Operation and maintenance scheme generation method and device for electric power well lid and storage medium

Abstract

The application discloses an operation and maintenance scheme generation method and device of a power well lid and a storage medium, and relates to the technical field of intelligent operation and maintenance of power equipment. The method constructs input features through multi-dimensional state data of well lids in each jurisdiction; then, according to the geographic coordinates of the well lids and the topological structure of the cable channel, a regional well lid relationship graph is constructed and a corresponding adjacency matrix is determined; then, the input features and the adjacency matrix are input into a graph time sequence neural network model to predict the health degree change trend and the failure probability of each well lid, and the emergency degree score of each well lid is calculated accordingly. Then, a risk propagation model is constructed in combination with the adjacency matrix to update the collaborative emergency degree of the well lid, and based on the collaborative emergency degree, the geographic road network and the operation and maintenance resource constraints, an optimal inspection route is dynamically planned and generated, a spare parts list is generated in combination with the simulation maintenance result, the spare parts preset position and quantity are determined through cluster analysis, and finally, the precise spare parts scheduling strategy is generated by adapting to the operation and maintenance scene, thereby improving the reliability of the operation and maintenance scheme.

Description

Technical Field

[0001] This application relates to the field of intelligent operation and maintenance technology for power equipment, and in particular to a method, equipment, and storage medium for generating operation and maintenance solutions for power manhole covers. Background Technology

[0002] Current methods for generating manhole cover maintenance plans only collect and analyze individual manhole operation monitoring data, relying on threshold judgments or individual prediction models to complete fault assessment and maintenance scheduling. However, power manhole covers are typically laid out in a continuous spatial pattern along cable lines, and manhole cover faults are prone to cascading transmission due to pipeline connections. Therefore, the data processing logic of the above-mentioned methods fragments the regional group relationships between manhole covers, making it difficult to adapt to the operating conditions of continuous fault transmission, ultimately resulting in low reliability of the generated maintenance plans.

[0003] The above content is only used to help understand the technical solution of this application and does not represent an admission that the above content is prior art. Summary of the Invention

[0004] The main purpose of this application is to provide a method, equipment and storage medium for generating operation and maintenance plans for power manhole covers, aiming to solve the technical problem of how to improve the reliability of manhole cover operation and maintenance plans.

[0005] To achieve the above objectives, this application proposes a method for generating an operation and maintenance plan for power manhole covers, the method comprising: The system acquires multidimensional status data of manhole covers in each jurisdiction collected by the regional gateway, and constructs input features based on the multidimensional status data. Based on the geographical coordinates of the manhole covers and the topology of the cable channels, a regional manhole cover relationship map is constructed, and an adjacency matrix is ​​determined based on the regional manhole cover relationship map. The input features and the adjacency matrix are input into a graph temporal neural network model, and the health trend and failure probability of each manhole cover are determined by the graph temporal neural network model. The urgency score of each manhole cover is determined based on the health status change trend and the failure probability; Based on the weighted connection weights in the adjacency matrix, a risk propagation model is constructed, and the collaborative urgency of each manhole cover is determined according to the urgency score and the risk propagation model. Using the collaborative urgency, the adjacency matrix, the driving distance between the manhole covers, and the maintenance resource constraints as input parameters for the preset path planning algorithm, the first round of inspection path planning is carried out in the geographical road network to generate an initial inspection sequence containing at least one manhole cover to be visited. The maintenance operation of the manhole covers in the initial inspection sequence is simulated sequentially. Based on the scores of each dimension in the health score of the manhole cover under maintenance, the spare parts demand type of the manhole cover under maintenance is determined, and the collaborative urgency of the remaining manhole covers in the geographical road network is updated according to the adjacency matrix. When the updated collaborative urgency meets the replanning condition, based on the updated collaborative urgency, the current maintenance location and the remaining manhole covers, the inspection path is replanned to generate the subsequent inspection sequence, and the maintenance operation of the manhole covers in the subsequent inspection sequence is simulated sequentially until the maintenance resource constraint is reached, and the inspection route is output. Based on the spare parts requirement types of all manhole covers performing simulated maintenance operations, a spare parts list is generated, and based on the spatial distribution of the collaborative urgency of each manhole cover and the adjacency matrix, the preset location and quantity of spare parts are determined by clustering. A spare parts scheduling strategy is generated based on the spare parts list, the preset locations of the spare parts, and the quantity of the spare parts.

[0006] In one embodiment, the step of constructing a regional manhole cover relationship map based on the geographic coordinates of the manhole cover and the cable channel topology includes: Each of the manhole covers is designated as a node in the manhole cover relationship diagram of the region; The distance between each manhole cover is determined based on its geographical coordinates. When the distance is less than a preset distance threshold, a topological adjacent edge is constructed between the nodes corresponding to the manhole covers. Based on the geographic coordinates of the manhole cover, determine the regional environment type to which the manhole cover belongs, and construct environmental similarity edges between nodes corresponding to manhole covers with the same regional environment type; Based on the cable channel topology, functional association edges are constructed between nodes corresponding to manhole covers that have channel topology dependencies.

[0007] In one embodiment, the step of determining the adjacency matrix based on the regional manhole cover relationship map includes: The initial connection weight of the edges is determined based on the type of edges between nodes in the regional manhole cover relationship diagram. Based on the time sequence data of the status of manhole covers corresponding to connected nodes within a preset time window, the dynamic correlation coefficient between the manhole covers corresponding to the connected nodes is determined. Based on the dynamic correlation coefficient, the initial connection weights of the edges between the connected nodes are adjusted to obtain the weighted connection weights; A weight matrix is ​​constructed based on the weighted connection weights, and the weight matrix is ​​normalized based on a preset weight threshold to generate the adjacency matrix.

[0008] In one embodiment, the step of determining the dynamic correlation coefficient between the manhole covers corresponding to the connected nodes based on the time-series data of the states of the manhole covers corresponding to the connected nodes within a preset time window includes: Based on the temporal characteristics of the state time series data, multiple event segments corresponding to the manhole cover are extracted from the state time series data; Based on the state time-series data corresponding to each event segment, an event feature sequence is constructed; Based on the feature sequence, the morphological similarity between event segments of the manhole covers corresponding to the connected nodes is determined within the preset time window; Based on the start time of the event segment, the event delay relationship between the manhole covers corresponding to the connected nodes is determined; Based on the morphological similarity and the event delay relationship, the dynamic correlation coefficient between the manhole covers corresponding to the connected nodes is determined.

[0009] In one embodiment, the weighted connection weight is used to quantify the state association strength between manhole covers; the collaborative urgency of the manhole cover is the weighted sum of its own urgency score and the urgency scores of other associated manhole covers.

[0010] In one embodiment, the step of determining the pre-positioned location and quantity of spare parts through clustering based on the spatial distribution of the collaborative urgency of each manhole cover and the adjacency matrix includes: A weighted graph is constructed using the collaborative urgency of each manhole cover as the node weight and the adjacency matrix as the connection relationship between nodes. Based on the geographical location of the pre-installed parts inventory, spatial clustering is performed on the weighted map to divide each manhole cover node into at least one high-risk area; Determine the geographical center of the high-risk area as the pre-positioned location of the spare parts corresponding to the high-risk area; Based on the total collaborative urgency value of manhole covers in the high-risk area corresponding to each of the aforementioned spare parts preset locations, the number of manhole covers, and the historical replacement frequency of similar spare parts, the quantity of spare parts corresponding to each of the aforementioned spare parts preset locations is determined.

[0011] In one embodiment, the replanning conditions include one or more of the following: The ratio of the decrease in the collaborative urgency of any of the remaining manhole covers to the decrease in the collaborative urgency of other remaining manhole covers is greater than a preset proportional threshold. The collaborative urgency level of any of the remaining manhole covers increases; The increase in the collaborative urgency level of any of the remaining manhole covers exceeds a preset threshold.

[0012] In addition, to achieve the above objectives, this application also proposes a device for generating operation and maintenance plans for power manhole covers. The device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor. The computer program is configured to implement the steps of the operation and maintenance plan generation method for power manhole covers as described above.

[0013] In addition, to achieve the above objectives, this application also proposes a storage medium, which is a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the steps of the above-described method for generating an operation and maintenance plan for power manhole covers.

