A remote intelligent monitoring system and method for a building internal power distribution system

Abstract

This invention belongs to the field of power distribution monitoring technology and discloses a remote intelligent monitoring system and method for building power distribution systems. The system includes a data fusion module, a perception map construction module, an operational health assessment module, a fault prediction module, an instruction optimization module, and a collaborative execution module. First, multimodal data fusion is performed on the electrical parameters and status information of power distribution nodes to obtain a standard monitoring dataset. Then, based on the node topology diagram, state space projection is performed on this dataset to obtain a situational awareness map. Subsequently, fuzzy comprehensive scoring of the operational parameters of the map nodes is performed, and fault early warning information is obtained through cascaded fault prediction. Combining the early warning information with the map-optimized basic monitoring and control instructions, remote monitoring and control instructions are obtained. Finally, edge collaborative control is performed on the execution terminal according to the instructions to obtain the monitoring operation optimization results, which are then sent to the cloud monitoring platform. This invention can improve the perception depth, judgment accuracy, and control precision of remote monitoring of building power distribution systems.

Description

Technical Field

[0001] This invention belongs to the field of power distribution monitoring technology, and in particular relates to a remote intelligent monitoring system and method for power distribution systems in buildings. Background Technology

[0002] The building's power distribution system is the core energy supply network of modern buildings. It is responsible for transmitting municipal high-voltage electricity through transformers and distribution to end-user circuits such as lighting, power, and low-voltage circuits, ensuring the continuous and reliable operation of various equipment within the building. Currently, the operation monitoring of building power distribution systems mainly relies on traditional manual inspections, local instrument monitoring, and basic SCADA systems. These systems collect basic electrical parameters such as voltage and current, and combine threshold alarms or simple statistical analysis to achieve fault response. While some advanced solutions introduce models such as fuzzy comprehensive evaluation and shallow neural networks to attempt to extend single-point fault monitoring to quantitative assessment of equipment health, they still rely on static parameter comparison and local state judgment as their core logic, and have not yet formed a complete technical system covering multi-source data fusion, dynamic trend inference, and full-link collaborative control.

[0003] The main problems with existing monitoring technologies for building power distribution systems are as follows: First, the perception dimension is one-sided. Most solutions only focus on collecting electrical quantities and do not integrate multimodal information such as environmental temperature and humidity, equipment topology association, and historical operation and maintenance records. This makes it difficult to comprehensively depict the true health status of the system and easily misses potential faults. Second, the assessment accuracy is insufficient. Static weight allocation and coarse level classification methods cannot smoothly reflect the continuous process of parameter deterioration. Key status information is easily lost, and the accuracy of assessment models is generally low, making it difficult to support accurate health trend prediction. Third, the control response is lagging. There is a lack of a closed-loop collaborative mechanism from status perception to fault prediction to command optimization. It is poorly adaptable to scenarios with dense circuits at the end of building power distribution and complex operating conditions. It cannot achieve efficient remote intelligent monitoring and preventive operation and maintenance, which restricts the improvement of power distribution system operation stability and energy utilization efficiency. Summary of the Invention

[0004] In view of the shortcomings of the prior art, the purpose of this invention is to provide a remote intelligent monitoring system and method for building power distribution systems, which comprehensively improves the perception depth, judgment accuracy and control precision of remote monitoring of building power distribution systems.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows: A remote intelligent monitoring system for building power distribution systems includes: The data fusion module is used to perform multimodal data fusion on the electrical parameters and status information of the power distribution nodes in the building to obtain a standard monitoring dataset of the power distribution nodes in the building. The perception map construction module is used to project the state space of a standard monitoring dataset based on the topology map of the power distribution nodes in the building, so as to obtain the situational perception map of the power distribution nodes in the building. The health assessment module is used to perform fuzzy comprehensive scoring on the node operation parameters of the situational awareness map to obtain the operation health score of the power distribution nodes in the building. The fault prediction module is used to predict cascaded faults of power distribution nodes within a building based on operational health scores, and to obtain fault warning information for power distribution nodes within the building. The instruction optimization module is used to perform multi-objective optimization of the basic monitoring and control instructions of the power distribution nodes in the building based on fault early warning information and situational awareness map, so as to obtain the remote monitoring and control instructions of the power distribution nodes in the building. The collaborative execution module is used to perform edge collaborative control of the execution terminals of the power distribution nodes in the building based on remote monitoring and control commands, obtain the monitoring and operation optimization results of the power distribution nodes in the building, and send the monitoring and operation optimization results to the cloud monitoring platform of the power distribution nodes in the building.

