Online monitoring operation and maintenance management method and system applied to sensor network

Abstract

The invention provides an online monitoring operation and maintenance management method and system applied to a sensor network, and the method comprises the steps: firstly linking the real-time operation information of each monitoring node of the sensor network with the historical operation and maintenance evolution data of the network, and generating a state evolution trajectory through a pre-trained AI correlation model; then starting an operation and maintenance demand intelligent deduction process based on the state evolution trajectory, and forming an operation and maintenance demand evolution graph through an AI deduction model; calling an operation and maintenance scheme evolution library, and generating an initial operation and maintenance scheme through an AI adaptation algorithm; inputting the state evolution trajectory and the operation and maintenance demand evolution graph into an AI evolution model, and performing adaptive evolution adjustment on the initial operation and maintenance scheme to obtain a target operation and maintenance scheme; and finally, synchronizing the target operation and maintenance scheme to an operation and maintenance execution end, and inputting the target operation and maintenance scheme into an operation and maintenance file module. According to the invention, intelligentization and dynamics of online monitoring operation and maintenance management of the sensor network are realized, the operation and maintenance accuracy and pertinence are improved, and the operation stability and reliability of the sensor network are enhanced.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence, and more specifically, to an online monitoring, operation and maintenance management method and system for sensor networks. Background Technology

[0002] In an era of widespread sensor network applications, their operational stability and efficiency are crucial for the normal operation of numerous fields. Currently, the monitoring and maintenance management of sensor networks mainly rely on traditional methods. These methods typically employ periodic inspections and preset threshold alarms. Periodic inspections require significant manpower and resources and struggle to capture the dynamic changes of each monitoring node in the sensor network in real time. Preset threshold alarms are too rigid and cannot be flexibly adjusted according to the dynamic evolution of sensor node states. When sensor node states are undergoing slow changes or complex fluctuations, traditional methods struggle to accurately determine their true operational status and predict potential maintenance needs. This results in maintenance work often lagging behind actual node state changes, failing to address node operational problems promptly and effectively, and consequently affecting the overall performance and reliability of the sensor network. Therefore, a more intelligent, dynamic, and efficient online monitoring and maintenance management method is urgently needed. Summary of the Invention

[0003] In view of the aforementioned problems, and in conjunction with the first aspect of the present invention, embodiments of the present invention provide an online monitoring and maintenance management method for sensor networks, the method comprising: By linking the real-time operation information of each monitoring node in the sensor network with the network's historical operation and maintenance evolution data, and using a pre-trained AI correlation model to perform dynamic evolution correlation processing of node states, the state evolution trajectory corresponding to each monitoring node is generated. The intelligent simulation process for operation and maintenance requirements is initiated based on the state evolution trajectory of each monitoring node. The AI ​​simulation model is used to mine the correlation evolution features in the state evolution trajectory and form an evolution map of operation and maintenance requirements corresponding to each monitoring node. The system retrieves a pre-defined operation and maintenance solution evolution library and uses an AI adaptation algorithm to dynamically associate the operation and maintenance requirement evolution map of each monitoring node with the solution evolution logic in the operation and maintenance solution evolution library to generate an initial operation and maintenance solution. The state evolution trajectory of each monitoring node and the evolution map of operation and maintenance requirements are input into the AI ​​evolution model to adaptively adjust the initial operation and maintenance plan and obtain the target operation and maintenance plan that is adapted to the node state in real time. The target operation and maintenance plan is synchronized to the corresponding operation and maintenance execution terminal, and the state evolution trajectory operation and maintenance requirement evolution map and target operation and maintenance plan are entered into the operation and maintenance file module of the sensor network.

[0004] In another aspect, embodiments of the present invention also provide an online monitoring and maintenance management system for sensor networks, including a processor and a machine-readable storage medium connected to the processor. The machine-readable storage medium is used to store programs, instructions, or code, and the processor is used to execute the programs, instructions, or code in the machine-readable storage medium to implement the above-described method.

[0005] Based on the above, this invention, by linking the real-time operational information of each monitoring node in the sensor network with historical network operation and maintenance evolution data, and using a pre-trained AI correlation model to generate the state evolution trajectory corresponding to each monitoring node, can comprehensively and accurately capture the dynamic changes in node states. Based on the state evolution trajectory, an intelligent operation and maintenance requirement inference process is initiated. Using the AI ​​inference model, an operation and maintenance requirement evolution map is formed, which can deeply explore the operation and maintenance requirements behind node state changes, achieving accurate prediction of operation and maintenance requirements. Through an AI adaptation algorithm, the operation and maintenance requirement evolution map is dynamically correlated with the operation and maintenance solution evolution library to generate an initial operation and maintenance solution. Then, the AI ​​evolution model is used to adaptively evolve and adjust the initial operation and maintenance solution to obtain a target operation and maintenance solution that is adapted to the node state in real time, greatly improving the accuracy and relevance of the operation and maintenance solution. Finally, the target operation and maintenance solution is synchronized to the operation and maintenance execution end and entered into the operation and maintenance file module, realizing the intelligent, dynamic, and efficient operation and maintenance management of the sensor network, effectively improving the operational stability and reliability of the sensor network. Attached Figure Description

[0006] Figure 1 This is a schematic diagram of the execution flow of the online monitoring and maintenance management method for sensor networks provided in this embodiment of the invention.

[0007] Figure 2 This is a schematic diagram of exemplary hardware and software components of an online monitoring and maintenance management system for sensor networks provided in an embodiment of the present invention. Detailed Implementation

[0008] The present invention will now be described in detail with reference to the accompanying drawings. Figure 1 This is a flowchart illustrating an online monitoring and maintenance management method for sensor networks according to an embodiment of the present invention. The following is a detailed description of this online monitoring and maintenance management method for sensor networks.

[0009] Step S110: Combine the real-time operation information of each monitoring node in the linkage sensor network with the network's historical operation and maintenance evolution data, and use a pre-trained AI correlation model to perform dynamic evolution correlation processing of node states, generating the state evolution trajectory corresponding to each monitoring node. This embodiment applies to the online monitoring, operation and maintenance management of an urban water supply network pressure monitoring sensor network. The monitoring sensor network includes pressure monitoring nodes deployed in different areas of the city. Each node is used to collect network pressure data and upload it to the management platform.

[0010] Step S111: Collect real-time operating information of each monitoring node in the sensor network. The real-time operating information includes node signal transmission information, device operating status information, data acquisition and feedback information.

[0011] Node signal transmission information includes signal strength, transmission delay, and packet loss rate; equipment operating status information includes power supply voltage, equipment temperature, and hardware module operating status; data acquisition feedback information includes collected pressure data, data acquisition frequency, and data upload success rate. The real-time operating information of each monitoring node exists in multi-dimensional data format. For example, a node's signal strength is within a certain range, transmission delay is within a certain range, packet loss rate is within a certain percentage range, power supply voltage is within a certain voltage range, equipment temperature is within a certain temperature range, pressure data is within a certain pressure range, data acquisition frequency is within a certain time interval range, and data upload success rate is within a certain percentage range.