[0014] This application provides a method for generating an operation and maintenance plan for power manhole covers. It acquires multi-dimensional status data of manhole covers in each jurisdiction collected by a regional gateway, and constructs input features based on this data. Next, based on the geographical coordinates of the manhole covers and the cable channel topology, a regional manhole cover relationship graph is constructed, and the corresponding adjacency matrix is ​​determined. Then, the input features and adjacency matrix are input into a time-series neural network model to predict the health trend and failure probability of each manhole cover, thereby calculating the urgency score of each manhole cover. Then, a risk propagation model is constructed using the adjacency matrix to update the collaborative urgency of the manhole covers. Based on the collaborative urgency, geographical road network, and operation and maintenance resource constraints, an optimal inspection route is dynamically planned and generated. Simultaneously, a spare parts list is generated based on simulated maintenance results. Cluster analysis is used to determine the pre-positioned location and quantity of spare parts. Finally, a precise spare parts scheduling strategy is generated to adapt to the operation and maintenance scenario, thus balancing the configuration of spare parts on-vehicle during a single inspection with pre-positioned spare parts reserves in the area. This aligns with the actual operation and maintenance characteristics of the chain propagation of manhole cover failures in different areas, improving the overall reliability of the operation and maintenance plan. Attached Figure Description

[0015] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0016] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 An architecture diagram of the operation and maintenance system for power manhole covers provided in Embodiment 1 of the method for generating operation and maintenance solutions for power manhole covers in this application; Figure 2 A flowchart illustrating the second embodiment of the method for generating an operation and maintenance plan for power manhole covers in this application; Figure 3 This is a flowchart of data acquisition and aggregation provided for Embodiment 2 of this application; Figure 4 This is a flowchart of a health assessment process provided in Embodiment 2 of this application; Figure 5 A flowchart illustrating the fourth embodiment of the method for generating an operation and maintenance plan for power manhole covers in this application; Figure 6 This is a schematic diagram of the equipment structure of the hardware operating environment involved in the method for generating the operation and maintenance plan for power manhole covers in this application embodiment.

[0018] The purpose, features, and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0019] It should be understood that the specific embodiments described herein are only used to explain the technical solutions of this application and are not intended to limit this application.

[0020] To better understand the technical solution of this application, a detailed description will be provided below in conjunction with the accompanying drawings and specific embodiments. It should be noted that all actions involving the acquisition of signals, information, or data in this application are performed in accordance with the relevant data protection laws and regulations of the country where the application is located, and with authorization from the owner of the corresponding device.

[0021] Current methods for generating manhole cover maintenance plans only collect and analyze individual manhole operation monitoring data, relying on threshold judgments or individual prediction models to complete fault assessment and maintenance scheduling. However, power manhole covers are typically laid out in a continuous spatial pattern along cable lines, and manhole cover faults are prone to cascading transmission due to pipeline connections. Therefore, the data processing logic of the above-mentioned methods fragments the regional group relationships between manhole covers, making it difficult to adapt to the operating conditions of continuous fault transmission, ultimately resulting in low reliability of the generated maintenance plans.

[0022] In view of the above problems, this application proposes a method for generating operation and maintenance plans for power manhole covers. First, it acquires multi-dimensional status data of manhole covers in each jurisdiction collected by a regional gateway, and constructs input features based on this data. Second, it constructs a regional manhole cover relationship graph and determines the corresponding adjacency matrix based on the geographical coordinates of the manhole covers and the cable channel topology. Then, it inputs the input features and adjacency matrix into a time-series neural network model to predict the health trend and failure probability of each manhole cover, thereby calculating the urgency score of each manhole cover. Next, it constructs a risk propagation model based on the adjacency matrix, updates the collaborative urgency of the manhole covers, and dynamically plans and generates the optimal inspection route based on the collaborative urgency, geographical road network, and operation and maintenance resource constraints. Simultaneously, it generates a spare parts list based on simulated maintenance results, determines the pre-positioned location and quantity of spare parts through cluster analysis, and finally generates a precise spare parts scheduling strategy adapted to the operation and maintenance scenario. This approach balances the configuration of spare parts on-vehicle during a single inspection with pre-positioned spare parts reserves in the area, aligning with the actual operation and maintenance characteristics of the chain propagation of manhole cover failures and improving the overall reliability of the operation and maintenance plan.

[0023] It should be noted that the executing entity in this embodiment can be a computing service device with data processing, network communication, and program execution functions, such as a tablet computer, personal computer, or mobile phone, or an electronic device capable of performing the above functions, such as an operation and maintenance system for power manhole covers. The following description uses an operation and maintenance system for power manhole covers as an example to illustrate this embodiment and the subsequent embodiments.

[0024] Example 1 Embodiment 1 of this application provides an operation and maintenance system for power manhole covers, referring to... Figure 1 , Figure 1 An architecture diagram of the operation and maintenance system for power manhole covers is provided, specifically: The overall architecture of the operation and maintenance system for power manhole covers is divided into three layers: the perception layer, the regional gateway, and the cloud decision-making layer.

[0025] The sensing layer is responsible for real-time and accurate collection of data on the physical status of each smart manhole cover and the underground environment. The sensing layer includes a status monitoring sensor group, a battery management unit, a low-power wireless communication module, and a microprocessor. The status monitoring sensor group includes a triaxial accelerometer, tilt sensor, water level sensor, temperature sensor, and gas sensor, used to detect the physical state and environmental parameters of the manhole cover. The battery management unit monitors data such as the voltage and remaining capacity of the manhole cover's power supply system. The microcontroller performs preliminary data filtering and outlier removal on the collected data at the manhole cover end, completing local preprocessing. The low-power wireless communication module is responsible for uploading the collected data to the regional gateway.

[0026] The regional gateway is deployed in the area near the manhole covers, responsible for aggregating and processing data from multiple manhole covers within a jurisdiction. The regional gateway includes a data aggregation module, a spatial coding module, a health assessment module, a population prediction model, and an edge reporting module. The data aggregation module periodically collects multi-dimensional status data reported by all manhole covers within the jurisdiction and organizes it into a structured state matrix. The spatial coding module constructs a regional manhole cover relationship map based on the geographic coordinates of the manhole covers and the cable channel topology, and determines the adjacency matrix based on this map. This regional manhole cover relationship map defines the spatial associations between manhole covers within the jurisdiction, such as proximity or environmental similarity, enabling the system to analyze using spatial correlations between populations, rather than treating each manhole cover as an isolated point. The health assessment module fuses the multi-dimensional status data to calculate an individual health score. The population prediction model uses a graph temporal neural network model combining graph convolutional networks and gated recurrent units to determine the health change trend and failure probability of each manhole cover. The graph convolutional network of the graph temporal neural network model uses a regional manhole cover relationship graph generated by the spatial coding module to learn the spatial dependency patterns among manhole cover groups; the gated recurrent unit learns the evolution of the state of each manhole cover over time. The edge reporting module compresses the local prediction results, namely the aforementioned health change trends and failure probabilities, as well as necessary key state data, and uploads them to the cloud. At the same time, it receives model update instructions from the cloud, realizing cloud-edge collaboration.

[0027] The cloud-based decision-making layer comprises a multi-region aggregation module, an inspection route optimization module, a spare parts scheduling module, a visualization and work order push module, and a model continuous learning module. The multi-region aggregation module receives and integrates data from all regional gateways, forming a unified view of the overall manhole cover status, providing support for macro-management. The inspection route optimization module calculates the urgency score of each manhole cover based on the health trend and failure probability reported by the regional gateways; and generates inspection routes by combining the urgency score with the geographical road network. The spare parts scheduling module determines the future spare parts list based on the predicted health trend and failure probability, and generates a spare parts scheduling strategy in conjunction with the inventory system. The visualization and work order push module displays the overall manhole cover status in the form of map heatmaps, and automatically encapsulates the optimized inspection routes and spare parts scheduling strategies into electronic work orders, pushing them to the mobile terminals of maintenance personnel to guide on-site operations. The model continuous learning module collects actual results from maintenance personnel, compares them with predicted results, periodically retrains and optimizes the graph temporal neural network model, and distributes it to regional gateways via the cloud, enabling the system to continuously improve itself and adapt to environmental changes.

[0028] Example 2 Based on the operation and maintenance system for power manhole covers proposed in Embodiment 1 above, Embodiment 1 of this application provides a method for generating operation and maintenance plans for power manhole covers, referring to... Figure 2In this embodiment, the method for generating the operation and maintenance plan for the power manhole cover includes steps S10 to S100: Step S10: Obtain multi-dimensional status data of manhole covers in each jurisdiction collected by the regional gateway, and construct input features based on the multi-dimensional status data.

[0029] It should be noted that the regional gateway is an edge computing node deployed along the cable channel, such as in ring main units and utility poles, responsible for aggregating and processing data from multiple smart manhole covers within its communication coverage area. Multidimensional status data refers to a collection of multidimensional data reflecting the operational status of each manhole cover, collected from its status monitoring sensor group. This data includes, but is not limited to, battery voltage, battery capacity, manhole cover tilt angle, water level, temperature, gas concentration, triaxial acceleration, communication signal strength, and reporting time.

