A substation autonomous inspection control method coordinated with an intelligent inspection robot

Abstract

The application discloses a kind of intelligent inspection robot collaborative substation autonomous inspection control method, it is related to substation inspection control field, comprising: by extracting contact temperature rise rate and interphase temperature difference dispersion rate from infrared thermal image time series data, equipment state deterioration degree is distinguished;According to equipment state deterioration degree, corresponding compensatory inspection task is matched, including inspection density, perspective and spectral segment configuration;Compensatory inspection task, inspection priority and expected completion time window are sent to robot as tender, and robot is based on own path delay cost and the insertion cost of current task sequence negotiates and elects winning robot, and drives winning robot to plan the insertion track satisfying time window constraint;Through adjacent robot task priority negotiation, path intersection conflict is solved;Realize equipment deterioration trend driven differentiated adaptive inspection, improve the resource utilization efficiency of multi-robot collaborative operation.

Description

Technical Field

[0001] This invention belongs to the field of substation inspection and control technology, and relates to a substation autonomous inspection and control method in collaboration with intelligent inspection robots. Background Technology

[0002] Substations are critical nodes in the power system, undertaking the functions of voltage transformation and power distribution. To ensure the safe and stable operation of equipment within the substation, inspection robots are gradually replacing manual inspections. The independent operation of a single robot is evolving into the collaborative operation of multiple robot clusters. When multiple robots share the physical path network of the substation, they will face the problem of path intersection and collision conflicts.

[0003] Existing technologies include collaborative collision avoidance schemes for power system inspection robots. For example, the invention patent with publication number CN121165716A proposes a collaborative collision avoidance method and device for robot swarms in power system inspection. By broadcasting their own motion status and planned paths, robots that identify potential collision conflicts form a temporary decision-making group. Within the group, each conflicting robot calculates and negotiates priorities based on its overall task priority, the urgency of the conflict, and its historical maneuver frequency, and then assigns collaborative collision avoidance actions to alleviate congestion caused by fixed priorities.

[0004] However, the aforementioned existing technologies still have the following shortcomings in practical applications: First, the status of equipment in substations changes in real time during actual operation. When the contact resistance of equipment contacts increases due to oxidation or loosening, the local temperature will rise abnormally, the fault will develop rapidly, and the risk will increase sharply. Existing technologies replace static priority allocation with calculation and negotiation priority, but the task priority used in the calculation still depends on the set static parameters, which may cause the robot cluster to delay the inspection of truly urgent equipment when facing sudden high-risk equipment.

[0005] Furthermore, the aforementioned existing technologies mainly focus on the path conflict resolution mechanism between robots, while ignoring the urgency of the real-time status of the equipment. Simply resolving path conflicts without establishing a dynamic correlation between task priority and real-time equipment degradation not only makes it difficult to prioritize high-risk equipment with limited robot resources, but may also lead to a disconnect between emergency inspection needs and task scheduling, thereby losing the best opportunity for early maintenance and increasing the risk of equipment failure. Summary of the Invention

[0006] In view of this, in order to solve the problems mentioned in the background art, the present invention provides a substation autonomous inspection control method in collaboration with intelligent inspection robots.

[0007] The objective of this invention can be achieved through the following technical solution: A substation autonomous inspection control method in collaboration with an intelligent inspection robot, comprising: S1, extracting the contact temperature rise rate and phase-to-phase temperature difference divergence rate based on the infrared thermal imaging time series data of key equipment nodes in the substation, determining the degree to which the equipment status deviates from the safety boundary, and outputting the equipment status deterioration degree. S2. Generate inspection priorities based on equipment condition deterioration, and match corresponding compensatory inspection tasks based on the inspection priorities, including inspection density, viewing angle and spectral band configuration. S3. The compensatory inspection task, inspection priority and expected completion time window are sent to the robot as bids. The robot negotiates and elects the winning robot based on its own path delay cost to reach the target area and the insertion cost to the current task sequence. S4. Drive the winning robot to plan the insertion trajectory that meets the time window constraints based on the substation path map, and resolve path intersection conflicts through task priority negotiation between adjacent robots.

[0008] Compared with the prior art, the beneficial effects of the present invention are as follows: (1) The present invention extracts the contact temperature rise rate and the phase temperature difference divergence rate from infrared thermal imaging time series data to determine the equipment condition deterioration degree, and ensures that high-risk equipment has priority scheduling rights over low-risk equipment during the task allocation cycle, thereby reducing the delay risk caused by static priority.

[0009] (2) The present invention matches the corresponding compensatory inspection tasks according to the degree of equipment condition deterioration, including inspection density, viewing angle and spectrum configuration, to solve the problem that the existing technology is difficult to prioritize high-risk equipment under limited robot resources, realize refined data collection of equipment with severe deterioration trend, and improve the ability to identify early faults.

[0010] (3) In this invention, the compensatory inspection task, inspection priority and expected completion time window are sent to the robot as bids. The robot negotiates and selects the winning robot based on its own path delay cost and the insertion cost of the current task sequence, and drives the winning robot to plan the insertion trajectory that meets the time window constraint. The task priority and the real-time degradation trend of the equipment are mapped and associated in real time, so as to ensure that the emergency inspection needs and task scheduling are matched in real time.