[0012] Step S112: Retrieve the historical operation and maintenance evolution data of the sensor network. The historical operation and maintenance evolution data includes the operation and maintenance operation information and post-operation and maintenance recovery information corresponding to the past state change information of each monitoring node.

[0013] Past status change information includes changes in signal strength, pressure data, and equipment temperature of the node over a past period. Corresponding maintenance operation information includes actions taken to address these changes, such as replacing hardware modules, adjusting power supply parameters, and optimizing signal transmission paths. Post-maintenance status recovery information includes the recovery status of signal strength, pressure data acquisition, and equipment temperature after the maintenance operations. For example, in the historical maintenance evolution data of a certain node, past status change information records that the node's signal strength gradually decreased from a high range to a low range, pressure data fluctuated from a stable range, and equipment temperature rose from a normal range to an abnormal range over a past period. The corresponding maintenance operation information records the operation of replacing the signal transmission module for that node. Post-maintenance status recovery information records that after the module replacement, signal strength recovered to a high range, pressure data stabilized, and equipment temperature returned to a normal range.

[0014] Step S113: Classify and associate real-time operation information with historical network operation and maintenance evolution data according to monitoring node identifiers to form an information association set corresponding to each monitoring node. Each monitoring node has a unique identifier, such as a node number. Based on this identifier, real-time operation information and historical network operation and maintenance evolution data belonging to the same node are associated. For example, if node A is identified by a specific number, the real-time operation information of the node corresponding to that number is associated with its historical operation and maintenance evolution data to form an information association set for node A. This information association set contains all data in node A's real-time operation information and all data in its historical operation and maintenance evolution data.

[0015] Step S114: Input each information association set into the feature input layer of the pre-trained AI association model, and extract the dynamic evolution features in each information association set through the built-in feature evolution extraction unit of the model.

[0016] The pre-trained AI association model is a deep learning model. The feature input layer receives data from the information association set, and the feature evolution extraction unit contains multiple convolutional layers and recurrent layers. The feature input layer transforms the multi-dimensional data in the information association set into a vector form that the model can process, and inputs it into the feature evolution extraction unit. The convolutional layers of the feature evolution extraction unit process the vectorized data to extract local features, and the recurrent layers perform temporal processing on the convolutional features to extract dynamic evolution features. For example, after the information association set of node A is input into the feature input layer, it is transformed into a high-dimensional vector. This high-dimensional vector contains all the features of node A's real-time operation information and historical operation and maintenance evolution data. After processing by the convolutional layers, local features such as signal intensity changes and pressure data fluctuations are extracted. Then, after processing by the recurrent layers, the dynamic evolution features of these local features over time are extracted.

[0017] Step S1141: The feature evolution extraction unit of the AI ​​association model performs multi-dimensional feature capture on the real-time running information in the information association set, and extracts the instantaneous state features in the real-time running information.

[0018] Multi-dimensional feature capture extracts features from each dimension of data in real-time operation information. For example, it extracts the current numerical and change features of signal strength, the current numerical and fluctuation features of pressure data, and the current numerical and change features of equipment temperature. Real-time state features are the set of features extracted from the current state of the real-time operation information.

[0019] Step S1142: Divide the network historical operation and maintenance evolution data in the information association set into stages, and extract the historical state characteristics of each stage and the state transition characteristics between stages.

[0020] The phase division is based on the state changes in historical operation and maintenance evolution data, which are divided into different phases. For example, based on the changes in the signal strength of nodes, historical data can be divided into signal strength stabilization phase, signal strength decline phase, and signal strength recovery phase. The historical state characteristics of each phase are the state characteristics of the nodes within that phase, such as the signal strength value range and pressure data value range in the signal strength stabilization phase. The state transition characteristics between phases are the characteristics of the state changes between different phases, such as the rate of change and the magnitude of change from the signal strength stabilization phase to the signal strength decline phase.

[0021] Step S1143: Align the instantaneous state features with the historical state features in time sequence to explore the similar evolutionary patterns between the instantaneous state features and the historical state features.

[0022] Time alignment matches the time point corresponding to the instantaneous state feature with the time point corresponding to the historical state feature. For example, it compares the instantaneous state feature at the current time point with the historical state feature within the same time interval in the historical data to explore the similarity between the two. For example, the signal strength change trend in the current instantaneous state feature is similar to the signal strength change trend in a certain stage of the historical state feature, and the pressure data fluctuation is similar to the pressure data fluctuation in a certain stage of the historical state feature.

[0023] Step S1144: Combine state transition features to analyze the potential evolution direction of the instantaneous state features, and generate a dynamic evolution feature set that includes instantaneous state features, historical state features, and potential evolution directions.

[0024] Step S11441: Classify the state transition features according to the evolution rate and sort out the state change patterns under different evolution rates.

[0025] Evolution rate refers to the speed at which state changes. For example, state transition characteristics can be divided into three categories: rapid evolution, medium-speed evolution, and slow evolution. Rapid evolution corresponds to a state change pattern in which the state changes significantly in a short period of time. Medium-speed evolution corresponds to a state change pattern in which the state changes moderately in a medium period of time. Slow evolution corresponds to a state change pattern in which the state changes slightly in a long period of time.

[0026] Step S11442: Associate the instantaneous state features with various state change patterns to identify the target change pattern corresponding to the instantaneous state features.

[0027] For example, the signal intensity change rate in instantaneous state features is relatively fast, which is associated with rapidly evolving state change patterns, thus identifying the target change pattern corresponding to the instantaneous state features as a rapidly evolving pattern.

[0028] Step S11443: Based on the target change pattern, deduce the possible state values ​​of the instantaneous state features in subsequent time periods and determine the potential evolution direction of the instantaneous state features.

[0029] For example, if the target change pattern is a rapid evolution pattern, and the signal strength in the historical data shows that under this rapid evolution pattern, the signal strength may drop rapidly from the current range to a lower range, then the signal strength in the inferred instantaneous state characteristics may take a lower value range in subsequent periods, and the potential evolution direction is determined to be a rapid decrease in signal strength.

[0030] Step S11444: Classify and integrate the instantaneous state features, historical state features, and potential evolution directions according to feature types, and label the correlation weights between each feature.

[0031] Feature types include signal features, pressure features, temperature features, etc. Features belonging to the same type are integrated together. For example, the instantaneous state features, historical state features, and potential evolution direction of signal strength are integrated into a signal feature class. The correlation weight is marked according to the correlation between each feature. For example, the instantaneous state features of signal strength are highly correlated with the historical state features and are marked with a higher correlation weight. The instantaneous state features of signal strength are highly correlated with the potential evolution direction and are marked with a higher correlation weight.

[0032] Step S11445: Arrange the classified features in an orderly manner according to the association weights to form a dynamic evolution feature set that includes immediate state features, historical state features, and potential evolution directions.