[0030] As an alternative approach to constructing input features, at the regional gateway side, the multidimensional state data of all manhole covers collected at a certain moment within the jurisdiction is arranged in order of the manhole covers to form a state matrix, which serves as the aforementioned input features. Each row of the state matrix represents a manhole cover, and each column represents a state dimension.

[0031] For example, the microcontroller inside each manhole cover triggers and reads the raw data collected by the status monitoring sensor group according to a preset fixed cycle, and performs preliminary verification on the collected raw data to remove invalid values ​​that are obviously out of range or abrupt values ​​that do not conform to physical laws caused by instantaneous sensor failure; and performs aggregation calculations on the continuous raw data within a preset time to generate statistical features, such as mean, maximum value and root mean square value, to smooth noise and extract effective information.

[0032] Then, the processed data from each dimension are packaged together with the manhole cover identifier and the reporting timestamp to obtain multi-dimensional status data of the manhole cover with timestamps. Next, the regional gateway receives and parses the multi-dimensional status data reported by each manhole cover within its jurisdiction. Data received from N manhole covers at the same collection time t is arranged in order of manhole cover identifier, forming an N×F two-dimensional status matrix. Here, N is the number of manhole covers, and F is the status dimension of each manhole cover, i.e., the number of parameters such as voltage, tilt angle, and water level. Simultaneously, the regional gateway also maintains a database containing the static attributes of each manhole cover, recording inherent information such as manhole cover identifier, latitude and longitude coordinates, cable trench to which it belongs, installation date, material, and well depth. The static attributes of the manhole cover are combined with the status matrix to form input features.

[0033] As another alternative approach to constructing input features, refer to Figure 3The provided data acquisition and aggregation flowchart, based on the aforementioned construction of the state matrix, further involves the regional gateway first aligning the state matrices of all manhole covers according to the reported timestamps. For cases where data for a particular manhole cover is missing at a specific time due to communication interruptions or other reasons, interpolation methods are used to fill the gaps. These filling methods specifically include spatial interpolation and temporal interpolation. Spatial interpolation refers to interpolating data using data from other manhole covers geographically adjacent to the missing cover or within the same environmental area at the same time; temporal interpolation refers to interpolating data from the manhole cover itself at historical time points before and after the missing time.

[0034] Furthermore, referring to Figure 4 The provided health assessment flowchart uses the current multidimensional state data of each manhole cover i, after time alignment and missing value imputation, as input parameters to preset scoring functions for five assessment dimensions: power supply, structure, environment, communication, and function. These preset scoring functions determine the current health score H_i(t) of the manhole cover. Specifically, the power supply health score is calculated using voltage attenuation rate and remaining capacity; the structural health score is calculated using absolute tilt angle and tilt angle change rate; the environmental health score is calculated using water level percentage and temperature deviation; the communication health score is calculated using signal strength and reporting success rate; and the functional health score is calculated using unlocking self-test and sensor self-test results. Finally, the individual scores calculated from these five assessment dimensions are weighted according to preset weight ratios to obtain the current health score of the manhole cover.

[0035] For example, the monitoring parameters and scoring functions corresponding to each evaluation dimension are shown in the table below:

[0036] The formula for the health score H_i can be expressed as:

[0037] in, manhole cover Health score, Assigning weights to each evaluation dimension. For the normalized preset scoring function, For the first The monitoring parameters are divided into several dimensions. The above health score can also be divided into the following 5 health levels to determine the corresponding operation and maintenance recommendations: When the score is ≥90, the health level is excellent, the body is functioning normally, and no intervention is required. 75≤ If the health level is less than 90, the health level is good; observe closely and include in routine inspections. 60≤ If the blood pressure is below 75, the health level is "concerned" and an upcoming check-up is required. 40≤ If the health level is below 60, it is a warning level and should be addressed as soon as possible. If the value is less than 40, the health level is considered severe and requires immediate attention, as there is a risk of malfunction.

[0038] Furthermore, based on the aforementioned health scores of the manhole covers, their health score change rate is calculated. Next, the health score, health score change rate, and the aforementioned multidimensional state data of the manhole covers at the current moment are normalized. The normalized features are then combined with environmental context, such as seasonal coding and rainfall levels, to form the node feature vector for each manhole cover at time t. For N manhole covers within the jurisdiction, the node feature vectors of each manhole cover over the past T consecutive time steps are extracted, forming a three-dimensional tensor T×N×F. This three-dimensional tensor is used as the aforementioned input feature. Here, F represents the feature dimension of each manhole cover at each time step in the three-dimensional tensor.

[0039] Step S20: Based on the geographical coordinates of the manhole cover and the topology of the cable channel, construct a regional manhole cover relationship map, and determine the adjacency matrix based on the regional manhole cover relationship map.

[0040] It should be noted that the geographic coordinates and cable channel topology of the aforementioned manhole covers can be extracted from the static attributes of the manhole covers. Specifically, the geographic coordinates of the manhole covers can be represented by latitude and longitude coordinates, and the cable channel topology can include the cable trench to which the manhole cover belongs, the power supply circuit number, and the uplink communication relay identifier. The purpose of constructing the aforementioned regional manhole cover relationship graph is to abstract the discretely distributed manhole covers in the physical world into a mathematical graph structure based on their spatial location and functional relationships, thereby providing a foundation for subsequent graph neural network models to capture the interdependencies between manhole covers.

[0041] Step S30: Input the input features and the adjacency matrix into the graph temporal neural network model, and determine the health trend and failure probability of each manhole cover through the graph temporal neural network model.

[0042] The health status trend is a predicted sequence of health scores for each manhole cover over a future period, output by a graph time-series neural network model, representing how the health scores of each manhole cover change over time. The failure probability represents the likelihood that the health score of a manhole cover will drop to a preset health level, such as the "warning" or "serious" level mentioned above, within a specific future time period.

[0043] For example, assuming the aforementioned three-dimensional tensor T×N×F is used as input feature X, the input feature X and the aforementioned N×N adjacency matrix A are converted into the tensor format required by the graph temporal neural network model before being input into the graph temporal neural network model. In the graph convolutional network, for each slice X_t of shape [N,F] in the historical time step t of the input feature X, the graph convolutional network calculates it together with the adjacency matrix A, weightedly aggregating the features of each manhole cover node with the features of its neighboring nodes, thereby encoding spatial association information into the feature representation of each node. Finally, the graph convolutional network outputs a node feature sequence enhanced with spatial information.

[0044] Alternatively, the propagation rules of graph convolutional networks can be expressed as:

[0045] in , It is an adjacency matrix. It is the identity matrix. It is a degree matrix, that is The sum of each row is used as the value on the diagonal of the degree matrix. Let H(l) be the trainable weight matrix, and H(l) be the input features of the l-th layer graph convolutional network. It is the output feature matrix of the (l+1)th layer graph convolutional network.

[0046] Furthermore, the aforementioned node feature sequences are input along the time dimension T into a gated recurrent unit (GRU). The GRU iterates sequentially according to time steps. Its internal gating mechanism enables it to capture and memorize long-term and short-term dependencies in the node feature sequences, thereby learning the dynamic pattern of the evolution of each manhole cover's state and ultimately outputting the hidden state of each manhole cover node at the end of the time series. These hidden states are then mapped through one or more fully connected layers in the output layer of the graph temporal neural network model. The trend prediction header of the output layer outputs a sequence of health prediction values ​​for the next K time steps, i.e., [H_i(t+1), H_i(t+2), ..., H_i(t+K)], with a shape of [N,K], representing the health change trend of each manhole cover. The fault probability header of the output layer outputs a scalar value. , representing the probability of each manhole cover malfunctioning within a specific time period in the future, with a shape of [N,1].

[0047] Step S40: Determine the urgency score of each manhole cover based on the health status change trend and the failure probability.

[0048] As an alternative method for determining the urgency score of manhole covers, the lowest health score is extracted from the above health score prediction value sequence and denoted as... The lower the minimum health value, the worse the condition of the manhole cover. Simultaneously, the average rate of change of health over a future period is calculated as the predicted rate of decline in health, denoted as Δ. A negative and larger absolute value of the predicted rate of decline indicates a more rapid deterioration trend. The urgency score E_i is obtained by weighting and fusing the minimum health value, predicted rate of decline, and failure probability using a linear weighting formula, which can be expressed as: E_i=α1×(1- )+α2×|Δ |+α3× , where α1, α2, and α3 are configurable weight coefficients and are all positive numbers.

[0049] As another alternative for determining the urgency score of a manhole cover, the urgency score corresponding to the minimum health value, predicted descent rate, and failure probability range can be predefined, and then the urgency score of the manhole cover can be determined through rule mapping.

[0050] Step S50: Based on the weighted connection weights in the adjacency matrix, a risk propagation model is constructed, and the collaborative urgency of each manhole cover is determined according to the urgency score and the risk propagation model.