[0011] (4) The present invention resolves path intersection conflicts through task priority negotiation between adjacent robots, and achieves conflict resolution without the need for central scheduling intervention. This not only ensures the priority of robots corresponding to high-risk equipment, but also avoids inspection delays caused by path conflicts, thereby reducing the risk of equipment failure and extending the safe operation cycle of substation equipment. Attached Figure Description

[0012] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0013] Figure 1 This is a diagram illustrating the implementation steps of the method of the present invention; Figure 2 This is a flowchart illustrating the process of determining the inspection task sequence for this invention. Figure 3 This is a flowchart illustrating the negotiation and handling process for path intersection conflicts of the inspection robot in this invention. Detailed Implementation

[0014] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0015] Please see Figure 1 As shown, the present invention provides a substation autonomous inspection control method in collaboration with intelligent inspection robots, including: S1, extracting the contact temperature rise rate and phase-to-phase temperature difference divergence rate based on the infrared thermal imaging time series data of key equipment nodes in the substation, determining the degree of equipment status deviation from the safety boundary, and outputting the equipment status deterioration degree.

[0016] Because the contacts of key equipment in substations can generate abnormal heat due to defects such as increased contact resistance, oxidation, or loosening during long-term operation, the temperature rise rate can detect the deterioration acceleration trend earlier. Therefore, extracting the contact temperature rise rate and the phase-to-phase temperature difference divergence rate can determine the dynamic degree of equipment deviation from the safety boundary.

[0017] Therefore, the specific steps for extracting the contact temperature rise rate and the phase-to-phase temperature difference divergence rate are as follows: S1-1, Collect infrared thermal images of key equipment nodes in continuous time sections, and segment the contact area and the conductor area connected to the contact from the infrared thermal images of each time section. S1-2. Calculate the rate of change of the temperature difference between the contact region and the conductor region between adjacent time sections, and denot it as the contact temperature rise rate. S1-3. Under the same time section, take the maximum value of the temperature difference between any two phase contact areas of the same equipment to obtain the interphase temperature difference; S1-4. Calculate the rate of change of the interphase temperature difference over time as the interphase temperature difference divergence rate.

[0018] The key equipment nodes mentioned in step S1-1 above refer to electrical connection points in the substation with a high probability of defects such as increased contact resistance, oxidation, or loosening. Specifically, they include, but are not limited to, the contact parts of the moving or stationary contacts of circuit breakers and disconnectors, as well as the overlapping surfaces between busbars and leads.

[0019] A continuous time profile refers to a snapshot of the temperature distribution that is equidistant over time, formed by collecting data at fixed sampling intervals, such as every 1 minute or 5 minutes. Because infrared thermal imagers can acquire two-dimensional temperature field distribution information of equipment surfaces in a non-contact manner in real time, without interfering with equipment operation and reflecting the overall heating status, infrared thermal images are obtained through continuous monitoring and data collection of equipment using online infrared thermal imagers.

[0020] The process of segmenting the contact region and the conductor region connected to the contact includes the following steps: First, the infrared thermal image acquired at the current time segment is spatially registered with the visible light image based on the outer contour of the equipment, and the region of interest where the key equipment node is located is located in the registered image; then, the temperature gradient amplitude along the horizontal and vertical directions is calculated pixel by pixel in the region of interest to generate a temperature gradient map, and the internal high-temperature area is separated into the contact region by using the edge of the steep gradient amplitude as the segmentation boundary; finally, starting from the edge of the contact region, the search is carried outward along the temperature gradient direction, and the strip-shaped region with continuously decreasing temperature and connected to the contact region is identified as the conductor region.

[0021] The conductor region refers to the area connected to the contacts and in the same electrical circuit where, under normal operating conditions, the temperature rise is mainly driven by heat conduction from the contacts.

[0022] Since the abnormal heating of the contacts originates from the local Joule heating caused by the increase in contact resistance, the heat will be conducted and diffused along the conductor to both ends. If only the contact area is divided, the heating of the contacts themselves and the heat conducted by the conductor will be superimposed, causing the calculation of the temperature rise rate to be mixed with the heat conduction lag factor, which affects the degradation judgment. Therefore, by dividing and distinguishing the contact area and the conductor area at the same time, the heat source and the heat conduction path are separated.

[0023] The specific steps of S1-2 are as follows: First, extract the temperature values ​​of all pixels in the contact area at each time segment and calculate the average to obtain the average temperature of the contact area. Simultaneously, obtain the average temperature of the conductor area in the same way. Then, subtract the average temperature of the conductor area from the average temperature of the contact area to obtain the temperature difference of the time segment. Second, obtain the temperature difference between two adjacent time segments. Subtract the temperature difference of the previous time segment from the temperature difference of the later time segment and divide by the sampling time interval between the adjacent time segments. The result is the contact temperature rise rate between the adjacent time segments. For three-phase equipment, calculate the contact temperature rise rate of each phase contact in the above manner, and select the maximum value from the three-phase calculation results as the contact temperature rise rate.