[0033] The classified features are arranged in descending order of their association weights. For example, features with the highest association weights in the signal feature class are placed at the front, and features with lower association weights are placed at the back, forming a dynamically evolving feature set.

[0034] Step S1145: Perform feature filtering on the dynamic evolution feature set, retain the core features that are directly related to the evolution of node states, and form the dynamic evolution features corresponding to each information association set.

[0035] Feature selection is performed using a built-in feature selection algorithm. This algorithm filters features based on their correlation with node state evolution, retaining core features with high correlation and removing non-core features with low correlation. For example, in the dynamic evolution feature set, features such as signal strength variation, pressure data fluctuation, and equipment temperature variation are directly related to node state evolution and are retained as core features, forming the dynamic evolution features corresponding to each information association set.

[0036] Step S115: By using the trajectory construction unit of the AI ​​association model, the dynamic evolution characteristics are associated with the state change patterns in the historical network operation and maintenance evolution data to generate a state evolution trajectory that can reflect the trend of node state changes.

[0037] The trajectory construction unit comprises a fully connected layer and an output layer. The fully connected layer correlates dynamic evolution characteristics with state change patterns in historical network operation and maintenance data, while the output layer generates the state evolution trajectory. State change patterns are the patterns of node state changes over time extracted from historical network operation and maintenance data. For example, historical data shows that a decrease in node signal strength is followed by fluctuations in pressure data, and an increase in equipment temperature is followed by signal transmission anomalies. The trajectory construction unit correlates dynamic evolution characteristics with these state change patterns to generate a state evolution trajectory. This trajectory exists in the form of time-series data, reflecting the state characteristics and trends of nodes at different points in time. For example, the state evolution trajectory of node A records the signal strength characteristics, pressure data characteristics, and equipment temperature characteristics of node A at the current time point, as well as the trends of signal strength change, pressure data change, and equipment temperature change at subsequent time points.

[0038] Step S120: Based on the state evolution trajectory of each monitoring node, start the intelligent simulation process of operation and maintenance requirements, and use the AI ​​simulation model to mine the correlation evolution features in the state evolution trajectory to form the operation and maintenance requirements evolution map corresponding to each monitoring node.

[0039] Step S121: Based on the state evolution trajectory of each monitoring node, start the intelligent simulation process of operation and maintenance requirements, load the pre-trained AI simulation model and complete the model operation configuration.

[0040] The intelligent simulation process for operational and maintenance requirements is automatically initiated by the management platform, loading a pre-trained AI simulation model. This AI simulation model is a deep learning model, containing trajectory parsing units, feature association units, requirement mapping units, and graph construction units. Model runtime configuration includes setting the model's input and output parameters, runtime environment parameters, etc., to ensure that the model can run normally.

[0041] Step S122: The trajectory of each monitoring node is segmented and decomposed by the trajectory analysis unit of the AI ​​inference model, and the key state features of each evolution stage in the trajectory are extracted.

[0042] The trajectory analysis unit includes a time-series segmentation module and a feature extraction module. The time-series segmentation module divides the trajectory into different evolution stages based on the state change points in the state evolution trajectory. For example, based on the point of decrease in signal strength, the point of fluctuation in pressure data, and the point of increase in equipment temperature, the trajectory is divided into an initial stable stage, a state change stage, and an abnormal stage. The feature extraction module extracts key state features from each evolution stage. For example, the key state features of the initial stable stage are stable signal strength, stable pressure data, and normal equipment temperature; the key state features of the state change stage are decreasing signal strength, fluctuating pressure data, and increasing equipment temperature; and the key state features of the abnormal stage are abnormal signal strength, abnormal pressure data, and abnormal equipment temperature.

[0043] Step S123: Mine the evolutionary relationship between key state features through the feature association unit of the AI ​​inference model, and sort out the evolutionary logic of different key state features.

[0044] The feature association unit comprises an attention mechanism layer and an association processing layer. The attention mechanism layer assigns weights to key state features, while the association processing layer mines the evolutionary relationships between these key state features. Evolutionary relationships refer to causal or temporal relationships between different key state features; for example, a decrease in signal strength leads to fluctuations in pressure data, and an increase in equipment temperature leads to signal transmission anomalies. Sequential evolutionary logic refers to the order in which different key state features appear; for example, a decrease in signal strength precedes fluctuations in pressure data, and an increase in equipment temperature precedes signal transmission anomalies. The feature association unit processes key state features to uncover these evolutionary relationships and sequential evolutionary logic.

[0045] Step S124: The evolutionary relationship is associated with the preset operation and maintenance requirement feature library through the requirement mapping unit of the AI ​​inference model, and the basic operation and maintenance requirements corresponding to each evolution stage are matched.

[0046] The demand mapping unit comprises a mapping layer and a matching layer. The mapping layer transforms evolutionary relationships into feature vectors that the model can process, and the matching layer matches these feature vectors with feature vectors in the operation and maintenance (O&M) demand feature library. The O&M demand feature library stores basic O&M demands corresponding to different evolutionary relationships. For example, a decrease in signal strength corresponds to the basic O&M demand of checking the signal transmission module, fluctuations in pressure data correspond to the basic O&M demand of checking the pressure sensor, and an increase in equipment temperature corresponds to the basic O&M demand of checking the heat dissipation module. The demand mapping unit matches the evolutionary relationships with these basic O&M demands to determine the basic O&M demands corresponding to each evolutionary stage. For example, in the state change stage, the evolutionary relationship is a decrease in signal strength leading to fluctuations in pressure data, and the matched basic O&M demands are checking the signal transmission module and checking the pressure sensor; in the anomaly stage, the evolutionary relationship is an increase in equipment temperature leading to signal transmission anomalies, and the matched basic O&M demands are checking the heat dissipation module and checking the signal transmission module.

[0047] Step S125: The basic operation and maintenance requirements of each evolution stage are connected in series according to the evolution logic through the graph construction unit of the AI ​​inference model, and the change trend of key state characteristics is combined to form an evolution graph of operation and maintenance requirements.

[0048] Step S1251: Determine the time sequence of each evolution stage through the graph construction unit of the AI ​​inference model, and sort the basic operation and maintenance requirements in order.

[0049] The chronological order is determined by the evolutionary stages divided by the trajectory analysis unit. For example, the initial stable stage comes first, the state change stage comes in the middle, and the abnormal stage comes last. The basic operation and maintenance requirements are initially sorted according to this order, with the basic operation and maintenance requirements of the initial stable stage coming first, the basic operation and maintenance requirements of the state change stage coming in the middle, and the basic operation and maintenance requirements of the abnormal stage coming last.

[0050] Step S1252: Extract the key state feature thresholds corresponding to each basic operation and maintenance requirement, and mark the state conditions triggered by each basic operation and maintenance requirement.