[0051] The risk propagation model uses the weighted connection weights between manhole covers as transmission coefficients to describe the spread or superposition of manhole cover failure risks or abnormal states within a regional group. For example, the weighted connection weights between all nodes are extracted from the generated adjacency matrix. These weighted connection weights are used as the transmission coefficients of risk between adjacent manhole covers to construct a risk propagation model with manhole covers as nodes and the transmission coefficients as propagation strength. Further, the urgency score calculated for each manhole cover is obtained. Through the risk propagation model, the urgency score of each manhole cover is propagated to its adjacent associated manhole covers according to the corresponding weighted connection weights. That is, for each manhole cover, its own urgency score is weighted and summed with the urgency scores of all associated manhole covers transmitted through the propagation. The result of this weighted summation is used as the collaborative urgency of that manhole cover. Here, the weighted connection weights are used to quantify the state association strength between manhole covers; the collaborative urgency of a manhole cover is the weighted sum of its own urgency score and the urgency scores of its other associated manhole covers.

[0052] For example, let the total number of manhole covers be N, and the collaborative urgency vector be... The urgency score vector of the manhole cover itself is The adjacency matrix is Elements in the matrix Let be the weighted connection weight between manhole cover node i and manhole cover node j. Let be the risk propagation coefficient, used to control the intensity of risk transmission from associated manhole covers. Then, the expression for calculating the collaborative urgency of the risk propagation model can be:

[0053] For a single manhole cover i, the scalar expression for calculating its collaborative urgency is:

[0054] in, For the collaborative urgency of manhole cover i, Rate the urgency of manhole cover i. The urgency score is given to manhole cover j associated with manhole cover i.

[0055] The above steps extract event fragments from state time-series data, construct event morphological feature sequences, and calculate morphological similarity. Combined with the event start time, they determine the event delay relationship and ultimately fuse them to obtain a dynamic correlation coefficient that better reflects the actual state transmission law. This allows the weighted weights of the adjacency matrix to reflect the state correlation strength between manhole covers in real time. This effectively improves the representation accuracy of spatial correlation between manhole covers, enhances the learning ability of graph time-series neural networks to the evolution of group states, and makes subsequent health prediction, fault probability judgment, and urgency assessment more in line with the actual situation on site, thereby generating more reliable operation and maintenance solutions.

[0056] Step S60: Using the collaborative urgency, the adjacency matrix, the driving distance between the manhole covers, and the maintenance resource constraints as input parameters for the preset path planning algorithm, the first round of inspection path planning is performed in the geographical road network to generate an initial inspection sequence containing at least one manhole cover to be visited.

[0057] A geographic road network is a data model that includes the geographic coordinates of all manhole covers and the actual road network information connecting these geographic coordinates. It is used to calculate the travel distance between manhole covers during inspections.

[0058] For example, the cloud-based decision layer first normalizes the collaborative urgency of all manhole covers, using the normalization result as the inspection priority weight for each manhole cover node; a higher inspection priority weight indicates a higher inspection priority. Then, it extracts the weighted connection weights between manhole covers based on the adjacency matrix, grouping manhole covers with weighted connection weights greater than a preset weight threshold into the same inspection cluster, and setting an association constraint that manhole covers within the same inspection cluster must be visited consecutively. Based on the geographical road network, it calculates the shortest path length between any two manhole covers to be inspected, using this as the travel distance between the manhole covers and the movement cost for path planning. Simultaneously, it obtains maintenance resource constraints, using them as solution constraints for path planning. Subsequently, the cloud-based decision layer synchronously inputs the above inspection priority weights, association constraints, movement costs, and solution constraints into a preset path planning algorithm, which can employ genetic algorithms or ant colony algorithms, etc. Based on the geographical road network, the cloud-based decision-making layer constructs a planar path search space for all manhole covers to be inspected and the starting positions of each maintenance team. The preset path planning algorithm is limited to path optimization within this space. At the same time, the algorithm optimization objectives are set to prioritize accessing manhole covers with high priority weights, minimize movement costs, and fully satisfy maintenance resource constraints and related constraints. Finally, the initial inspection sequence that satisfies all constraints and optimization objectives is output.

[0059] Step S70: Sequentially simulate the maintenance operations of the manhole covers in the initial inspection sequence, determine the spare parts demand type of the current manhole cover based on the scores of each dimension in the health score of the current manhole cover being maintained, and update the collaborative urgency of the remaining manhole covers in the geographical road network based on the adjacency matrix.

[0060] For example, the cloud-based decision layer selects the current manhole cover to be repaired sequentially according to the initial inspection sequence. Within the power manhole cover maintenance system, it simulates the repair operation on the manhole cover and reads its health score, including sub-scores for five dimensions: power supply, structure, environment, communication, and function. Each dimension score is compared with preset dimension thresholds to identify abnormal dimensions. Further, based on these abnormal dimensions, the required spare parts type for the current manhole cover is determined. Then, the cloud-based decision layer reads the weighted connection weights between the current manhole cover and the remaining manhole covers in the adjacency matrix to obtain the state association strength between the manhole covers. Based on the health recovery status of the current manhole cover after repair, it corrects the risk propagation strength using the weighted connection weights. Using the corrected risk propagation strength, it recalculates the collaborative urgency of all remaining manhole covers in the geographical network, completing the update of the collaborative urgency of the remaining manhole covers. Here, "remaining manhole covers" refers to those manhole covers that did not undergo simulated repair operations in the initial inspection sequence.

[0061] Furthermore, taking the risk propagation model proposed above as an example, the specific steps of updating the collaborative urgency of the remaining manhole covers in the geographical road network according to the adjacency matrix may include: firstly, extracting the weighted connection weights between the currently repaired manhole cover and each remaining manhole cover from the adjacency matrix. The weighted connection weight is used as a quantitative parameter for the strength of the state correlation between manhole covers. Then, the health scores of the manhole covers before and after maintenance are compared to determine the health recovery difference. This health recovery difference is then multiplied by a preset risk attenuation coefficient to generate a corresponding risk propagation correction factor. Finally, the risk propagation coefficient is adjusted based on this risk propagation correction factor. The health recovery difference is positively proportional to the risk propagation correction factor, while the risk propagation correction factor is inversely proportional to the risk propagation coefficient. Simultaneously, the urgency level of the currently repaired manhole cover is updated based on its health score after repair. Finally, the risk propagation coefficient, corrected for the updated urgency level of the currently repaired manhole cover, is substituted back into the aforementioned collaborative urgency calculation formula, and the collaborative urgency level of all remaining manhole covers in the geographical road network is updated using the risk propagation model.

[0062] Step S80: When the updated collaborative urgency meets the replanning condition, based on the updated collaborative urgency, the current maintenance location and the remaining manhole covers, the inspection path is replanned to generate the subsequent inspection sequence, and the maintenance operation of the manhole covers in the subsequent inspection sequence is simulated sequentially until the maintenance resource constraint is reached, and the inspection route is output.

[0063] For example, the cloud-based decision layer retrieves two sets of collaborative urgency data for all remaining manhole covers before and after the update, and calculates the decrease or increase in the collaborative urgency of each individual remaining manhole cover. It then iterates through the amplitude data for all remaining manhole covers, sequentially verifying the replanning judgment rules.

[0064] Furthermore, the aforementioned replanning conditions include one or more of the following: The ratio of the decrease in the collaborative urgency of any remaining manhole cover to the decrease in the collaborative urgency of other remaining manhole covers is greater than a preset proportional threshold. The urgency level of any remaining manhole cover increases; The increase in the collaborative urgency level of any remaining manhole cover exceeds a preset threshold.

[0065] When the collaborative urgency of remaining manhole covers meets one or more of the above replanning conditions, the updated collaborative urgency, the current maintenance team's repair location, and all uninspected remaining manhole covers are used as input parameters. The preset path planning algorithm corresponding to step S60 is reused to perform the inspection path optimization calculation within the geographical road network again, generating the subsequent inspection sequence. According to the order of the subsequent inspection sequence, simulated maintenance operations are carried out on each manhole cover in the sequence, and the processing flow of spare parts demand determination and remaining manhole cover collaborative urgency update in step S70 is repeated until resource consumption reaches the upper limit of maintenance resource constraints. At this point, the iterative calculation is terminated, all inspection sequences generated throughout the process are summarized, and the final inspection route is output.

[0066] Step S90: Generate a spare parts list based on the spare parts requirement types of all manhole covers performing simulated maintenance operations, and determine the preset location and quantity of spare parts by clustering based on the spatial distribution of the collaborative urgency of each manhole cover and the adjacency matrix.