[0024] When the equipment is defect-free, the temperature difference between the contact and conductor regions approaches zero. However, as the contact resistance increases, Joule heating mainly concentrates in the contact region, leading to a widening temperature difference between the contact and conductor. Therefore, using the conductor region temperature as a reference standard can eliminate the interference of ambient temperature and load current fluctuations on absolute temperature measurement.

[0025] Since a defect in one of the three phases can cause a rapid temperature rise, taking the maximum value of the temperature difference between any two phases in steps S1-3 above can capture the extreme temperature difference between the most abnormal phase and the normal phase, thus avoiding diluting the abnormal signal by taking the average or minimum value.

[0026] The specific steps for calculating the rate of change of interphase temperature difference over time in step S1-4 are as follows: First, take the maximum value of the temperature difference between any two phase contact areas of the same equipment as the interphase temperature difference between two adjacent time sections. Subtract the interphase temperature difference of the previous time section from the interphase temperature difference of the later time section, and then divide the resulting difference by the sampling time interval between adjacent time sections. The calculation result is the interphase temperature difference divergence rate between adjacent time sections. A positive rate of change indicates divergence, while zero or negative indicates no divergence.

[0027] The above-mentioned determination of the degree to which the equipment status deviates from the safety boundary, and the output of the equipment status deterioration degree include: for the monitored equipment, taking the maximum value of the contact temperature rise rate of each phase as the contact temperature rise rate; Contact degradation usually starts from a single phase. An abnormal increase in contact resistance causes the heat generation power of that phase to deviate from the normal level first. By taking the maximum value of the temperature rise rate of each phase, the degradation dynamics of the phase that first experiences accelerated defect can be captured, avoiding the masking of local fault signals by averaging the low rate values ​​of the normal phase.

[0028] First, determine the direction of the contact temperature rise rate. If the contact temperature rise rate is in the positive direction, it is determined that the equipment is in a trend of temperature deterioration. If the contact temperature rise rate is zero or in the negative direction, it is determined that the equipment is in a trend of temperature stabilization or recovery. The method for determining the direction of the contact temperature rise rate is as follows: take the change in temperature difference between adjacent time sections. If the result of subtracting the temperature difference from the previous time section from the temperature difference at the next time point is greater than zero, then it is the positive direction; if the result is equal to zero or less than zero, then it is the non-positive direction.

[0029] A positive contact temperature rise rate indicates that the temperature difference between the contact area and the conductor area is continuously increasing, indicating that the Joule heat power generated by the defect point is greater than the heat dissipation power, the energy is in a net accumulation state, and the deterioration is accelerating; if it is zero or negative, the heat dissipation power is greater than or equal to the heat generation power, the equipment is in thermal equilibrium or recovery state, and the deterioration has not continued to worsen.

[0030] Based on the determination of a rising temperature deterioration trend, the interphase temperature difference divergence rate is judged. If the interphase temperature difference divergence rate is greater than zero, it is determined that the equipment condition has deteriorated due to imbalance. If the interphase temperature difference divergence rate is less than or equal to zero, it is determined that the condition has deteriorated due to equilibrium. Given that the overall contact is confirmed to be in a trend of increasing temperature and deterioration, if the temperature difference divergence rate between phases is greater than zero, it indicates that the temperature difference between the phases is increasing over time, indicating that the heat generation rate of one phase has diverged from the other two phases, the three-phase thermal balance has been broken and the imbalance continues to worsen, so it is judged as unbalanced deterioration; if it is less than or equal to zero, the temperature difference between the phases has not increased or even decreased, indicating that although the three phases are heating up synchronously, the thermal balance has not been broken, and the fault is developing in a balanced manner, so it is judged as balanced deterioration.

[0031] The device that simultaneously has a positive contact temperature rise rate and a greater than zero interphase temperature difference divergence rate will output the highest level of degradation. The simultaneous presence of a positive contact temperature rise rate and a phase-to-phase temperature difference divergence rate greater than zero indicates that the equipment is not only experiencing net energy accumulation and continuous deterioration overall, but this acceleration is also concentrated in a single phase. The three-phase thermal balance has been broken and the degree of imbalance is expanding. This is the most dangerous physical state, characterized by accelerated deterioration of the single phase and a widening gap with other phases, and is closest to the fault critical point. Therefore, the output is the highest level of deterioration.

[0032] For devices that only have a positive contact temperature rise rate but a phase-to-phase temperature difference divergence rate of less than or equal to zero, the output is medium-level degradation. A positive rate of temperature rise indicates that the equipment is in a state of net energy accumulation and accelerated deterioration. However, a negative divergence rate indicates that the three-phase temperature difference is converging. The risk of deterioration is lower than that of a positive divergence rate, but still higher than that of a zero divergence rate. Therefore, it is classified as intermediate.

[0033] Devices with a non-positive contact temperature rise rate output a low level of degradation.

[0034] A non-positive contact temperature rise rate indicates that the temperature difference has not increased or is decreasing, the heat generation power of the defect point is not greater than the heat dissipation power, energy no longer accumulates, the equipment is in a stable thermal equilibrium or self-recovery state, the deterioration process has not started or has stopped, so it is judged as a low degree of deterioration.