[0006] Preferably, in the data fusion module, the process of obtaining the standard monitoring dataset of the building's power distribution nodes is as follows: Real-time status perception of power distribution nodes within the building is performed to obtain the electrical parameters and status information of the power distribution nodes within the building. The electrical parameters and status information are then aligned with time-series features to obtain synchronous monitoring data of the power distribution nodes within the building. The synchronous monitoring data is normalized to obtain standardized node monitoring data of the power distribution nodes in the building; Heterogeneous data correlation is performed on the electrical parameters and status information in the standardized node monitoring data to obtain multimodal data records of the power distribution nodes in the building; Based on the node number of the power distribution nodes in the building, semantic alignment and fusion are performed on the multimodal data records to obtain a standard monitoring dataset of the power distribution nodes in the building.

[0007] Preferably, in the perception map construction module, the process of obtaining the situational awareness map of the power distribution nodes within the building is as follows: The topology of the power distribution nodes within the building is encoded using graph structure encoding to obtain the node topology adjacency matrix of the power distribution nodes within the building. Based on the node topology adjacency matrix, spatial index binding is performed on the node monitoring data in the standard monitoring dataset to obtain the electrical status label node set of the power distribution nodes in the building. Based on the connection relationship of the node topology adjacency matrix, coupling feature extraction is performed on the electrical parameters of adjacent nodes in the electrical status label node set to obtain the electrical coupling relationship edge set of the power distribution nodes in the building. Knowledge graph embedding is performed on the electrical status label node set and the electrical coupling relationship edge set to obtain the situational awareness graph of the power distribution nodes in the building.

[0008] Preferably, in the operational health assessment module, the process of obtaining the operational health score of the building's power distribution nodes is as follows: Data mining is performed on the situational awareness map to obtain the node operation parameters of the power distribution nodes within the building; Based on the preset node health level standard, the node operation parameters of the power distribution nodes in the building are subjected to multi-level fuzzy mapping to obtain the node dimension membership vector of the power distribution nodes in the building. Based on the node health level standard, the node location and power supply importance of the situational awareness map are centrally weighted to obtain the node dimension weight vector of the power distribution node in the building. The operational health score of the power distribution nodes in the building is obtained by weighting and comprehensively evaluating the node membership vector and the node weight vector.

[0009] Preferably, the formula for calculating the operational health score is: ; In the formula, H represents the operational health score of the power distribution node within the building, m represents the total number of preset electrical operating parameter categories for the node health level standard, and n represents the total number of preset discrete state intervals for the node health level standard. For the i-th weight coefficient of the node dimension weight vector, Let represent the fuzzy membership degree of the i-th actual running parameter value in the node dimension membership vector, belonging to the j-th health level. Let be the quantitative score corresponding to the j-th health level in the node health level standard. This is the preset scaling factor.

[0010] Preferably, in the fault prediction module, the process of obtaining fault early warning information for power distribution nodes within the building is as follows: Based on the operational health score, change point detection is performed on the power distribution nodes within the building to identify potential risk nodes in the power distribution nodes within the building. Based on potential risk nodes, risk neighborhood mining is performed on the situational awareness map to obtain the risk node topology and electrical coupling degree of risk nodes in the building's power distribution nodes. Based on the topological relationship of risk nodes and potential risk nodes, load transfer tracing is performed to obtain the cascading fault propagation path of power distribution nodes within the building; Based on the node operating parameters, operational health scores, and electrical coupling degree of risk nodes in the cascaded fault propagation path, a multi-source information fusion assessment is conducted on the probability of fault occurrence and the scope of fault impact in the propagation path of the cascaded fault propagation path to obtain fault early warning information for power distribution nodes within the building.