[0067] Furthermore, the steps described above, based on the spatial distribution of the collaborative urgency of each manhole cover and the adjacency matrix, to determine the pre-positioned location and quantity of spare parts through clustering, include steps S91 to S94: Step S91: Using the collaborative urgency of each manhole cover as the node weight and the adjacency matrix as the connection relationship between nodes, a weighted graph is constructed.

[0068] A weighted graph is a topological graph where manhole covers are nodes, collaborative urgency is the node weight, and an adjacency matrix represents the connections between nodes. For example, each manhole cover is treated as an independent node in the weighted graph. The updated collaborative urgency of each manhole cover is obtained and set as the node weight of the corresponding node. Then, the weighted connection weights of each element in the adjacency matrix are read, and the edge connections and weights between different nodes are determined based on these weighted connection weights. Furthermore, the geographic coordinates corresponding to the manhole covers are bound to the corresponding nodes, completing the construction of a weighted graph with spatial attributes.

[0069] Step S92: Based on the geographical location of the pre-installed parts inventory, perform spatial clustering on the weighted graph to divide each manhole cover node into at least one high-risk area.

[0070] High-risk areas are groups of manhole covers that have been clustered together. Within high-risk areas, the overall urgency of manhole cover coordination is higher, and the probability of needing spare parts for malfunctions is greater.

[0071] For example, the cloud-based decision layer imports the geographical location of the pre-installed parts inventory, uses the geographical location of the pre-installed parts inventory as the clustering anchor point, and then uses an inter-clustering algorithm to combine the node weights of the weighted graph and the topological connection relationships between nodes to cluster all manhole cover nodes, and delineates each set of manhole covers formed after clustering as a single high-risk area.

[0072] Step S93: Determine the geographical center of the high-risk area as the pre-positioned location of the spare parts corresponding to the high-risk area.

[0073] The pre-positioned spare parts location refers to the selected storage location for spare parts within a high-risk area. For example, the geographic coordinates of all manhole covers within a single high-risk area are extracted, and the geographic center of the high-risk area is determined based on the geographic coordinates of all manhole covers. The calculated geographic center is then set as the pre-positioned spare parts location corresponding to the high-risk area.

[0074] Understandably, spatial clustering to delineate high-risk areas based on the spatial distribution and topological relationships of manhole covers' collaborative urgency can break the extensive model of scattered stockpiling for individual manholes. By combining the concentrated deployment of spare parts locations for regional overall fault risks, it aligns with the on-site operation and maintenance characteristics of power manhole cover faults being regionally correlated and risk transmission occurring in contiguous areas. It also takes into account the needs of both vehicle-mounted spare parts for inspections and pre-positioned spare parts for regional reserves, matching the actual working conditions of concentrated fault outbreaks in the area. This allows spare parts resources to shift from uniform and blind deployment to fixed-point pre-positioning following the risk distribution.

[0075] Step S94: Determine the number of spare parts corresponding to each of the pre-positioned spare parts locations based on the total collaborative urgency value of manhole covers in the high-risk area corresponding to each of the pre-positioned spare parts locations, the number of manhole covers, and the historical replacement frequency of similar spare parts.

[0076] Step S100: Generate a spare parts scheduling strategy based on the spare parts list, the preset location of the spare parts, and the quantity of the spare parts.

[0077] For example, the coordinated urgency of all manhole covers within a single high-risk area is calculated separately, and the total coordinated urgency value for the area is obtained by summing them up. Simultaneously, the number of manhole covers within each high-risk area is counted, and the historical replacement frequency of similar spare parts in the spare parts list for each type of spare part in its corresponding high-risk area is obtained. A weighted calculation is performed on the total coordinated urgency value, the number of manhole covers, and the historical replacement frequency of similar spare parts to determine the spare parts quantity for each type of spare part at the corresponding preset location within the high-risk area. The spare parts list, preset locations, and quantities are then packaged and integrated into a preset format to generate a spare parts scheduling strategy.

[0078] The above steps are based on a risk propagation model. Further, they combine the urgency score of the manhole cover itself to calculate the collaborative urgency of the associated risk transmission. Using this collaborative urgency, the travel distance between manhole covers, and maintenance resource constraints as input parameters, an initial inspection sequence is obtained through the first round of inspection path planning. Then, maintenance operations on the manhole covers in the initial inspection sequence are simulated sequentially. Following this, based on the changes in the health of the manhole covers after maintenance, the collaborative urgency of the remaining manhole covers in the geographical road network is dynamically updated using the risk propagation model. When the updated collaborative urgency meets the replanning conditions, the inspection path planning is iteratively re-performed to generate subsequent inspection sequences. Maintenance operations are continuously simulated sequentially until maintenance resource constraints are met, at which point the inspection route is output. Finally, a spare parts list is generated by summarizing the spare parts requirements of all manhole covers that have undergone simulated maintenance operations. At the same time, the spatial distribution of the collaborative urgency of each manhole cover and the adjacency matrix are integrated. High-risk areas are divided through spatial clustering, and the geographical center of each high-risk area is determined as the pre-positioned location for spare parts. The number of spare parts corresponding to each pre-positioned location is determined based on the total collaborative urgency of manhole covers in the corresponding high-risk area, the number of manhole covers, and the historical replacement frequency of similar spare parts. This balances the configuration of spare parts on the vehicle during a single inspection with the pre-positioned spare parts reserve in the area, and conforms to the actual operation and maintenance characteristics of the chain propagation of manhole cover failure areas. This not only improves the rationality of inspection layout and the efficiency of on-site fault handling, but also enables the fine-grained deployment of spare parts resources on demand, effectively reducing inventory redundancy and material allocation costs, and improving the reliability of the operation and maintenance plan.

[0079] Example 3 Based on the above embodiment two, in the method for generating operation and maintenance schemes for power manhole covers proposed in this embodiment three, the step of constructing a regional manhole cover relationship map based on the geographical coordinates of the manhole cover and the cable channel topology includes steps S21 to S24: Step S21: Each of the manhole covers is designated as a node in the regional manhole cover relationship diagram.

[0080] Step S22: Determine the distance between each manhole cover based on its geographical coordinates. When the distance is less than a preset distance threshold, construct a topological adjacent edge between the nodes corresponding to the manhole covers.

[0081] Topological adjacent edges refer to the edges that connect two manhole cover nodes that are physically adjacent along the cable channel and whose distance between them is within a preset distance threshold in the regional manhole cover relationship diagram. This directly reflects the physical connection and spatial continuity of the cable network itself.

[0082] For example, a list of identifiers for all manhole covers managed by the regional gateway is obtained, and a corresponding node is created for each manhole cover in the identifier list. Each node will serve as a node in the regional manhole cover relationship graph and will be assigned a unique index number. Then, all manhole cover node pairs are traversed, and the distance between each pair of manhole covers is calculated based on their latitude and longitude coordinates. The calculated distance between each pair of manhole covers is compared with a preset distance threshold. For any two manhole cover nodes i and j, if the distance between them is less than the preset distance threshold, a topologically adjacent edge is created between these two nodes in the regional manhole cover relationship graph.

[0083] Step S23: Determine the regional environment type of the manhole cover based on its geographical coordinates, and construct environmental similarity edges between nodes corresponding to manhole covers with the same regional environment type.

[0084] Regional environment type is a classification label for the external environment in which the manhole cover is located, including but not limited to low-lying areas, main traffic arteries, and areas with soft soil. Manhole covers of the same environment type may face similar external risks, while environmental similarity edges are edges connecting two manhole cover nodes with the same regional environment type in the regional manhole cover relationship diagram, indicating that these manhole covers may be affected by similar external environmental factors, such as rainwater accumulation or vehicle load, and their state changes may be correlated.

[0085] For example, a regional environment type classification map is obtained, which defines the environment types to which different coordinate regions belong. The geographic coordinates of each manhole cover are matched with the regional environment type classification map to determine the specific regional environment type to which the manhole cover belongs. All manhole cover node pairs are traversed. For any two manhole cover nodes i and j, if manhole cover nodes i and j are matched to the same regional environment type, an environment similarity edge is created between these two nodes in the regional manhole cover relationship graph. The environment similarity edge and the aforementioned topological adjacency edge can coexist; that is, two nodes can be connected simultaneously by an environment similarity edge and a topological adjacency edge.

[0086] Step S24: Based on the cable channel topology, construct functional association edges between nodes corresponding to manhole covers that have channel topology dependencies.

[0087] Channel topology dependency refers to the functional or logical dependencies between manhole covers in a cable channel system, including sharing the same power circuit, uploading data through the same communication relay node, and / or belonging to the upstream and downstream of the same physical cable trench. Functional association edges, on the other hand, are edges connecting two manhole cover nodes with the aforementioned channel topology dependencies in the regional manhole cover relationship diagram, reflecting the coupling relationship of manhole covers at the system functional level, such as power supply or communication.