[0035] S2. Generate inspection priorities based on equipment condition deterioration, and match corresponding compensatory inspection tasks based on the inspection priorities, including inspection density, viewing angle, and spectral band configuration.

[0036] Given that different levels of equipment degradation correspond to varying rates of failure development and severity, performing inspections on all equipment at the same frequency and in the same manner would lead to untimely inspections of high-risk equipment and excessive inspections of low-risk equipment, resulting in resource waste. Therefore, the aforementioned method of generating inspection priorities based on equipment degradation includes: pre-dividing the monitored equipment within the substation into several zones according to the electrical bays of the substation. The specific implementation steps for pre-dividing the monitored equipment in the substation into several areas based on the electrical bays of the substation are as follows: First, obtain the primary electrical main wiring diagram of the substation and identify the division boundaries of each electrical bay in the diagram. The electrical bays are independent functional units that divide the primary equipment in the substation, with circuit breakers and disconnectors as the dividing lines.

[0037] Secondly, the location of each monitored device in the substation's primary electrical main wiring diagram is obtained, the electrical bay where each monitored device is located is determined, and the device identifier of each monitored device is recorded under the corresponding electrical bay number, thus completing the classification of devices and bays.

[0038] Then, based on the three-dimensional spatial model or plan layout of the substation, the spatial coordinate range of each electrical bay within the station is obtained, and the physical boundary of the bay is used as the area boundary.

[0039] Finally, the spatial range corresponding to each electrical bay is treated as an independent inspection area, forming several inspection areas based on electrical bays.

[0040] Obtain the equipment condition deterioration degree of all devices under inspection in each region, and determine the highest equipment condition deterioration degree among them; Using regions as the granularity, regions containing the highest level of equipment degradation are prioritized as the current inspection regions, followed by regions containing medium-level equipment degradation, and finally regions containing only low-level equipment degradation. By using electrical intervals as the regional granularity and sorting them according to the highest degree of degradation within each region, spatial clustering of inspection resources is achieved: high-risk hotspot areas are prioritized for handling, avoiding ineffective cross-regional movement of robots and frequent switching of spectrum segments. At the same time, high-risk equipment is forced to be scheduled first under the time window constraints, thereby maximizing the overall risk suppression efficiency with limited resources.

[0041] Within the selected area, the shortest path sequence is obtained based on the equipment location information, and the inspection task sequence is determined. See Figure 2 As shown, the specific implementation steps for determining the inspection task sequence are as follows: obtain the current position of the robot performing this inspection task as the starting point of the path, and the positions of all devices to be inspected in the corresponding area, and map them as nodes in the path search; Using the starting point of the path and the location of each device to be inspected as nodes, a feasible moving path connecting the starting point and all nodes of the devices to be inspected is constructed based on the substation path map. The above steps for constructing a feasible moving path connecting the starting point and all nodes of the equipment to be inspected are as follows: retrieve the two-dimensional path topology map of the substation. The path topology map represents the area that the robot can pass through with nodes and edges. The nodes include road intersections, channel endpoints and equipment inspection docking points. The edges represent the barrier-free driving path connecting two nodes, and the edge weight is the path length.

[0042] Map the starting point of the path and the location of each device to be inspected to the nearest node in the path topology graph to obtain the node corresponding to the starting point and the node corresponding to each device; if the device location itself is a node, then take that node directly.

[0043] Calculate the shortest feasible path and its cumulative path length between the following two types of node pairs in the path topology graph: the shortest path between the starting node and each device node, and the shortest path between any two device nodes.

[0044] Using the starting point node and all device nodes as vertices, and the shortest path between each pair of vertices as connecting edges, a complete graph is constructed. Each edge in the complete graph is associated with the corresponding sequence of actual driving trajectory points and path length, forming a set of feasible movement paths connecting the starting point and all device nodes to be inspected.

[0045] For multiple candidate paths formed by accessing the device nodes under inspection in different orders, the total moving distance required for each candidate path to traverse all the device nodes under inspection starting from the starting point is determined. The shortest feasible path is selected from multiple candidate paths based on the shortest total travel distance. The order in which the equipment is passed along the shortest feasible path is taken as the inspection task sequence.

[0046] Choosing the path with the shortest total travel distance minimizes the total time it takes for the robot to traverse all the devices to be inspected in the current area, while satisfying the area inspection priority. This shortens the duration of the current task sequence, allowing the robot to release and respond to compensatory inspection tasks in other areas as soon as possible. It also reduces the risk of defaulting on the expected completion time window due to path redundancy and improves the overall timeliness of multi-robot collaborative scheduling.

[0047] After completing the inspection of the current area, select the next area according to the area priority order.

[0048] Considering that the infrared signatures of minor equipment anomalies are extremely weak and transient, the detection of early degradation relies heavily on sophisticated methods. If only a one-size-fits-all inspection approach is adopted, it is highly likely that potential hazards will only be discovered after they have developed into accidents. Therefore, a differentiated compensation mechanism is established. This mechanism, which matches corresponding compensatory inspection tasks based on inspection priority, includes: determining the insertion frequency of supplementary compensatory inspections between regular inspections based on the highest equipment degradation level in the current inspection area. The frequency of the highest degradation level is higher than that of the medium degradation level, and the frequency of the medium degradation level is higher than that of the lowest degradation level. Specifically, when the highest equipment condition deterioration level in the inspection area is the highest level, a compensatory inspection is inserted between two consecutive regular inspections; when the highest equipment condition deterioration level is medium level, a compensatory inspection is inserted every two regular inspections; when the highest equipment condition deterioration level is low level, no compensatory inspection is inserted, and only the regular inspection cycle is performed.