[0051] Key state characteristic thresholds are critical values ​​extracted from historical network operation and maintenance evolution data that trigger basic operation and maintenance requirements. For example, signal strength below a certain value triggers a basic operation and maintenance requirement to check the signal transmission module; pressure data fluctuation exceeding a certain range triggers a basic operation and maintenance requirement to check the pressure sensor; and equipment temperature above a certain value triggers a basic operation and maintenance requirement to check the heat dissipation module. The state conditions triggering each basic operation and maintenance requirement are labeled. For example, the state condition for checking the signal transmission module is that the signal strength is below a specific threshold; the state condition for checking the pressure sensor is that the pressure data fluctuation exceeds a specific range; and the state condition for checking the heat dissipation module is that the equipment temperature exceeds a specific threshold.

[0052] Step S1253: Analyze the relationship between basic operation and maintenance requirements in adjacent evolution stages, and determine the logic of the impact of operation and maintenance requirements in the previous stage on operation and maintenance requirements in the next stage.

[0053] The correlation analysis is performed using a built-in correlation analysis algorithm within the model. This algorithm analyzes the relationships between basic maintenance requirements at different stages in historical maintenance data. For example, if a maintenance requirement for the signal transmission module in a previous stage is not addressed promptly, it can lead to signal transmission anomalies in the subsequent stage, triggering maintenance requirements for both the signal transmission module and the pressure sensor. Similarly, if a maintenance requirement for the heat dissipation module in a previous stage is not addressed promptly, it can cause a continuous rise in equipment temperature in the subsequent stage, triggering maintenance requirements for both the heat dissipation module and its replacement. The algorithm determines the logic of how maintenance requirements from previous stages affect those in subsequent stages. For instance, if basic maintenance requirements from a previous stage are addressed, subsequent basic maintenance requirements may not be triggered; however, if basic maintenance requirements from a previous stage are not addressed, subsequent basic maintenance requirements will be triggered and their level will increase.

[0054] Step S1254: Based on the changing trends of key status characteristics, predict the duration and evolution / upgrade path of each basic operation and maintenance requirement.

[0055] The changing trends of key state characteristics are extracted from the state evolution trajectory, such as a continuous decreasing trend in signal strength, a continuous fluctuating trend in pressure data, and a continuous increasing trend in equipment temperature. Based on these trends, the duration of basic maintenance requirements is predicted. For example, under a continuous decreasing signal strength trend, the duration of basic maintenance requirements for inspecting the signal transmission module is within a certain time range; under a continuous fluctuating pressure data trend, the duration of basic maintenance requirements for inspecting the pressure sensor is within a certain time range. The evolution and upgrade path refers to the evolution path of basic maintenance requirements from low-level to high-level. For example, if the basic maintenance requirement for inspecting the signal transmission module is not addressed in a timely manner, it will escalate to the maintenance requirement of replacing the signal transmission module; if the basic maintenance requirement for inspecting the heat dissipation module is not addressed in a timely manner, it will escalate to the maintenance requirement of replacing the heat dissipation module.

[0056] Step S1255: Integrate the sorted basic operation and maintenance requirements status conditions associations and evolution upgrade paths into structured graph data. Basic operation and maintenance requirements are represented by graph nodes, and associations and evolution paths are represented by edges, forming an operation and maintenance requirements evolution graph.

[0057] Structured graph data exists in the form of nodes and edges. Graph nodes represent basic maintenance requirements, and each node contains information such as the name of the basic maintenance requirement, the triggering state condition, and the duration. Edges represent the relationships and evolution paths between basic maintenance requirements in adjacent evolution stages, and each edge contains information such as the type of relationship and the direction of the evolution path. For example, in the maintenance requirement evolution graph, the basic maintenance requirement in the initial stable stage is to check the equipment status. The corresponding graph node contains the name of the equipment status to be checked, the triggering state condition is that the equipment status is normal, and the duration is a certain time range. The basic maintenance requirements in the state change stage are to check the signal transmission module and check the pressure sensor. The corresponding graph node contains the names of the signal transmission module and pressure sensor to be checked, the triggering state conditions are that the signal strength decreases and the pressure data fluctuates, and the duration is a certain time range. The basic maintenance requirements in the abnormal stage are to replace the signal transmission module and replace the pressure sensor. The corresponding graph node contains the names of the signal transmission module and pressure sensor to be replaced, the triggering state conditions are that the signal strength is abnormal and the pressure data is abnormal, and the duration is a certain time range. Connecting a node in the initial stable phase to a node in the state change phase indicates that unresolved maintenance requirements for checking the equipment status in the previous phase will trigger maintenance requirements for checking the signal transmission module and pressure sensor in the next phase. Connecting a node in the state change phase to a node in the abnormal phase indicates that unresolved maintenance requirements for checking the signal transmission module and pressure sensor in the previous phase will trigger maintenance requirements for replacing the signal transmission module and pressure sensor in the next phase.

[0058] Step S130: Retrieve the preset operation and maintenance solution evolution library, and dynamically associate the operation and maintenance requirement evolution map of each monitoring node with the solution evolution logic in the operation and maintenance solution evolution library using an AI adaptation algorithm to generate an initial operation and maintenance solution. Step S131: Retrieve the preset operation and maintenance scheme evolution library. The operation and maintenance scheme evolution library stores the scheme evolution logic corresponding to various basic operation and maintenance schemes. The scheme evolution logic includes the adjustment rules, evolution trigger conditions, and adaptation state range of the scheme module.

[0059] The basic operation and maintenance (O&M) plan includes plans for checking equipment status, checking signal transmission modules, checking pressure sensors, replacing signal transmission modules, and replacing pressure sensors. The adjustment rules for these plan modules refer to the rules for adjusting modules such as operation procedures, tool adaptation, execution order, and parameter configuration within the basic O&M plan. For example, the operation procedure module in the signal transmission module check plan can add a signal strength test step, and the tool adaptation module can be replaced with a more precise testing tool. Evolution trigger conditions refer to the conditions that trigger the plan's evolution. For example, a continuous decrease in signal strength for more than a certain period will trigger the signal transmission module check plan to evolve into a signal transmission module replacement plan. The adaptation state range refers to the range of node states to which the plan's evolution logic applies. For example, the adaptation state range for the signal transmission module check plan is when the signal strength is within a certain range, while the adaptation state range for the signal transmission module replacement plan is when the signal strength is at an even lower range.

[0060] Step S132: Use AI adaptation algorithm to encode the features of the operation and maintenance requirement evolution map and generate a requirement encoding vector that can reflect the evolution law of requirements.

[0061] The AI ​​adaptation algorithm comprises a feature extraction layer and an encoding layer. The feature extraction layer extracts the evolutionary features of maintenance requirements from the maintenance requirement evolution graph, such as the type of basic maintenance requirements, triggering state conditions, correlations, and evolutionary upgrade paths. The encoding layer transforms these evolutionary features into requirement encoding vectors, which are high-dimensional vectors reflecting the evolutionary patterns of requirements. For example, the maintenance requirement evolution graph may contain basic maintenance requirements for inspecting signal transmission modules and pressure sensors. The triggering state conditions are signal strength decline and pressure data fluctuation. The correlation is that the current requirement is triggered because the previous stage's unresolved equipment status inspection requirement was not addressed. The evolutionary upgrade path is that the current unresolved requirement will escalate to the requirement of replacing the signal transmission module and pressure sensor. The feature extraction layer extracts these features, and the encoding layer transforms them into requirement encoding vectors.