[0011] Preferably, in the instruction optimization module, the process of obtaining remote monitoring and control instructions for the building's power distribution nodes is as follows: By performing reverse reasoning on fault early warning information and situational awareness maps, the operating constraints and key performance indicators of power distribution nodes within the building are obtained; The operational constraints and key performance indicators are transformed into target planning to obtain the optimization target set of the power distribution nodes in the building; Command strategies are constructed for the power distribution nodes within the building to obtain the basic monitoring and control commands for the power distribution nodes within the building. Adjustable parameters are identified for the basic monitoring and control commands to obtain the optimized variables for the basic monitoring and control commands. The optimization objective set, operational constraints, and optimization variables are encoded into candidate optimization instructions for the building's power distribution nodes; Based on the node operating status and key performance indicators of the situational awareness map, multi-attribute decision-making is performed on candidate optimization commands to obtain remote monitoring and control commands for the power distribution nodes in the building.

[0012] Preferably, the process for obtaining the operational constraints and key performance indicators of the power distribution nodes within a building is as follows: Key elements of fault warning information are extracted to obtain the fault nodes and risk propagation paths of the fault warning information; Based on the faulty nodes and risk propagation paths, local graph focusing is performed on the situational awareness map to obtain the real-time operating parameters and topological connections of the faulty nodes. The real-time operating parameters and topological connections are then encapsulated in a structured manner to obtain the initial fact set of the faulty nodes. Based on the fault phenomena in the initial fact set, the perturbation propagation inversion of the risk propagation path is performed to obtain the candidate causal variables of the power distribution nodes in the building. Causal characteristic analysis was performed on the candidate causal variables to obtain the key constraint variables of the candidate causal variables; Based on the correlation between key constraint variables and fault nodes, implicit constraint inversion is performed on the situational awareness map to obtain the operational constraints and key performance indicators of the power distribution nodes within the building.

[0013] Preferably, in the collaborative execution module, the process of obtaining the monitoring and operation optimization results of the building's power distribution nodes and sending these results to the cloud monitoring platform of the building's power distribution nodes is as follows: The remote monitoring and control commands are orchestrated and decomposed to obtain the target execution terminal and edge execution command set of the power distribution nodes in the building. Probe detection is performed on the target execution terminal to obtain the real-time operating status and communication response latency of the target execution terminal; Dynamic timing planning is performed on the real-time running status and communication response delay to obtain the collaborative control timing and action execution priority of the target execution terminal. Resource scheduling optimization is then performed on the collaborative control timing and action execution priority to obtain the edge collaborative scheduling strategy of the target execution terminal. Based on the edge collaborative scheduling strategy, the control actions of the edge execution instruction set are sent to the target execution terminal for distributed consensus recording, thereby obtaining the node control execution record of the target execution terminal; Based on the node control execution records, the control effectiveness of the real-time operating parameters of the nodes in the situational awareness map is evaluated to obtain the monitoring and operation optimization results of the power distribution nodes in the building, and the monitoring and operation optimization results are sent to the cloud monitoring platform of the power distribution nodes in the building.