[0049] Furthermore, the insertion time of compensatory inspections is located at the midpoint between two adjacent regular inspections, so that compensatory inspections and regular inspections are equally distributed on the time axis, ensuring that the monitoring density of high-risk equipment is uniformly covered within the cycle.

[0050] When there are multiple monitored devices with different degradation levels in the same inspection area, the frequency of compensatory inspection insertion is determined based on the highest degradation level of the device in the inspection area, and the inspection items of each device are executed sequentially according to the shortest path.

[0051] When the equipment condition deterioration level of any monitored device changes, the frequency of compensatory inspection insertion for the area where the device is located is re-determined at the beginning of the next inspection cycle, and the inspection is performed according to the updated frequency.

[0052] If the selected area has the highest level of degradation, then the frontal, top-down, and side-tilt viewing angles are selected, and narrow-band and wide-band temperature measurement spectrums are used alternately. If it is of medium level, then the frontal and top-down viewing angles are selected, and wide-band temperature measurement spectrum is used. If it is of low level, then the frontal viewing angle is selected, and narrow-band temperature measurement spectrum is used. This ensures that the most urgent equipment receives the most sufficient data support, while avoiding system waste caused by excessive inspections.

[0053] The highest level of degradation equipment simultaneously exhibits a positive temperature rise rate and a positive interphase temperature difference divergence rate, indicating the potential for localized overheating and three-phase imbalance. A single frontal viewing angle is easily obstructed and cannot capture the true hot spots. It is necessary to add top-down and side-tilt viewing angles to eliminate obstruction and obtain the full picture of the temperature field. Narrow-band spectral bands have high thermal sensitivity and can accurately measure minute temperature differences, while wide-band spectral bands have high heat flux and can quickly locate hot spots. Alternating between the two can balance positioning accuracy and measurement efficiency.

[0054] The medium-deterioration equipment only shows a positive temperature rise rate but the phase temperature difference divergence rate is zero, which is a uniform temperature rise. The front and top viewing angles are sufficient to eliminate obstruction interference. The wide band spectrum can meet the conventional temperature measurement accuracy, and there is no need to increase the processing overhead by using a narrow band.

[0055] When the equipment temperature stabilizes or recovers at a low level of degradation, the risk is minimal. Basic monitoring can be completed with a standard configuration that uses a frontal view and narrow spectral bands, thus avoiding waste of resources.

[0056] S3. The compensatory inspection task, inspection priority, and expected completion time window are sent to the robot as bids. The robot negotiates and selects the winning robot based on its own path delay cost to reach the target area and the insertion cost to the current task sequence.

[0057] The path delay cost for the robot to reach the target area and the insertion cost into the current task sequence include: dividing the shortest path length from the robot's current position to the target area by the robot's current average speed to obtain the path delay cost; The path length divided by the average speed gives the shortest time required for the robot to physically move from its current position to the target area. This reflects the delay cost of the robot responding to this inspection task, so as to compare the relative cost of each robot reaching the site in multi-robot negotiation.

[0058] Calculate the shortest path length from the position of the last task in the current task sequence to the target area, and divide it by the average speed to obtain the additional travel time; Extract the original planned completion time of the last scheduled task in the robot's current task sequence, add the additional travel time and the expected operation time of the compensating inspection task to the original planned completion time to obtain the new completion time, and use the difference between the new completion time and the original planned completion time as the insertion cost.

[0059] Insertion cost refers to the length of time that the completion time of the original last task is delayed due to the insertion of a new task.

[0060] Since robots already have a predetermined task sequence, each task has a planned completion time. Inserting a new task means adding extra travel distance and operation time to the existing task sequence, inevitably delaying the scheduling of subsequent tasks. The difference between the original planned completion time and the new planned completion time quantifies the degree to which the insertion of a new task crowds out and disrupts the existing task plan. Using this difference as the insertion cost, the planning adjustment cost that each robot needs to bear to undertake a new task can be measured from a time perspective. Together with the path delay cost, this constitutes the comprehensive cost of the robot responding to a new task.

[0061] Since each robot has different existing task load and available time, their response capabilities and execution costs for new tasks vary. Therefore, a negotiation and election mechanism is used to allow each robot to compete for tasks based on its real-time status, and the most suitable robot is selected for inspection.

[0062] Therefore, the above-mentioned negotiation to select the winning robot includes: when each robot receives the bidding information, if its own promised arrival time is within the expected completion time window, it will send back the sum of the path delay cost and the insertion cost along with its own promised arrival time as bidding information based on its current location and the sequence of tasks it has undertaken; otherwise, it will abandon the bid. This ensures the real-time performance and optimal resource allocation under the collaborative scheduling of multiple robots, and avoids blind bidding that could lead to large-scale timeouts or path congestion in subsequent tasks.