[0062] Step S133: Perform feature encoding on the evolution logic of each scheme in the operation and maintenance scheme evolution library to generate a scheme encoding vector corresponding to each scheme evolution logic.

[0063] Feature encoding is performed using the same feature extraction and encoding layers as in step S132. The feature extraction layer extracts scheme evolution features from the scheme evolution logic, such as adjustment rules for scheme modules, evolution triggering conditions, and adaptation state ranges. The encoding layer transforms these scheme evolution features into a scheme encoding vector, which is a high-dimensional vector reflecting the scheme evolution logic. For example, the scheme evolution logic might evolve from checking the signal transmission module to replacing the signal transmission module. The adjustment rules for the scheme modules might be replacing the checking step in the operation process with a replacement step, replacing the tool adaptation module with a tool replacement, the evolution triggering condition being a continuous decrease in signal strength for a certain period of time, and the adaptation state range being a lower signal strength range. The feature extraction layer extracts these features, and the encoding layer transforms them into a scheme encoding vector.

[0064] Step S134: Dynamically associate the demand encoding vector with the encoding vectors of each solution to discover the points of convergence between the demand evolution pattern and the solution evolution logic.

[0065] Dynamic association is performed using a built-in similarity calculation algorithm. This algorithm calculates the similarity between the requirement encoding vector and the encoding vectors of each solution. Solution encoding vectors with higher similarity have a higher degree of fit between their evolution logic and the requirement evolution pattern. A fit point refers to the same or similar features in the requirement evolution pattern and the solution evolution logic. For example, the evolution upgrade path in the requirement evolution pattern might be an upgrade from checking the signal transmission module requirement to replacing the signal transmission module requirement; the evolution trigger condition in the solution evolution logic might be an evolution from checking the signal transmission module solution to replacing the signal transmission module solution. These two features constitute a fit point.

[0066] Step S135: Select the optimal matching solution evolution logic based on the matching point, and generate an initial operation and maintenance solution that adapts to the current operation and maintenance requirements evolution map based on the solution evolution logic.

[0067] The selection process utilizes a built-in algorithm within the model. This algorithm filters based on the number and similarity of matching points, selecting the evolutionary logic with the most matching points and the highest similarity as the optimal matching evolutionary logic. An initial operation and maintenance plan is then generated based on this evolutionary logic. For example, the optimal matching evolutionary logic might evolve from checking the signal transmission module to replacing it. The generated initial operation and maintenance plan includes modules such as the operation flow for checking the signal transmission module, tool adaptation, execution order, and parameter configuration, as well as adjustment rules, evolution trigger conditions, and adaptation state ranges for evolving to replacing the signal transmission module.

[0068] Step S140: Input the state evolution trajectory of each monitoring node and the evolution map of operation and maintenance requirements into the AI ​​evolution model, and make adaptive evolution adjustments to the initial operation and maintenance plan to obtain the target operation and maintenance plan that is adapted to the node state in real time.

[0069] Step S141: Input the state evolution trajectory and operation and maintenance requirement evolution map of each monitoring node into the AI ​​evolution model, and load the initial operation and maintenance plan at the same time.

[0070] The AI ​​evolution model is a deep learning model that includes an input layer, a feature processing layer, a module evolution unit, and a solution integration unit. The input layer receives data such as state evolution trajectory, operation and maintenance requirement evolution map, and initial operation and maintenance solution. The feature processing layer extracts and fuses features from the input data, such as extracting node state change trend features from the state evolution trajectory, extracting requirement evolution features from the operation and maintenance requirement evolution map, and extracting solution module features from the initial operation and maintenance solution. These features are then fused together.

[0071] Step S142: The initial operation and maintenance plan is decomposed into operation process module, tool adaptation module, execution sequence module and parameter configuration module through the solution decomposition unit of the AI ​​evolution model.

[0072] The solution decomposition unit includes a decomposition layer, which breaks down the initial operation and maintenance solution into four modules based on its structure. The operation flow module contains the operational steps in the initial operation and maintenance solution; for example, checking the signal transmission module involves opening the device casing, testing the signal strength, checking the module connection, and closing the device casing. The tool adaptation module contains the operation and maintenance tools used in the initial operation and maintenance solution, such as a signal tester for testing signal strength and a multimeter for checking module connection. The execution order module contains the execution order of the operational steps in the initial operation and maintenance solution, such as opening the device casing, testing the signal strength, checking the module connection, and closing the device casing. The parameter configuration module contains the execution parameters for each operational step in the initial operation and maintenance solution, such as the test time for testing signal strength within a certain time range and the test frequency within a certain number of times.

[0073] Step S143: The key state features in the state evolution trajectory are associated with the adaptation conditions of each module through the feature matching unit of the AI ​​evolution model, and the adaptation differences between each module and the current node state are identified.

[0074] The feature matching unit includes a matching layer that compares key state features in the state evolution trajectory with the adaptation conditions of each module to identify adaptation differences. The adaptation conditions of each module refer to the range of node states to which the module is applicable. For example, the adaptation condition for the test signal strength step in the operation flow module is that the signal strength is within a certain range; the adaptation condition for the signal tester in the tool adaptation module is that the signal strength is within a certain range; the adaptation condition for the execution order in the execution order module is that the node state is stable; and the adaptation condition for the test time in the parameter configuration module is that the rate of change of signal strength is within a certain range. For example, the key state characteristic in the state evolution trajectory is that the signal strength continuously decreases and changes rapidly. The adaptation condition for the test signal strength step in the operation process module is that the signal strength is within a certain range, but the current signal strength is below that range, resulting in an adaptation difference. Similarly, the adaptation condition for the signal tester in the tool adaptation module is that the signal strength is within a certain range, but the current signal strength is below that range, again resulting in an adaptation difference. The adaptation condition for the execution order in the execution order module is that the node state is stable, but the current node state is in a changing phase, resulting in an adaptation difference. Furthermore, the adaptation condition for the test time in the parameter configuration module is that the signal strength change rate is within a certain range, but the current signal strength change rate is too fast and exceeds that range, resulting in an adaptation difference.

[0075] Step S144: Based on the adaptation differences and the evolution trend of the evolution map of the module evolution unit of the AI ​​evolution model, the module is adjusted in a targeted manner to generate the evolved module content.

[0076] Step S1441: For the operation process module, supplement specific operation steps that can alleviate state evolution deviations based on adaptation differences, and adjust the execution timing of operation steps in combination with the evolution trend of the operation and maintenance requirement evolution map.