[0014] A remote intelligent monitoring method for a building's power distribution system, applied to the aforementioned system, includes the following steps: S1. Perform multimodal data fusion on the electrical parameters and status information of the power distribution nodes in the building to obtain the standard monitoring dataset of the power distribution nodes in the building; S2. Based on the topology diagram of the power distribution nodes in the building, the standard monitoring dataset is projected into the state space to obtain the situational awareness map of the power distribution nodes in the building. S3. Perform fuzzy comprehensive scoring on the node operation parameters of the situational awareness map to obtain the operational health score of the power distribution nodes in the building. S4. Based on the operational health score, perform cascade fault prediction on the power distribution nodes within the building to obtain fault warning information for the power distribution nodes within the building. S5. Based on the fault warning information and situational awareness map, perform multi-objective optimization on the basic monitoring and control instructions of the power distribution nodes in the building to obtain the remote monitoring and control instructions of the power distribution nodes in the building. S6. Based on remote monitoring and control commands, perform edge collaborative control on the execution terminals of the power distribution nodes in the building to obtain the monitoring and operation optimization results of the power distribution nodes in the building, and send the monitoring and operation optimization results to the cloud monitoring platform of the power distribution nodes in the building.

[0015] The present invention has the following beneficial effects: This invention achieves standardized processing of electrical parameters and status information of power distribution nodes within buildings through multimodal data fusion. It combines topology diagrams to complete the state space projection of standard monitoring datasets, constructing a situational awareness map that accurately reflects the electrical coupling relationships of nodes. Simultaneously, it uses a multi-dimensional weighted fuzzy comprehensive scoring method to complete node operational health scoring. Based on the health score, it conducts cascading fault prediction, accurately identifies potential risk nodes, and tracks fault propagation paths. Multi-source information fusion evaluation ensures that fault warning information is both accurate and comprehensive. From data perception to fault prediction, it achieves precise implementation of technical actions throughout the entire process, improving the perception depth and judgment accuracy of remote monitoring of power distribution systems within buildings.

[0016] This invention combines fault early warning information with situational awareness maps to conduct reverse reasoning, allowing the optimization of control commands to align with the actual operational constraints and performance indicators of power distribution nodes. Multi-objective optimization enables precise upgrades of basic monitoring and control commands. Edge collaborative control uses probes to detect terminal status and formulate targeted scheduling strategies, clarifying control timing and execution priorities. Distributed consensus recording and control effectiveness evaluation enable full traceability and effect verification of control actions. Real-time data synchronization from the cloud monitoring platform allows for real-time feedback of monitoring results. The seamless integration of technologies throughout the entire process improves the execution efficiency and adaptability of remote control of the power distribution system, ensuring the stability and intelligence level of the power distribution system operation. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the modules of the system of the present invention; Figure 2 This is a flowchart illustrating the method of the present invention. Detailed Implementation

[0018] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings.

[0019] Example 1: As Figure 1 As shown, a remote intelligent monitoring system for building power distribution systems includes: The data fusion module is used to perform multimodal data fusion on the electrical parameters and status information of the power distribution nodes in the building to obtain a standard monitoring dataset of the power distribution nodes in the building. The perception map construction module is used to project the state space of a standard monitoring dataset based on the topology map of the power distribution nodes in the building, so as to obtain the situational perception map of the power distribution nodes in the building. The health assessment module is used to perform fuzzy comprehensive scoring on the node operation parameters of the situational awareness map to obtain the operation health score of the power distribution nodes in the building. The fault prediction module is used to predict cascaded faults of power distribution nodes within a building based on operational health scores, and to obtain fault warning information for power distribution nodes within the building. The instruction optimization module is used to perform multi-objective optimization of the basic monitoring and control instructions of the power distribution nodes in the building based on fault early warning information and situational awareness map, so as to obtain the remote monitoring and control instructions of the power distribution nodes in the building. The collaborative execution module is used to perform edge collaborative control of the execution terminals of the power distribution nodes in the building based on remote monitoring and control commands, obtain the monitoring and operation optimization results of the power distribution nodes in the building, and send the monitoring and operation optimization results to the cloud monitoring platform of the power distribution nodes in the building.

[0020] The system in this embodiment can be hosted on a cloud server. In terms of implementation, it can function as one or more service devices, or as an application installed in the cloud (e.g., a mobile service operator's server, server cluster, etc.), or it can be developed into a website. A module, also called a unit, refers to a series of computer program segments that can be executed by the processor of an electronic device and perform a fixed function, stored in the memory of the electronic device. Each module can be implemented independently and can call other modules.