[0063] Compensatory inspection tasks target equipment that has already deteriorated. The deterioration process is time-sensitive; if the robot arrives too late, the equipment may have deteriorated further or even malfunctioned, rendering the inspection task meaningless. Therefore, using the expected completion time window as a constraint prevents the optimal inspection opportunity from being missed due to task assignment delays.

[0064] The above-mentioned process for calculating the promised arrival time and determining the time window is as follows: The robot obtains its current position and the execution time and in-station movement time of each task in the already accepted task sequence. Starting from the current moment, it executes the task sequence sequentially until the last task. Then, it calculates the travel time required to move from the last task position to the target area. The current moment is added to the sum of all the durations to obtain its promised arrival time.

[0065] The task trigger node acquires the current contact temperature rise rate and interphase temperature difference divergence rate of the monitored equipment. Based on the current values ​​of these two parameters and their recent trends, it extrapolates the estimated remaining time before the equipment reaches the safety critical point. The current moment is used as the start time of the time window, and the current moment plus the estimated remaining time multiplied by a safety factor less than 1 is used as the end time of the time window, forming the expected completion time window. The steps for obtaining the estimated remaining time of the safety critical point are as follows: the highest softening temperature of the contact material is used as the safety critical value; the remaining temperature rise space obtained by subtracting the current contact temperature is divided by the sum of the current contact temperature rise rate and the interphase temperature difference divergence rate to obtain the estimated remaining time. The safety factor ranges from 0.3 to 0.7, with a default value of 0.5. For critical equipment, such as the main transformer incoming line bay, the safety factor is 0.3; for general equipment, it is 0.7. The smaller the safety factor, the tighter the time window, and the more urgent the inspection response.

[0066] Each robot calculates its path delay cost and insertion cost, then sums them to obtain the total cost, while recording its own promised arrival time. If the promised arrival time falls between the earliest and latest times within the expected completion time window, the robot sends the total cost, promised arrival time, and its own identifier back to the task triggering node as bidding information; otherwise, it abandons the bid and does not send any information.

[0067] The task trigger node collects all bidding information and selects the robot with the smallest cumulative result as a candidate. If multiple robots have the same cumulative result, the robot with the highest remaining power is selected as the winning robot.

[0068] The goal of selecting the robot with the smallest cumulative result as the winner is to achieve the optimal allocation of inspection tasks by minimizing path disturbance and task delay costs, thereby reducing the overall time and energy costs in the multi-robot collaborative process.

[0069] When the results are the same, the robot with the highest remaining power is selected. The purpose is to balance the energy load of each robot and avoid the robot losing its working ability due to premature power depletion caused by repeatedly assigning the same robot. This extends the overall sustainable working time of the robot team and improves the task completion rate under long-term unattended conditions.

[0070] It should be added that if no bidding information is received within the waiting time threshold, the task triggering node will execute the following strategy: increase the priority of the current inspection task to be higher than the lowest priority task in the current task sequence of each robot, and force the robot to recalculate the insertion cost and bid. The waiting time threshold is directly set to a fixed 1 second or 2 seconds, and then fine-tuned between 10% and 20% based on the frequency of triggering priority increases during actual operation.

[0071] S4. Drive the winning robot to plan the insertion trajectory that meets the time window constraints based on the substation path map, and resolve path intersection conflicts through task priority negotiation between adjacent robots.

[0072] The above-mentioned driving robot to plan the insertion trajectory that meets the time window constraint based on the substation path map includes: receiving the expected completion time window issued by the task triggering node, and searching for an idle period on the generated and locked path time line, the continuous length of the intersection interval with the expected completion time window is not less than the shortest operation time. The idle period within the expected completion time window refers to a time interval between two adjacent existing tasks on the robot's locked path timeline, during which the robot has no operations. It is obtained by extracting the time gaps between all adjacent tasks from the robot's currently locked task sequence, determining whether each gap intersects with the expected completion time window, and if so, calculating the intersection interval and checking if its length is not less than the minimum operation time required to perform the compensatory inspection task. If satisfied, the intersection interval is considered a candidate idle period.

[0073] It should be added that if no idle period with a length not less than the shortest job time is found within the expected completion time window, the robot returns a scheduling failure flag to the task triggering node. After receiving the scheduling failure flag, the task triggering node either reduces the expected job time of the current compensation task or re-executes the negotiation and election process to select other candidate robots.

[0074] Based on the static path map of the substation, a travel path is planned from the current position corresponding to the start of the idle period to the target area, and after the task is completed, it is connected to the next task point to generate an updated trajectory plan.

[0075] Because multiple robots operate simultaneously on a shared substation path network, there is an inevitable potential conflict due to path intersections or overlaps, where they may simultaneously occupy the same road segment or intersection. Therefore, see [link to relevant documentation]. Figure 3 As shown, the above-mentioned negotiation to resolve path intersection conflicts includes: when any robot, based on its updated trajectory, predicts that it will occupy the same road segment or intersection with another robot at a certain time in the future, it initiates a conflict negotiation request to the other party. The specific implementation steps for predicting that a robot will occupy the same road segment or intersection with another robot at a certain future time are as follows: First, each robot maintains its own locked and effective driving trajectory in real time. The trajectory is represented in the form of a time and position sequence, where the position is subdivided into path segment identifiers or intersection identifiers.