[0077] The adaptation discrepancy stems from a mismatch between the adaptation conditions of the test signal strength step in the operation process module and the current node status. The supplementary specific operation step is to add a signal transmission path check step, which can mitigate the state evolution deviation of continuously decreasing signal strength. The evolution trend of the operation and maintenance requirement evolution map is that unresolved current requirements will escalate into requirements to replace the signal transmission module. The execution timing of the operation step is adjusted to place the signal transmission path check step after the test signal strength step and before the module connection check step, so as to promptly identify problems with the signal transmission path.

[0078] Step S1442: For the tool adaptation module, based on the mismatch points between the tool and the node status in the adaptation differences, replace them with operation and maintenance tool types that can adapt to the current status and future evolution trends.

[0079] The compatibility discrepancy is that the compatibility conditions of the signal tester in the tool compatibility module do not match the current node status. The current node status is that the signal strength is continuously decreasing and the rate of change is relatively fast. The future evolution trend is that the signal strength will continue to decrease. The replacement maintenance tool is a more accurate signal analyzer, which can adapt to the current signal strength range and the future lower signal strength range.

[0080] Step S1443: For the execution order module, refer to the adjusted operation process module and the updated tool adaptation module, and rearrange the operation order according to the principle of optimal operation and maintenance efficiency.

[0081] The revised operation process module includes steps such as opening the device casing, testing signal strength, checking the signal transmission path, checking module connections, and closing the device casing. The updated tool adaptation module includes tools such as a signal analyzer and a multimeter. Based on the principle of optimal operation and maintenance efficiency, the operation sequence has been rearranged to: opening the device casing, using a signal analyzer to test signal strength, using a multimeter to check the signal transmission path, checking module connections, and closing the device casing, in order to improve operation and maintenance efficiency.

[0082] Step S1444: For the parameter configuration module, adjust the execution parameters of each operation step by combining the characteristic change rate in the state evolution trajectory and the upgrade path of the operation and maintenance requirement evolution map.

[0083] For example, step S14441: Extract the characteristic change rate in the state evolution trajectory and analyze the speed and magnitude of change of each key state characteristic.

[0084] The rate of change of a feature is the rate at which key state features extracted from the state evolution trajectory change over time, such as the rate of change of signal strength, the rate of change of pressure data, and the rate of change of equipment temperature. The analysis focuses on the speed and magnitude of these rates of change.

[0085] Step S14442: Analyze the upgrade path of the operation and maintenance requirements evolution map and clarify the changes in the requirements of operation parameters for operation and maintenance at each stage.

[0086] The upgrade path is the evolution path of basic operation and maintenance requirements extracted from the operation and maintenance requirement evolution map from low level to high level. For example, if the current requirement is to check the signal transmission module, the upgrade path is that if this requirement is not handled, it will be upgraded to the requirement to replace the signal transmission module. It clarifies the changes in the requirements of operation parameters for operation and maintenance requirements at each stage. For example, the requirement to check the signal transmission module requires the test signal strength parameters to be a fixed test time and a fixed test frequency, while the requirement to replace the signal transmission module requires the test signal strength parameters to be extended and the test frequency increased.

[0087] Step S14443: Correlate the feature change rate with the parameter requirement changes in the upgrade path to determine the adjustment direction of the execution parameters for each operation step.

[0088] For example, if the characteristic change rate is a fast signal strength change rate, and the parameter requirements in the upgrade path are to extend the test time and increase the test frequency, then by correlating the two, the direction of adjusting the execution parameters in the signal strength test stage is to extend the test time and increase the test frequency; the direction of adjusting the execution parameters in the signal transmission path inspection stage is to extend the inspection time and increase the inspection frequency.

[0089] Step S14444: Based on the adjustment direction and the parameter configuration experience in similar states in historical operation and maintenance data, initially set the adjustment parameter values ​​for each operation step.

[0090] Historical operation and maintenance data shows a similar state where the signal strength changes rapidly. Based on the parameter configuration experience under this state, the test time for the signal strength test is initially set to the extended time range, and the test frequency is set to the increased number of times. The check time for the signal transmission path check is also set to the extended time range, and the check frequency is set to the increased number of times.

[0091] Step S14445: Substitute the initially set adjustment parameter values ​​into the simulated operation and maintenance scenario to conduct evolution simulation and observe the impact of parameter adjustment on node state evolution.

[0092] The simulated operation and maintenance scenario is a virtual scenario built based on the current node status. The initially set adjustment parameter values ​​are substituted into this scenario to simulate the execution of operation and maintenance operations and observe the evolution of the node status. For example, after simulating the adjustment parameters of the test signal strength step, observe the changes in signal strength and whether it can alleviate the trend of continuous decline in signal strength.

[0093] Step S14446: Fine-tune the parameter values ​​based on the simulation results to form the final parameter configuration and complete the adjustment of the parameter configuration module.

[0094] If the simulation results show that the downward trend in signal strength has been alleviated, it means that the adjusted parameter values ​​are appropriate and no fine-tuning is needed. If the simulation results show that the downward trend in signal strength has not been alleviated, it means that the adjusted parameter values ​​are inappropriate and fine-tuning is needed, such as further extending the test time or increasing the test frequency, until the simulation results show that the downward trend in signal strength has been alleviated, thus forming the final parameter configuration and completing the adjustment of the parameter configuration module.

[0095] Step S1445: Perform logical verification on each adjusted module to ensure smooth connection between modules and form the evolved module content.

[0096] Logical verification is performed using a built-in verification algorithm within the model. This algorithm checks the connectivity between the adjusted operation flow module, tool adaptation module, execution order module, and parameter configuration module. For example, it checks whether the operation steps in the operation flow module match the tools in the tool adaptation module, whether the execution order in the execution order module matches the operation steps in the operation flow module, and whether the execution parameters in the parameter configuration module match the operation steps in the operation flow module. If the connectivity is smooth, the evolved module content is formed; if the connectivity is not smooth, steps S1441 to S1444 are returned for readjustment until the connectivity is smooth.

[0097] Step S145: The evolved modules are reorganized according to the operation and maintenance logic through the solution integration unit of the AI ​​evolution model to form a target operation and maintenance solution that is adapted to the node status in real time.

[0098] The solution integration unit comprises an integration layer and an output layer. The integration layer reorganizes the evolved operation flow module, tool adaptation module, execution sequence module, and parameter configuration module according to the operation and maintenance logic, which includes the rationality of the operation flow, the accuracy of tool adaptation, the efficiency of the execution sequence, and the applicability of the parameter configuration. The output layer generates a target operation and maintenance solution, which includes the content of each evolved module and is adapted to the node status in real time. For example, the operation flow module in the target operation and maintenance solution includes steps such as opening the device casing, using a signal analyzer to test the signal strength, using a multimeter to check the signal transmission path, checking the module connection status, and closing the device casing. The tool adaptation module includes tools such as a signal analyzer and a multimeter. The execution sequence module includes the execution order of the above steps. The parameter configuration module includes parameters such as the test time and test frequency for testing signal strength, and the check time and check frequency for checking the signal transmission path. This solution is adapted to the current node's continuously decreasing signal strength and rapid change rate in real time.