[0088] For example, the operation and maintenance system for power manhole covers reads the static attributes of each manhole cover, identifies the fields defining functional dependencies, and establishes connections between manhole cover nodes with specific channel topology dependencies. For instance, if two manhole covers belong to the same cable trench, they are considered to be upstream and downstream of the same physical cable trench, and a functional association edge is established. If two manhole covers have the same power supply circuit number, they are considered to share the same power supply circuit, and a functional association edge is established. If two manhole covers have the same uplink communication relay identifier, they are considered to upload data through the same communication relay node, and a functional association edge is established. Similarly, functional association edges, environmental similarity edges, and topological adjacency edges can coexist; that is, two nodes can be connected simultaneously by functional association edges, environmental similarity edges, and topological adjacency edges.

[0089] Furthermore, the above-mentioned step of determining the adjacency matrix based on the regional manhole cover relationship diagram includes steps S25 to S28: Step S25: Determine the initial connection weight of the edge according to the type of edge between each node in the regional manhole cover relationship diagram.

[0090] In the regional manhole cover relationship diagram, the types of edges between nodes, namely the topological adjacent edges, environmentally similar edges, and functionally related edges mentioned above, can be set with different or the same initial connection weights.

[0091] Step S26: Based on the time sequence data of the status of the manhole covers corresponding to the connected nodes within a preset time window, determine the dynamic correlation coefficient between the manhole covers corresponding to the connected nodes.

[0092] Status time-series data refers to a time series of status data of a manhole cover arranged chronologically within a preset time window. This may include the aforementioned health score, battery voltage, battery capacity, manhole cover tilt angle, water level, temperature, gas concentration, triaxial acceleration, and communication signal strength. The dynamic correlation coefficient is used to quantify the degree of synchronization between the status time-series data of two connected manhole cover nodes within the preset time window.

[0093] For example, for any two manhole cover nodes i and j connected by an edge in the regional manhole cover relationship graph, extract the state time series data of manhole cover nodes i and j within the same preset time window, representing the same state dimension. Calculate the statistical correlation, such as the Pearson correlation coefficient, between the two extracted time series data to reflect the similarity of the recent historical trends in the states of the two manhole covers. Use the aforementioned Pearson correlation coefficient as the dynamic correlation coefficient between manhole cover node i and manhole cover node j. .

[0094] Step S27: Adjust the initial connection weights of the edges between the connected nodes according to the dynamic correlation coefficient to obtain the weighted connection weights.

[0095] For example, a pre-defined weight fusion function is used to fuse the static initial connection weights and the dynamic correlation coefficients to obtain the weighted connection weights. The weight fusion function can be expressed as: = ×(α+β×| |), where, It is the weighted connection weight between manhole cover node i and manhole cover node j; α is the initial connection weight between manhole cover node i and manhole cover node j; α and β are adjustable parameters used to control the relative importance of static prior weights and dynamic correlation coefficients.

[0096] Step S28: Construct a weight matrix based on the weighted connection weights, and normalize the weight matrix based on a preset weight threshold to generate the adjacency matrix.

[0097] For example, initialize an N×N zero matrix W, traverse each edge in the regional manhole cover relationship graph, and for the edge connecting manhole cover node i and manhole cover node j, set the element values ​​of the i-th row and j-th column of matrix W as the weighted connection weight of that edge. To avoid noise from very weak connections, a preset weight threshold can be set. Iterate through all non-zero elements in the weight matrix W; if its value is less than the weight threshold, set it to 0. Then, normalize the weight matrix W, and use the normalized weight matrix A as the adjacency matrix, which can be expressed as A = ×W× Where D1 is a diagonal matrix, and its diagonal elements are... It equals the sum of all elements in the i-th row of the weight matrix W.

[0098] Example 4 Based on the above embodiments, in the method for generating an operation and maintenance plan for power manhole covers proposed in this embodiment, refer to Figure 5 Step S26 above further includes steps S261 to S265: Step S261: Based on the temporal characteristics of the state time series data, extract multiple event segments corresponding to the manhole cover from the state time series data.

[0099] State time series data refers to a sequence of multidimensional state data of a manhole cover arranged chronologically within a preset time window. Time series features are quantitative indicators extracted from state time series data to describe its change patterns, such as mean, variance, trend slope, and peak value. An event segment is a subsequence extracted from state time series data. This subsequence typically corresponds to a meaningful state change process, such as the complete process of water level rising from its initial stage to its peak and then falling back, or a sustained and significant decline in health. An event segment includes a start time, an end time, and the original state time series data within that period.

[0100] For example, the temporal characteristics of the state time series data within a preset sliding time window are determined, resulting in a temporal feature sequence arranged chronologically. Then, the temporal feature difference values ​​between adjacent time points are calculated sequentially along the time dimension to quantify the magnitude of change in the temporal features within a unit time step. Next, the temporal feature difference value corresponding to each time point is compared with a preset change threshold. If the temporal feature difference value is greater than or equal to the change threshold, that time point is marked as a candidate change point. Temporal domain clustering is performed on all candidate change points, merging consecutive candidate change points with time intervals less than a preset minimum time interval into a change point cluster to eliminate isolated and unrelated single-point candidate change points. The earliest time point in each change point cluster is then determined as the starting time point of a significant state change in the state time series data. The temporal feature difference values ​​after this starting time point are traversed along the time dimension. When the temporal feature difference values ​​for a preset number of consecutive time steps are all less than the aforementioned change threshold, the last time point among the preset number of time steps is determined as the end time point. Finally, using the above start and end times as boundaries, the state time sequence data of the intermediate time period is obtained to obtain the event fragment.

[0101] Step S262: Construct an event feature sequence based on the state time sequence data corresponding to each event segment.

[0102] For example, the state time-series data corresponding to a single event segment is normalized along the time dimension to eliminate differences in units and numerical amplitudes, resulting in normalized event segment time-series data. For the normalized event segment time-series data, event features are extracted step-by-step. These features may include the aforementioned time-series features of the state time-series data corresponding to the event segment, as well as features such as instantaneous amplitude, instantaneous rate of change, cumulative change, fluctuation amplitude, and trend direction. The event features are then arranged in chronological order according to the event segments to obtain an event feature sequence. This event feature sequence is used to quantitatively describe the overall change pattern of the event.

[0103] Step S263: Based on the feature sequence, determine the morphological similarity between event segments of the manhole covers corresponding to the connected nodes within the preset time window.

[0104] Morphological similarity is a quantitative indicator used to measure the degree of similarity in morphology, such as the trend, magnitude, and continuity, between two different manhole cover event segments.

[0105] For example, within a preset time window, the event feature sequences corresponding to the event segments of manhole covers corresponding to two connected nodes are obtained. The two event feature sequences are time-aligned to ensure that the starting point and step size of their time dimensions are consistent. Then, using sequence similarity measurement methods, such as calculating Euclidean distance, Manhattan distance, and Pearson correlation coefficient, the feature value difference at each corresponding time step of the two event feature sequences is calculated, and the absolute value of the feature value difference is taken to obtain the numerical difference of a single time step. The numerical differences of all single time steps are accumulated sequentially along the time dimension to obtain the cumulative difference value of the entire event feature sequence. The morphological similarity between the two event segments is determined based on this cumulative difference value. The cumulative difference value is inversely proportional to the morphological similarity.

[0106] Step S264: Based on the start time of the event segment, determine the event delay relationship between the manhole covers corresponding to the connected nodes.

[0107] Event delay relationships describe the temporal sequence and intervals of events occurring on two different manhole covers, reflecting the possible direction and speed of propagation of state changes among the manhole cover group.

[0108] For example, the event delay relationship between manhole covers corresponding to connected nodes may include the absolute value of the event delay and the event propagation direction. The absolute value of the event delay is the time difference between the start times of two time segments. Based on the positive or negative sign of the time difference, the order in which the events occur is determined, and thus the event propagation direction is determined.

[0109] Step S265: Determine the dynamic correlation coefficient between the manhole covers corresponding to the connected nodes based on the morphological similarity and the event delay relationship.

[0110] In this embodiment, the dynamic correlation coefficient is a composite correlation strength index that integrates morphological similarity and event delay relationship. It not only indicates whether the state changes of two manhole covers are similar, but also includes whether such similarity follows a specific temporal sequence pattern, thereby enabling a more accurate characterization of the causal or transmissive dynamic correlation between manhole cover groups.

[0111] For example, based on the above-mentioned relationship between morphological similarity and event delay, a delay attenuation coefficient is determined according to the absolute value of the event delay; the smaller the absolute value of the delay, the larger the delay attenuation coefficient. The morphological similarity is multiplied by the delay attenuation coefficient to obtain a preliminary correlation value. Then, according to the event propagation direction, a direction identifier is added to the preliminary correlation value to retain the propagation direction information. Finally, the preliminary correlation value with the added direction identifier is normalized to obtain the dynamic correlation coefficient between the corresponding manhole covers of two connected nodes.