[0076] Secondly, each robot periodically broadcasts its future trajectory information, including the expected start and end times of occupying each road segment and intersection, through vehicle-to-vehicle communication or via a central dispatch node.

[0077] Then, once any robot generates a new inserted trajectory based on the updated trajectory, it compares the new trajectory with the broadcast trajectories of other robots that it has already received.

[0078] Next, check segment by segment on the timeline whether there is any overlap between the time intervals occupied by the same path segment or the same intersection point by the local machine and other robots. If there is an overlap, it is determined that simultaneous occupation will occur at some point in the future.

[0079] Finally, the specific time and location of the overlap, as well as the identifier of the other robot involved, are extracted to trigger a conflict negotiation request.

[0080] During negotiations, the two sides in the conflict compare the inspection priorities of the tasks they will be performing when entering the conflict section. The side with lower priority gives way, while the side with higher priority passes through the original route first. Inspection priority reflects the urgency of equipment deviating from safety boundaries. High-priority tasks correspond to equipment with more severe degradation trends and tighter time windows; any delay could lead to missed faults or accidents. Low-priority tasks correspond to equipment with lower risks and have sufficient time to wait. Therefore, in the event of path conflicts, prioritizing the original path of high-priority tasks and actively avoiding low-priority tasks ensures that the most valuable right-of-way is allocated to the most urgent inspection needs.

[0081] If both parties have the same inspection priority, the one that arrives at the conflict point first shall pass first, and the one that arrives later shall give way. If the one that arrives later cannot give way, the dispatch node shall be notified to re-plan. When both robots have the same level of urgency, there is no basis for prioritization. The robot that arrives at the conflict point first has already invested time and energy. If it is required to give way, it will need to slow down, stop, or even reverse, resulting in unnecessary energy consumption and wasted time. The robot that arrives later only needs to wait a short while to pass. This avoids deadlocks caused by robots giving way or locking each other due to equal priority, and reduces unnecessary acceleration, deceleration, and stopping operations.

[0082] If the arrival times are the same, the robot with the smaller number will pass first, and the other will give way.

[0083] Conflict negotiation employs a recursion depth limit, with a default maximum of 3 negotiation attempts. If the conflict cannot be resolved after exceeding this limit, it is determined to be due to path network congestion. The high-priority robot currently inserting the task requests a temporary lock on the conflict segment from the scheduling node, while the low-priority robot waits in place until the high-priority robot passes through.

[0084] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product.

[0085] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0086] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0087] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0088] Finally, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for autonomous substation inspection control using intelligent inspection robots in collaboration, characterized in that: include: Based on the infrared thermal imaging time series data of key equipment nodes in the substation, the contact temperature rise rate and phase-to-phase temperature difference divergence rate are extracted to determine the degree of equipment status deviation from the safety boundary and output the equipment status deterioration degree. Inspection priorities are generated based on the degree of equipment condition deterioration, and corresponding compensatory inspection tasks are matched according to the inspection priorities, including inspection density, viewing angle and spectral band configuration. The compensatory inspection task, inspection priority, and expected completion time window are sent to the robot as bids. The robot negotiates and elects the winning robot based on its own path delay cost to reach the target area and the insertion cost to the current task sequence. The robot that won the bid plans an insertion trajectory that meets the time window constraints based on the substation path map, and resolves path intersection conflicts through task priority negotiation between adjacent robots. The degradation of the output device status includes: For the monitored equipment, the maximum value of the contact temperature rise rate for each phase is taken as the contact temperature rise rate; First, determine the direction of the contact temperature rise rate. If the contact temperature rise rate is in the positive direction, it is determined that the equipment is in a trend of temperature deterioration. If the contact temperature rise rate is zero or in the negative direction, it is determined that the equipment is in a trend of temperature stabilization or recovery. Based on the determination of a rising temperature deterioration trend, the interphase temperature difference divergence rate is judged. If the interphase temperature difference divergence rate is greater than zero, it is determined that the equipment condition has deteriorated due to imbalance. If the interphase temperature difference divergence rate is less than or equal to zero, it is determined that the condition has deteriorated due to equilibrium. The device that simultaneously has a positive contact temperature rise rate and a greater than zero interphase temperature difference divergence rate will output the highest level of degradation. For devices that only have a positive contact temperature rise rate but a phase-to-phase temperature difference divergence rate of less than or equal to zero, the output is medium-level degradation. Devices with a non-positive contact temperature rise rate output a low level of degradation.

2. The method for autonomous substation inspection control using intelligent inspection robots in collaboration with the method described in claim 1, characterized in that: The extraction contact temperature rise rate and interphase temperature difference divergence rate include: Collect infrared thermal images of key equipment nodes in continuous time sections, and segment the contact area and the conductor area connected to the contact from the infrared thermal images of each time section. The rate of change of the temperature difference between the contact region and the conductor region between adjacent time sections is calculated and denoted as the contact temperature rise rate. Under the same time section, the temperature difference between any two phase contact areas of the same equipment is taken as the maximum value to obtain the interphase temperature difference; The rate of change of the interphase temperature difference over time is calculated as the interphase temperature difference divergence rate.