[0021] In the data fusion module, the process of obtaining the standard monitoring dataset for the building's power distribution nodes is as follows: Real-time status perception of power distribution nodes within the building is performed to obtain the electrical parameters and status information of the power distribution nodes within the building. The electrical parameters and status information are then aligned with time-series features to obtain synchronous monitoring data of the power distribution nodes within the building. The synchronous monitoring data is normalized to obtain standardized node monitoring data of the power distribution nodes in the building; Heterogeneous data correlation is performed on the electrical parameters and status information in the standardized node monitoring data to obtain multimodal data records of the power distribution nodes in the building; Based on the node number of the power distribution nodes in the building, semantic alignment and fusion are performed on the multimodal data records to obtain a standard monitoring dataset of the power distribution nodes in the building.

[0022] The sensor acquisition devices deployed at the power distribution nodes within the building are continuously activated. These devices periodically collect electrical parameters such as voltage, current, and power, as well as status information such as on / off, operation, and standby status of the power distribution nodes. The collection interval is set to 1 second. The electrical parameters and status information collected at the same timestamp for the same power distribution node are matched one by one. For data with missing timestamps, the data is supplemented by the average of the data collected at the two adjacent valid timestamps of that node. After completing the data matching and supplementation for all power distribution nodes, the synchronous monitoring data of the power distribution nodes within the building is obtained.

[0023] The different types of electrical parameters and status information in the synchronous monitoring data are converted according to the preset dimension unification rules. The electrical parameters are converted into dimensionless values ​​based on their own rated value range. The status information is converted into fixed numerical values ​​in the 0-1 range according to the preset digital encoding rules. After the dimension and representation of all data are unified, the standardized node monitoring data of the power distribution nodes in the building are obtained.

[0024] Using the unique identifier of each power distribution node in the building as the core association basis, all electrical parameters and status information belonging to the same power distribution node in the standardized node monitoring data are structured and integrated. Various electrical parameters are set as numerical fields and various status information are set as identifier fields, and integrated into the same independent structured data unit. Each power distribution node corresponds to a dedicated structured data unit. After the structured integration of all power distribution nodes is completed, the multimodal data record of the power distribution nodes in the building is obtained.

[0025] Using the unique node number of the power distribution node in the building as the basis for global matching, global semantic calibration processing is carried out on the multimodal data records. The content consistency of the multimodal data records corresponding to the same node number under different acquisition channels and different data formats is verified. The verification standard is that the deviation of the acquired values ​​of the same data field does not exceed the preset 5%. For fields with deviation within the preset threshold, the average value of the acquired values ​​is taken and merged. For fields with deviation exceeding the preset threshold, abnormal acquired data is removed and valid acquired data is retained. After completing the data verification and merging of all power distribution nodes, the standard monitoring dataset of the power distribution nodes in the building is obtained.

[0026] In the data fusion module, multimodal data fusion of electrical parameters and status information of building power distribution nodes is achieved through phased and standardized processing actions. Each processing step has clear preset operation rules and quantification thresholds, ensuring the reproducibility of the technical solution. From real-time acquisition to time-series alignment, data temporal consistency is ensured. Normalization processing achieves uniformity in the form of different types of data. Heterogeneous data association completes the structured integration of cross-type data. Semantic alignment fusion based on node numbers ensures the global accuracy and standardization of the data. The final standard monitoring dataset has the characteristics of being structured, standardized, and complete, and can directly provide accurate and effective data source support for subsequent technical processing links related to building power distribution nodes.