[0099] Step S150: Synchronize the target operation and maintenance plan to the corresponding operation and maintenance execution terminal, and at the same time, enter the state evolution trajectory operation and maintenance requirement evolution map and the target operation and maintenance plan into the operation and maintenance file module of the sensor network.

[0100] First, the target maintenance plan is synchronized to the corresponding maintenance execution terminal via network transmission. The maintenance execution terminal is a mobile terminal or maintenance device used by maintenance personnel, such as their mobile phone or tablet, or the control terminal of the maintenance device. After synchronization is complete, the maintenance execution terminal receives the target maintenance plan and displays its contents so that maintenance personnel can perform maintenance operations.

[0101] Then, the state evolution trajectory, maintenance requirement evolution map, and target maintenance plan are entered into the sensor network's maintenance file module. The maintenance file module is a database storing maintenance-related data. The data entry process is performed through the management platform's interface. The management platform converts the state evolution trajectory, maintenance requirement evolution map, and target maintenance plan into a database-storable format, such as structured data format, and then enters it into the maintenance file module. After the data entry is complete, the maintenance file module stores the state evolution trajectory, maintenance requirement evolution map, and target maintenance plan for each monitoring node for subsequent querying and analysis.

[0102] Figure 2 The illustration shows exemplary hardware and software components of an online monitoring and maintenance management system 100 for sensor networks, which can implement the ideas of this application, according to some embodiments of this application. For example, processor 120 can be used in the online monitoring and maintenance management system 100 for sensor networks and to perform the functions in this application.

[0103] The online monitoring and maintenance management system 100 for sensor networks can be a general-purpose server or a special-purpose server; both can be used to implement the online monitoring and maintenance management method for sensor networks described in this application. Although only one server is shown in this application, for convenience, the functions described in this application can be implemented in a distributed manner on multiple similar platforms to balance the load.

[0104] For example, an online monitoring and maintenance management system 100 for sensor networks may include a network port 110 connected to the network, one or more processors 120 for executing program instructions, a communication bus 130, and various forms of storage media 140, such as a disk, ROM, or RAM, or any combination thereof. Exemplarily, the online monitoring and maintenance management system 100 for sensor networks may also include program instructions stored in ROM, RAM, or other types of non-transitory storage media, or any combination thereof. The methods of this application can be implemented according to these program instructions. The online monitoring and maintenance management system 100 for sensor networks also includes an I / O interface 150 between the computer and other input / output devices.

[0105] For ease of explanation, only one processor is described in the online monitoring and maintenance management system 100 for sensor networks. However, it should be noted that the online monitoring and maintenance management system 100 for sensor networks in this application may also include multiple processors. Therefore, the steps performed by one processor as described in this application may also be performed jointly by multiple processors or individually. For example, if the processor of the online monitoring and maintenance management system 100 for sensor networks performs steps A and B, it should be understood that steps A and B may also be performed jointly by two different processors or individually by one processor. For example, the first processor performs step A, the second processor performs step B, or the first processor and the second processor jointly perform steps A and B.

[0106] Furthermore, this embodiment of the invention also provides a readable storage medium, which has computer-executable instructions pre-set in it. When the processor executes the computer-executable instructions, the above-mentioned online monitoring and maintenance management method applied to sensor networks is implemented.

[0107] It should be noted that, in order to simplify the description of the present invention and thus help to understand one or more embodiments of the invention, multiple features may sometimes be grouped into one embodiment, drawing or description thereof in the foregoing description of the embodiments of the present invention.

Claims

1. A method for online monitoring, operation, and maintenance management of sensor networks, characterized in that, The method includes: By linking the real-time operation information of each monitoring node in the sensor network with the network's historical operation and maintenance evolution data, and using a pre-trained AI correlation model to perform dynamic evolution correlation processing of node states, the state evolution trajectory corresponding to each monitoring node is generated. The intelligent simulation process for operation and maintenance requirements is initiated based on the state evolution trajectory of each monitoring node. The AI ​​simulation model is used to mine the correlation evolution features in the state evolution trajectory and form an evolution map of operation and maintenance requirements corresponding to each monitoring node. The system retrieves a pre-defined operation and maintenance solution evolution library and uses an AI adaptation algorithm to dynamically associate the operation and maintenance requirement evolution map of each monitoring node with the solution evolution logic in the operation and maintenance solution evolution library to generate an initial operation and maintenance solution. The state evolution trajectory of each monitoring node and the evolution map of operation and maintenance requirements are input into the AI ​​evolution model to adaptively adjust the initial operation and maintenance plan and obtain the target operation and maintenance plan that is adapted to the node state in real time. The target operation and maintenance plan is synchronized to the corresponding operation and maintenance execution terminal, and the state evolution trajectory operation and maintenance requirement evolution map and target operation and maintenance plan are entered into the operation and maintenance file module of the sensor network.

2. The online monitoring and maintenance management method for sensor networks according to claim 1, characterized in that, The real-time operational information of each monitoring node in the linked sensor network and the network's historical operation and maintenance evolution data are used to perform dynamic evolution correlation processing of node states through a pre-trained AI correlation model, generating the state evolution trajectory corresponding to each monitoring node, including: Collect real-time operating information of each monitoring node in the sensor network, including node signal transmission information, device operating status information, data acquisition and feedback information; Retrieve historical network operation and maintenance evolution data of the sensor network, which includes operation and maintenance operation information and post-operation and maintenance state recovery information corresponding to the past state change information of each monitoring node; Real-time operational information and historical network operation and maintenance evolution data are classified and associated according to monitoring node identifiers to form an information association set corresponding to each monitoring node. Each set of information associations is input into the feature input layer of the pre-trained AI association model, and the dynamic evolution features in each set of information associations are extracted by the feature evolution extraction unit built into the model. By using the trajectory construction unit of the AI ​​association model, dynamic evolution characteristics are associated with the state change patterns in the historical operation and maintenance evolution data of the network to generate a state evolution trajectory that can reflect the trend of node state changes.

3. The online monitoring and maintenance management method for sensor networks according to claim 1, characterized in that, The process of intelligently inferring operation and maintenance requirements based on the state evolution trajectory of each monitoring node is described. This process uses an AI inference model to mine the correlation evolution features in the state evolution trajectory, forming an operation and maintenance requirement evolution map corresponding to each monitoring node, including: The intelligent simulation process for operation and maintenance requirements is initiated based on the state evolution trajectory of each monitoring node, and the pre-trained AI simulation model is loaded and the model operation configuration is completed. The trajectory analysis unit of the AI ​​inference model is used to segment and decompose the state evolution trajectory of each monitoring node, and extract the key state features of each evolution stage in the trajectory. By using the feature association units of the AI ​​inference model, we can explore the evolutionary relationships between key state features and sort out the sequential evolution logic of different key state features. The AI-driven model uses a demand mapping unit to associate evolutionary relationships with a pre-defined database of operational and maintenance requirements, matching the basic operational and maintenance requirements corresponding to each evolutionary stage. The basic operation and maintenance requirements of each evolution stage are connected in accordance with the evolution logic through the graph construction unit of the AI ​​inference model, and the change trend of key state characteristics is combined to form an evolution graph of operation and maintenance requirements.