3. The method for autonomous substation inspection control using intelligent inspection robots in collaboration with the method described in claim 1, characterized in that: The process of generating inspection priorities based on equipment condition deterioration includes: The monitored equipment within the substation is pre-divided into several areas based on the electrical bays of the substation; Obtain the equipment condition deterioration degree of all devices under inspection in each region, and determine the highest equipment condition deterioration degree among them; Using regions as the granularity, regions containing the highest level of equipment degradation are prioritized as the current inspection regions, followed by regions containing medium-level equipment degradation, and finally regions containing only low-level equipment degradation. Within the selected area, the shortest path sequence is obtained based on the equipment location information, and the inspection task sequence is determined. After completing the inspection of the current area, select the next area according to the area priority order.

4. The method for autonomous substation inspection control using intelligent inspection robots in collaboration with the method described in claim 3, characterized in that: The determined inspection task sequence includes: Obtain the current position of the robot performing this inspection task as the starting point of the path, and the positions of all devices to be inspected in the corresponding area, and map them as nodes in the path search; Using the starting point of the path and the location of each device to be inspected as nodes, a feasible moving path connecting the starting point and all nodes of the devices to be inspected is constructed based on the substation path map. For multiple candidate paths formed by accessing the device nodes under inspection in different orders, the total moving distance required for each candidate path to traverse all the device nodes under inspection starting from the starting point is determined. The shortest feasible path is selected from multiple candidate paths based on the shortest total travel distance. The order in which the equipment is passed along the shortest feasible path is taken as the inspection task sequence.

5. The method for autonomous substation inspection control using intelligent inspection robots in collaboration with the method described in claim 1, characterized in that: The matching of corresponding compensatory inspection tasks based on inspection priority includes: The frequency of supplementary inspections between regular inspections is determined based on the highest equipment condition deterioration level in the current inspection area. The frequency of the highest deterioration level is higher than that of the medium deterioration level, and the frequency of the medium deterioration level is higher than that of the low deterioration level. If the selected area is at the highest level of degradation, then the frontal, top-down, and side-tilt viewing angles are selected, and narrow-band and wide-band temperature measurement spectrums are used alternately. If it is at the medium level, then the frontal and top-down viewing angles are selected, and wide-band temperature measurement spectrums are used. If it is at the low level, then the frontal viewing angle is selected, and narrow-band temperature measurement spectrums are used.

6. The method for autonomous substation inspection control using intelligent inspection robots in collaboration with the method described in claim 1, characterized in that: The path delay cost for the robot to reach the target area and the insertion cost into the current task sequence include: Divide the shortest path length from the robot’s current position to the target area by the robot’s current average speed to obtain the path delay cost. Calculate the shortest path length from the position of the last task in the current task sequence to the target area, and divide it by the average speed to obtain the additional travel time; Extract the original planned completion time of the last scheduled task in the robot's current task sequence, add the additional travel time and the expected operation time of the compensating inspection task to the original planned completion time to obtain the new completion time, and use the difference between the new completion time and the original planned completion time as the insertion cost.

7. The method for autonomous substation inspection control using intelligent inspection robots in collaboration with the method described in claim 1, characterized in that: The robots selected through negotiation include: When each robot receives the bidding information, if its promised arrival time is within the expected completion time window, it will send back the sum of the path delay cost and the insertion cost along with its promised arrival time as the bidding information based on its current location and the task sequence it has already undertaken; otherwise, it will abandon the bidding. The task trigger node collects all bidding information and selects the robot with the smallest cumulative result as a candidate. If multiple robots have the same cumulative result, the robot with the highest remaining power is selected as the winning robot.

8. The method for autonomous substation inspection control using intelligent inspection robots in collaboration with the method described in claim 1, characterized in that: The insertion trajectory planned by the robot that won the bid based on the substation path map and satisfying the time window constraints includes: Receive the expected completion time window issued by the task triggering node, and search for an idle period on the generated and locked path timeline. The continuous length of the intersection interval with the expected completion time window is not less than the shortest operation time. Based on the static path map of the substation, a travel path is planned from the current position corresponding to the start of the idle period to the target area, and after the task is completed, it is connected to the next task point to generate an updated trajectory plan.

9. The method for autonomous substation inspection control using intelligent inspection robots in collaboration with the method described in claim 1, characterized in that: The conflict resolution paths mentioned include: When any robot, based on its updated trajectory, predicts that it will occupy the same road segment or intersection as another robot at some point in the future, it initiates a conflict negotiation request to the other robot. During negotiations, the two sides in the conflict compare the inspection priorities of the tasks they will be performing when entering the conflict section. The side with lower priority gives way, while the side with higher priority passes through the original route first. If both parties have the same inspection priority, the one that arrives at the conflict point first shall pass first, and the one that arrives later shall give way. If the one that arrives later cannot give way, the dispatch node shall be notified to re-plan. If the arrival times are the same, the robot with the smaller number will pass first, and the other will give way.

Images