[0027] In the perception map construction module, the process of obtaining the situational awareness map of the power distribution nodes within the building is as follows: The topology of the power distribution nodes within the building is encoded using graph structure encoding to obtain the node topology adjacency matrix of the power distribution nodes within the building. Based on the node topology adjacency matrix, spatial index binding is performed on the node monitoring data in the standard monitoring dataset to obtain the electrical status label node set of the power distribution nodes in the building. Based on the connection relationship of the node topology adjacency matrix, coupling feature extraction is performed on the electrical parameters of adjacent nodes in the electrical status label node set to obtain the electrical coupling relationship edge set of the power distribution nodes in the building. Knowledge graph embedding is performed on the electrical status label node set and the electrical coupling relationship edge set to obtain the situational awareness graph of the power distribution nodes in the building.

[0028] Each node in the topology diagram of the building's power distribution nodes is uniquely identified. The identification rule is to encode the nodes digitally according to the physical arrangement order of the building's power distribution layout. Then, all power distribution nodes are arranged in rows and columns according to the topology identifier order to form a matrix frame. Nodes with direct physical connections are marked with a preset connection identifier 1 at the matrix frame position, and nodes without direct physical connections are marked with a preset no-connection identifier 0 at the matrix frame position. After verifying and marking the physical connection relationships of all power distribution nodes in the topology diagram one by one, a matrix data with the number of rows and columns completely consistent with the total number of power distribution nodes in the building is formed, resulting in the node topology adjacency matrix of the power distribution nodes in the building.

[0029] The unique topology identifier of each power distribution node in the node topology adjacency matrix is ​​extracted as the spatial index primary key. All monitoring data of the power distribution node corresponding to the topology identifier in the standard monitoring dataset are used as spatial index values. The binding operation of each node topology identifier and the corresponding monitoring data is completed according to the one-to-one matching rule. Then, based on the actual collected values ​​in the standard monitoring dataset, a status label is added to each bound data unit. The status labels are divided into two categories: normal and abnormal. The labeling is based on the preset rated operating parameter range of the power distribution node. If the collected value is within the range, it is labeled as normal status; if the collected value exceeds the range, it is labeled as abnormal status. After completing the index binding and status label addition of all nodes, a structured node data set is formed, resulting in the electrical status label node set of the power distribution nodes in the building.

[0030] The system analyzes the location information of all nodes marked with connection identifier 1 in the node topology adjacency matrix to determine all adjacent distribution nodes corresponding to each distribution node under the physical connection relationship. Using three preset electrical parameter association dimensions—voltage transmission, current distribution, and power coupling—it extracts the corresponding dimension electrical parameters of each distribution node and its adjacent distribution nodes at the same timestamp. The system sorts out and records the association changes of the extracted electrical parameters. For each pair of nodes with physical connections and completed coupling feature sorting, a unique edge identifier is configured. The edge identifier contains two types of information: the topology identifier of the node pair and the coupling feature type. All node pairs with configured edge identifiers are integrated to form a structured edge data set, thus obtaining the electrical coupling relationship edge set of the distribution nodes in the building.

[0031] A pre-defined two-layer graph data framework is constructed, comprising two basic layers: a node layer and an edge layer. The node layer carries electrical status label node set data, while the edge layer carries electrical coupling relationship edge set data. Each node data in the electrical status label node set is precisely mapped to its corresponding position in the node layer of the graph data framework according to a unique topological identifier. Similarly, each edge data in the electrical coupling relationship edge set is precisely mapped to its corresponding position in the edge layer of the graph data framework according to a unique edge identifier. During the mapping process, the topological connection relationship between the edge layer data and the node layer data is kept completely consistent. After all data mapping is completed, the graph data framework is fused and verified. The verification criteria are that there is no duplication between node identifiers and edge identifiers, and no deviation in the connection relationship between nodes and edges. After the verification is passed, a complete power distribution node operation status association graph is formed, resulting in a situational awareness graph of the power distribution nodes within the building.

