state test voltage for the detection component type in the preset attachment verification database; The detection elements of the switchgear are tested and adjusted based on the operating state test current and operating state test voltage, and the test temperature of the detection elements is used to generate temperature change curves. The curvature is calculated based on the temperature change curve to determine the test temperature change rate. When the test temperature change rate exceeds the preset standard change rate, the detection element is marked as an abnormal element, and abnormal element feedback information is generated.

9. The automatic inspection method for switchgear based on an inspection robot according to claim 8, characterized in that, The component anomaly verification strategy also includes an anomaly level evaluation sub-strategy, including: Multiple anomaly levels are obtained through analysis of multi-source sensing detection by the detection elements, and a comprehensive anomaly level is calculated based on the anomaly level evaluation model, which uses the following formula: ； in, To determine the overall anomaly level, to These are preset weighting coefficients for circuit abnormalities, component aging, partial discharge abnormalities, and temperature abnormalities. The line connection is at an abnormal level. The component is considered to be at an abnormal aging level. It is classified as a partial discharge abnormality. This indicates an abnormal temperature level.

10. An automatic inspection system for switchgear based on an inspection robot, employing the automatic inspection method for switchgear based on an inspection robot as described in any one of claims 1-9, comprising: The inspection path allocation module analyzes the preset path generation strategy and inspection target to allocate the inspection path corresponding to the inspection target and triggers the inspection command to instruct the inspection robot to match and execute the inspection operation. The unlocked environment analysis module is configured with a switch cabinet environment analysis strategy to detect and analyze the safety of the switch cabinet environment in order to determine whether to trigger the switch cabinet information collection and analysis steps. The unlocking and recognition adaptation module performs point cloud image analysis based on the collected switch and lock images to determine the keyhole area and trigger a mechanical unlocking command to unlock the switch cabinet; The information acquisition and analysis module determines the component detection type and elements of the switchgear based on the detection target through multi-source sensing detection, in order to generate a detection instruction sequence, collect detection information, and output the detection results.

Images