[0112] This embodiment detects change points and extracts event segments from the time-series data of manhole cover status, constructs a feature sequence representing the event morphology, calculates the morphological similarity of event segments between connected manhole covers, and determines the event delay relationship by combining the event start time. Finally, it fuses the morphological similarity and event delay relationship to obtain a dynamic correlation coefficient, realizing the dynamic quantification of the state association strength between manhole covers. This processing flow can not only accurately capture the temporal characteristics and spatial transmission laws of manhole cover status changes, but also effectively improve the accuracy and timeliness of the adjacency matrix weights, thereby improving the prediction accuracy of subsequent graph temporal neural network models and generating more reliable operation and maintenance solutions.

[0113] Example 5 Based on the above embodiments, in the method for generating operation and maintenance plans for power manhole covers proposed in this embodiment, after step S40, another optional scheme for generating spare parts scheduling strategies is also provided: Specifically, the operation and maintenance resource constraints are first obtained, including the number of available operation and maintenance teams, the starting location of each team, the maximum daily working hours, and vehicle capacity. The inspection task planning is modeled as a Vehicle Routing Problem with Time Windows (VRPTW). Here, "vehicle" corresponds to the operation and maintenance team; "customer point" corresponds to each manhole cover, and the "service time" for each manhole cover can be the average historical maintenance time; "demand" corresponds to the urgency score of each manhole cover; and "distance" or "cost" is the travel distance between two manhole covers calculated based on the geographical road network. Then, optimization algorithms such as genetic algorithms or ant colony algorithms are used to solve the VRPTW. The optimization objective is to minimize the total inspection distance or total travel time while maximizing coverage of high-urgency manhole covers, while satisfying the operation and maintenance resource constraints. Finally, an optimal access sequence is generated for each operation and maintenance team, i.e., the aforementioned inspection route.

[0114] For example, assuming the objective function is solved using a genetic algorithm, the optimization objective can be expressed as:

[0115] in, This is the collection of all inspection routes. An inspection route consists of the complete path taken by a maintenance team starting from the origin, visiting several manhole covers in sequence, and returning. For the first Total travel time for the route. The set of manhole covers to be visited for the r-th inspection route. A preset penalty coefficient is used to quantify the cost of "high-urgency manhole covers not being actually visited". The larger the penalty coefficient, the more the power manhole cover operation and maintenance system tends to force these manhole covers to be included in the inspection route. The urgency threshold can be set to 60 points. When... At that time, the manhole cover was marked as "high urgency" and, in principle, must be visited during this inspection. This is an indicator function; it takes the value 1 if the condition within the square brackets is true, and 0 otherwise. Here, the condition is "the urgency of manhole cover i exceeds the threshold and the manhole cover has not been actually accessed in the inspection route r". This represents the number of manhole covers with high urgency that should have been visited but were not actually visited in the statistical inspection route r.

[0116] Furthermore, the spare parts scheduling strategy includes a spare parts list for each manhole cover. For example, in the spare parts scheduling module, based on multi-dimensional status data of each manhole cover reported from the regional gateway, such as battery voltage, manhole cover tilt, and water level, it can be determined which status data meet preset maintenance condition thresholds. Each triggered maintenance condition corresponds to a specific status dimension anomaly. The power manhole cover operation and maintenance system will match and generate the corresponding spare parts list based on the triggered status dimension. For example, assuming that the battery capacity is less than 30% and the manhole cover tilt angle is greater than 30° in the multi-dimensional status data, the spare parts list will at least include the battery components corresponding to the battery status dimension and the manhole cover structure leveling and fixing component package corresponding to the manhole cover tilt status dimension.

[0117] Furthermore, in addition to the aforementioned spare parts list, the spare parts scheduling strategy may also include target warehouse-receiving point pairs corresponding to the spare parts scheduling. For example, the target warehouse-receiving point pairs corresponding to the spare parts scheduling can be determined based on real-time inventory system data, the aforementioned inspection routes, and the starting positions of each maintenance team.

[0118] For example, in determining the above inspection route, the starting position of each maintenance team G is already known, assuming it is the starting warehouse. Furthermore, its departure time is T_depart. The access sequence (inspection route) of each maintenance team G, as well as the service time corresponding to each manhole cover along the inspection route and the travel distance between manhole covers, are also known. Therefore, the estimated arrival time ETA_i for each manhole cover i can be determined. Simultaneously, the warehouse location and inventory quantity of each warehouse are obtained from the real-time inventory system data. For manhole cover i, its spare parts must be in place before maintenance team G begins service. Therefore, its demand time window is [T_depart, service start time]. For simplicity, a latest acceptable time T_latest_i can be set, typically ETA_i minus a safety buffer such as 30 minutes, to allow the maintenance team to receive and count the parts.

[0119] Next, iterate through each manhole cover i with spare parts requirements and its required spare parts list. Find the set of warehouses {W} whose inventory quantity can meet the spare parts list. For each warehouse Wj in the warehouse set {W}, calculate the time T_transport(Wj,Pk) required to ship the spare parts from Wj to a receiving point on the inspection route of the maintenance team G. The receiving point Pk can be: the starting warehouse W0 of the maintenance team, an intermediate meeting point on the inspection route, or the geographical coordinates of manhole cover i. Only retain warehouse-receiving point pairs (Wj,Pk) that can satisfy T_depart+T_transport<=T_latest_i. From all feasible warehouse-receiving point pairs (Wj,Pk), select the target warehouse-receiving point pair with the shortest travel path.

[0120] In this embodiment, the regional gateway first collects and preprocesses multi-dimensional status data of manhole covers to construct input features. Then, it combines geographical coordinates and cable channel topology to construct a regional manhole cover relationship graph and generate a dynamic adjacency matrix. Through a graph temporal neural network model, it predicts the health trend and failure probability of manhole covers. Based on this, it calculates the urgency score and combines it with the geographical road network to complete the inspection route planning and spare parts scheduling strategy generation. This upgrades single manhole cover monitoring to regional group collaborative prediction and decision-making, effectively utilizing the spatial correlation and temporal evolution law between manhole covers, realizing the transformation from passive emergency repair to predictive operation and maintenance, improving the accuracy of fault early warning, inspection efficiency and resource scheduling rationality, thereby improving the reliability of operation and maintenance solutions.

[0121] It should be noted that the above examples are only for understanding this application and do not constitute a limitation on the method for generating operation and maintenance solutions for power manhole covers in this application. Any simple modifications based on this technical concept are within the protection scope of this application.

[0122] This application provides a device for generating operation and maintenance plans for power manhole covers. The device includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the method for generating operation and maintenance plans for power manhole covers in the above embodiment 1.

[0123] The following is for reference. Figure 6 The diagram illustrates a structural schematic of a device suitable for generating operation and maintenance solutions for power manhole covers in accordance with embodiments of this application. The device for generating operation and maintenance solutions for power manhole covers in these embodiments may include, but is not limited to, mobile terminals such as laptops and tablets (PADs, Portable Application Description) and fixed terminals such as desktop computers. Figure 6 The power manhole cover maintenance solution generation device shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of this application.

[0124] like Figure 6 As shown, the power manhole cover maintenance plan generation device may include a processing unit 1001 (e.g., a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 1002 or the program loaded from the storage device 1003 into the random access memory (RAM) 1004. The random access memory 1004 also stores various programs and data required for the operation of the power manhole cover maintenance plan generation device. The processing unit 1001, the read-only memory 1002, and the random access memory 1004 are interconnected via a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems can be connected to I / O interface 1006: input devices 1007 including, for example, touchscreens, touchpads, keyboards, mice, image sensors, microphones, accelerometers, gyroscopes, etc.; output devices 1008 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 1003 including, for example, magnetic tapes, hard disks, etc.; and communication devices 1009. Communication device 1009 allows the power manhole cover maintenance solution generation device to communicate wirelessly or wiredly with other devices to exchange data. Although the figure shows a power manhole cover maintenance solution generation device with various systems, it should be understood that it is not required to implement or possess all the systems shown. More or fewer systems can be implemented alternatively.

[0125] Specifically, according to the embodiments disclosed in this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments disclosed in this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device, or installed from storage device 1003, or installed from read-only memory 1002. When the computer program is executed by processing device 1001, it performs the functions defined in the methods of the embodiments disclosed in this application.

[0126] The power manhole cover operation and maintenance scheme generation device provided in this application, employing the power manhole cover operation and maintenance scheme generation method in the above embodiments, can solve the technical problem of how to improve the reliability of manhole cover operation and maintenance schemes. Compared with the prior art, the beneficial effects of the power manhole cover operation and maintenance scheme generation device provided in this application are the same as the beneficial effects of the power manhole cover operation and maintenance scheme generation method provided in the above embodiments, and other technical features in the power manhole cover operation and maintenance scheme generation device are the same as those disclosed in the previous embodiment method, and will not be repeated here.