4. The online monitoring and maintenance management method for sensor networks according to claim 1, characterized in that, The process involves inputting the state evolution trajectory of each monitoring node and the evolution map of operation and maintenance requirements into the AI ​​evolution model to adaptively adjust the initial operation and maintenance plan, thereby obtaining a target operation and maintenance plan that is adapted to the node state in real time. This includes: Input the state evolution trajectory of each monitoring node and the evolution map of operation and maintenance requirements into the AI ​​evolution model, and load the initial operation and maintenance plan at the same time; The initial operation and maintenance plan is decomposed into operation process module, tool adaptation module, execution sequence module and parameter configuration module through the solution decomposition unit of the AI ​​evolution model; The feature matching unit of the AI ​​evolution model associates the key state features in the state evolution trajectory with the adaptation conditions of each module, and identifies the adaptation differences between each module and the current node state. Based on the adaptation differences and the evolution trend of the evolution map of the AI ​​evolution model, the module evolution units are adjusted in a targeted manner to generate the evolved module content. The AI ​​evolution model's solution integration unit reorganizes the evolved modules according to the operation and maintenance logic to form a target operation and maintenance solution that adapts to the node status in real time.

5. The online monitoring and maintenance management method for sensor networks according to claim 2, characterized in that, The feature evolution extraction unit of the AI ​​association model extracts dynamic evolution features from each information association set, including: The feature evolution extraction unit of the AI ​​association model captures multi-dimensional features of real-time running information in the information association set and extracts the instantaneous state features in the real-time running information. The network historical operation and maintenance evolution data in the information association set is divided into stages, and the historical state characteristics of each stage and the state transition characteristics between stages are extracted. By aligning the instantaneous state features with the historical state features in a time sequence, we can uncover the similar evolutionary patterns between the instantaneous state features and the historical state features. By combining state transition characteristics to analyze the potential evolution direction of instantaneous state characteristics, a dynamic evolution feature set containing instantaneous state characteristics, historical state characteristics, and potential evolution direction is generated. Feature filtering is performed on the dynamic evolution feature set to retain the core features directly related to the evolution of node states, forming dynamic evolution features corresponding to each information association set.

6. The online monitoring and maintenance management method for sensor networks according to claim 3, characterized in that, The graph construction unit using the AI ​​inference model connects the basic operation and maintenance requirements of each evolution stage according to the evolutionary logic, and combines the changing trends of key state characteristics to form an evolution graph of operation and maintenance requirements, including: The time sequence of each evolution stage is determined by the graph construction unit of the AI ​​inference model, and the basic operation and maintenance requirements are initially sorted according to the sequence. Extract the key state feature thresholds corresponding to each basic operation and maintenance requirement, and label the state conditions that trigger each basic operation and maintenance requirement. Analyze the relationships between basic operation and maintenance requirements in adjacent evolution stages to determine the logic of the impact of operation and maintenance requirements in the previous stage on operation and maintenance requirements in the next stage. Based on the changing trends of key status characteristics, predict the duration and evolution / upgrade path of each basic operation and maintenance requirement; The sorted basic operation and maintenance requirements status conditions, relationships, and evolutionary upgrade paths are integrated into structured graph data. Basic operation and maintenance requirements are represented by graph nodes, and relationships and evolutionary paths are represented by edges, forming an operation and maintenance requirements evolution graph.

7. The online monitoring and maintenance management method for sensor networks according to claim 4, characterized in that, The module evolution unit, through the AI ​​evolution model, makes targeted adjustments to each module based on the evolution trend of the adaptation differences and operational needs evolution map, generating the evolved module content, including: For the operation process module, special operation steps that can alleviate state evolution deviations are added based on adaptation differences, and the execution timing of operation steps is adjusted in combination with the evolution trend of the operation and maintenance demand evolution map. For the tool adaptation module, based on the mismatch between the tool and the node status in the adaptation differences, replace it with an operation and maintenance tool type that can adapt to the current status and future evolution trend; For the execution order module, referencing the adjusted operation process module and the updated tool adaptation module, the operation order was rearranged according to the principle of optimal operation and maintenance efficiency; For the parameter configuration module, the execution parameters of each operation are adjusted based on the characteristic change rate in the state evolution trajectory and the upgrade path of the operation and maintenance requirement evolution map. Logical verification is performed on each adjusted module to ensure smooth connection between modules and form the evolved module content.

8. The online monitoring and maintenance management method for sensor networks according to claim 1, characterized in that, The process involves dynamically associating the evolution map of operation and maintenance requirements for each monitoring node with the solution evolution logic in the operation and maintenance solution evolution library using an AI adaptation algorithm to generate an initial operation and maintenance solution, including: Retrieve a preset operation and maintenance scheme evolution library, which stores the scheme evolution logic corresponding to various basic operation and maintenance schemes. The scheme evolution logic includes the adjustment rules, evolution trigger conditions, and adaptation state range of the scheme module. The AI ​​adaptation algorithm is used to encode the features of the operation and maintenance demand evolution map, generating a demand encoding vector that can reflect the demand evolution pattern. Feature encoding is performed on the evolution logic of each scheme in the operation and maintenance scheme evolution library to generate a scheme encoding vector corresponding to each scheme evolution logic; The demand encoding vector is dynamically associated with the encoding vectors of each solution to uncover the points of convergence between the demand evolution pattern and the solution evolution logic. The optimal matching solution evolution logic is selected based on the matching points, and an initial operation and maintenance solution is generated based on this solution evolution logic to adapt to the current operation and maintenance requirements evolution map.

9. The online monitoring and maintenance management method for sensor networks according to claim 5, characterized in that, The method combines state transition feature analysis with the potential evolution direction of the instantaneous state features to generate a dynamic evolution feature set containing instantaneous state features, historical state features, and potential evolution directions, including: The state transition features are classified according to the evolution rate, and the state change patterns under different evolution rates are sorted out. Associating real-time state features with various state change patterns, and identifying the target change patterns corresponding to the real-time state features; Based on the target change pattern, the possible state values ​​of the instantaneous state characteristics in subsequent periods are deduced to determine the potential evolution direction of the instantaneous state characteristics. The instantaneous state features, historical state features, and potential evolution directions are classified and integrated according to feature type, and the correlation weights between each feature are labeled. The classified features are arranged in an orderly manner according to the association weights to form a dynamic evolution feature set that includes immediate state features, historical state features, and potential evolution directions.

10. An online monitoring and maintenance management system for sensor networks, characterized in that, The online monitoring and maintenance management system for sensor networks includes a processor and a memory, the memory and the processor being connected. The memory is used to store programs, instructions or code, and the processor is used to execute the programs, instructions or code in the memory to implement the online monitoring and maintenance management method for sensor networks as described in any one of claims 1-9.

Images