[0127] It should be understood that the various parts disclosed in this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0128] 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.

[0129] This application provides a computer-readable storage medium having computer-readable program instructions (i.e., a computer program) stored thereon, which are used to execute the operation and maintenance scheme generation method for power manhole covers in the above embodiments.

[0130] The computer-readable storage medium provided in this application may be, for example, a USB flash drive, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatuses, or devices, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), or flash memory, optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. The program code contained on the computer-readable storage medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, radio frequency (RF), etc., or any suitable combination thereof.

[0131] The aforementioned computer-readable storage medium may be included in the power manhole cover operation and maintenance solution generation device; or it may exist independently and not be assembled into the power manhole cover operation and maintenance solution generation device.

[0132] The aforementioned computer-readable storage medium carries one or more programs that, when executed by the power manhole cover maintenance solution generation device, enable the power manhole cover maintenance solution generation device to write computer program code for performing the operations of this application in one or more programming languages ​​or a combination thereof. The programming languages ​​include object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as C or similar languages. The program code can be executed entirely on the user's computer, partially on the user's computer, or as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0133] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation that may be implemented in systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0134] The modules described in the embodiments of this application can be implemented in software or hardware. The names of the modules do not necessarily limit the functionality of the unit itself.

[0135] The readable storage medium provided in this application is a computer-readable storage medium that stores computer-readable program instructions (i.e., a computer program) for executing the above-described method for generating an operation and maintenance plan for power manhole covers. This method can solve the technical problem of how to improve the reliability of manhole cover operation and maintenance plans. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in this application are the same as those of the power manhole cover operation and maintenance plan generation method provided in the above embodiments, and will not be repeated here.

[0136] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the above-described method for generating an operation and maintenance plan for power manhole covers.

[0137] The computer program product provided in this application can solve the technical problem of how to improve the reliability of manhole cover operation and maintenance schemes. Compared with the prior art, the beneficial effects of the computer program product provided in this application are the same as those of the power manhole cover operation and maintenance scheme generation method provided in the above embodiments, and will not be repeated here.

[0138] The above description is only a part of the embodiments of this application and does not limit the patent scope of this application. All equivalent structural transformations made under the technical concept of this application and using the contents of the specification and drawings of this application, or direct / indirect applications in other related technical fields, are included in the patent protection scope of this application.

Claims

1. A method for generating an operation and maintenance plan for power manhole covers, characterized in that, The method for generating the operation and maintenance plan for the power manhole cover includes: The system acquires multidimensional status data of manhole covers in each jurisdiction collected by the regional gateway, and constructs input features based on the multidimensional status data. Based on the geographical coordinates of the manhole covers and the topology of the cable channels, a regional manhole cover relationship map is constructed, and an adjacency matrix is ​​determined based on the regional manhole cover relationship map. The input features and the adjacency matrix are input into a graph temporal neural network model, and the health trend and failure probability of each manhole cover are determined by the graph temporal neural network model. The urgency score of each manhole cover is determined based on the health status change trend and the failure probability; Based on the weighted connection weights in the adjacency matrix, a risk propagation model is constructed, and the collaborative urgency of each manhole cover is determined according to the urgency score and the risk propagation model. Using the collaborative urgency, the adjacency matrix, the driving distance between the manhole covers, and the maintenance resource constraints as input parameters for the preset path planning algorithm, the first round of inspection path planning is carried out in the geographical road network to generate an initial inspection sequence containing at least one manhole cover to be visited. The maintenance operation of the manhole covers in the initial inspection sequence is simulated sequentially. Based on the scores of each dimension in the health score of the manhole cover under maintenance, the spare parts demand type of the manhole cover under maintenance is determined, and the collaborative urgency of the remaining manhole covers in the geographical road network is updated according to the adjacency matrix. When the updated collaborative urgency meets the replanning condition, based on the updated collaborative urgency, the current maintenance location and the remaining manhole covers, the inspection path is replanned to generate the subsequent inspection sequence, and the maintenance operation of the manhole covers in the subsequent inspection sequence is simulated sequentially until the maintenance resource constraint is reached, and the inspection route is output. Based on the spare parts requirement types of all manhole covers performing simulated maintenance operations, a spare parts list is generated, and based on the spatial distribution of the collaborative urgency of each manhole cover and the adjacency matrix, the preset location and quantity of spare parts are determined by clustering. Based on the spare parts list, the preset locations of the spare parts, and the quantity of the spare parts, a spare parts scheduling strategy is generated; The step of determining the adjacency matrix based on the regional manhole cover relationship diagram includes: The initial connection weight of the edges is determined based on the type of edges between nodes in the regional manhole cover relationship diagram. Based on the time sequence data of the status of manhole covers corresponding to connected nodes within a preset time window, the dynamic correlation coefficient between the manhole covers corresponding to the connected nodes is determined. Based on the dynamic correlation coefficient, the initial connection weights of the edges between the connected nodes are adjusted to obtain the weighted connection weights; A weight matrix is ​​constructed based on the weighted connection weights, and the weight matrix is ​​normalized based on a preset weight threshold to generate the adjacency matrix. The step of determining the dynamic correlation coefficient between manhole covers corresponding to connected nodes based on the state time-series data of manhole covers corresponding to connected nodes within a preset time window includes: Based on the temporal characteristics of the state time series data, multiple event segments corresponding to the manhole cover are extracted from the state time series data; Based on the state time-series data corresponding to each event segment, an event feature sequence is constructed; Based on the event feature sequence, the morphological similarity between event segments of the manhole covers corresponding to the connected nodes is determined within the preset time window; Based on the start time of the event segment, the event delay relationship between the manhole covers corresponding to the connected nodes is determined; Based on the morphological similarity and the event delay relationship, the dynamic correlation coefficient between the manhole covers corresponding to the connected nodes is determined.

2. The method for generating an operation and maintenance plan for power manhole covers as described in claim 1, characterized in that, The step of constructing a regional manhole cover relationship map based on the geographical coordinates of the manhole covers and the cable channel topology includes: Each of the manhole covers is designated as a node in the manhole cover relationship diagram of the region; The distance between each manhole cover is determined based on its geographical coordinates. When the distance is less than a preset distance threshold, a topological adjacent edge is constructed between the nodes corresponding to the manhole covers. Based on the geographic coordinates of the manhole cover, determine the regional environment type to which the manhole cover belongs, and construct environmental similarity edges between nodes corresponding to manhole covers with the same regional environment type; Based on the cable channel topology, functional association edges are constructed between nodes corresponding to manhole covers that have channel topology dependencies.

3. The method for generating an operation and maintenance plan for power manhole covers as described in claim 1, characterized in that, The weighted connection weights are used to quantify the state association strength between manhole covers; the collaborative urgency of a manhole cover is the weighted sum of its own urgency score and the urgency scores of other associated manhole covers.

4. The method for generating an operation and maintenance plan for power manhole covers as described in claim 1, characterized in that, The step of determining the pre-positioned location and quantity of spare parts through clustering based on the coordinated emergency spatial distribution of each manhole cover and the adjacency matrix includes: A weighted graph is constructed using the collaborative urgency of each manhole cover as the node weight and the adjacency matrix as the connection relationship between nodes. Based on the geographical location of the pre-installed parts inventory, spatial clustering is performed on the weighted map to divide each manhole cover node into at least one high-risk area; Determine the geographical center of the high-risk area as the pre-positioned location of the spare parts corresponding to the high-risk area; Based on the total collaborative urgency value of manhole covers in the high-risk area corresponding to each of the aforementioned spare parts preset locations, the number of manhole covers, and the historical replacement frequency of similar spare parts, the quantity of spare parts corresponding to each of the aforementioned spare parts preset locations is determined.

5. The method for generating an operation and maintenance plan for power manhole covers as described in claim 1, characterized in that, The replanning conditions include one or more of the following: The ratio of the decrease in the collaborative urgency of any of the remaining manhole covers to the decrease in the collaborative urgency of other remaining manhole covers is greater than a preset proportional threshold. The collaborative urgency level of any of the remaining manhole covers increases; The increase in the collaborative urgency level of any of the remaining manhole covers exceeds a preset threshold.

6. A device for generating operation and maintenance solutions for power manhole covers, characterized in that, The device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the method for generating an operation and maintenance scheme for power manhole covers as described in any one of claims 1 to 5.

7. A storage medium, characterized in that, The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, it implements the steps of the method for generating an operation and maintenance scheme for power manhole covers as described in any one of claims 1 to 5.

Images