[0032] In the perception graph construction module, through clear operation rules and quantitative annotation standards, the state space projection of the standard monitoring dataset based on the topology diagram of the power distribution nodes in the building is realized. The operation of each step is reproducible. The node topology adjacency matrix accurately restores the physical connection relationship of the power distribution nodes. Spatial index binding realizes the accurate matching of monitoring data and topology nodes. Coupled feature extraction fully explores the electrical correlation characteristics between adjacent nodes. Knowledge graph embedding builds a structured graph system in which nodes and edges are interconnected. The resulting situational awareness graph can comprehensively and accurately reflect the topological connection status and electrical operation coupling relationship of the power distribution nodes in the building. It provides structured, visualized and highly correlated graph data support for subsequent operational health assessment and fault prediction of the power distribution nodes in the building, ensuring the accuracy and effectiveness of subsequent technical steps.

[0033] In the operational health assessment module, the process of obtaining the operational health score of the building's power distribution nodes is as follows: Data mining is performed on the situational awareness map to obtain the node operation parameters of the power distribution nodes within the building; Based on the preset node health level standard, the node operation parameters of the power distribution nodes in the building are subjected to multi-level fuzzy mapping to obtain the node dimension membership vector of the power distribution nodes in the building. Based on the node health level standard, the node location and power supply importance of the situational awareness map are centrally weighted to obtain the node dimension weight vector of the power distribution node in the building. The operational health score of the building's power distribution nodes is obtained by weighting and comprehensively evaluating the node membership vector and node weight vector. The calculation formula is as follows: ; In the formula, H represents the operational health score of the power distribution node within the building, m represents the total number of preset electrical operating parameter categories for the node health level standard, and n represents the total number of preset discrete state intervals for the node health level standard. For the i-th weight coefficient of the node dimension weight vector, Let represent the fuzzy membership degree of the i-th actual running parameter value in the node dimension membership vector, belonging to the j-th health level. Let be the quantitative score corresponding to the j-th health level in the node health level standard. This is the preset scaling factor.

[0034] Data is extracted node by node from the situational awareness map based on the unique topological identifier of the power distribution nodes within the building. The extraction scope includes all electrical operating parameter fields and status identifier fields at the node layer of the map. The integrity of the extracted single-node data is verified, with the verification standard being that the number of missing electrical operating parameter fields in a single node's data does not exceed one. For node data with missing single fields, the average value of the same field from the node's last 10 consecutive collection cycles is used to complete the data. For node data without missing fields, the original collected values ​​are directly retained. After completing the data extraction, verification, and completion for all nodes, a structured parameter set is formed to obtain the node operating parameters of the power distribution nodes within the building. The node health level standard is formulated based on industry technical specifications for the operation of power distribution nodes within the building and long-term historical operating data of the power distribution nodes. The preset electrical operating parameter categories are the core operating parameters of the power distribution nodes, including categories such as voltage, current, and power. All preset electrical operating parameter categories in this standard are statistically analyzed one by one, and the resulting statistical quantity is m.

[0035] Data processing is carried out based on the preset node health level standard. The standard presets four health levels: excellent, good, medium and poor. For each level, it matches the rated value range of various electrical operating parameters of the power distribution node in the building. The interval division is set according to the rated operating value range and fault critical value range of the corresponding electrical parameter. The discrete state intervals divided for all electrical operating parameters in the node health level standard are statistically analyzed as a whole, and the statistical quantity obtained is n.

[0036] For each type of electrical parameter in the node's operating parameters, a level matching determination is performed. According to a preset membership determination rule, the degree to which the actual value belongs to one of the four health levels is labeled. The determination rule is that if the actual value falls within a certain level range, the corresponding membership degree is labeled; if it falls outside a range, the membership degree is labeled according to the percentage of the value. This membership degree value is the [value to be filled in]. Simultaneously, based on the practical application needs of power distribution node operational health assessment, a unique and continuous quantitative score is assigned to each health level; this score is... The membership degrees of various health levels corresponding to different electrical parameters under each power distribution node are arranged sequentially according to the preset parameter and level order to form an ordered set of values, thus obtaining the node dimension membership vector of the power distribution node in the building.

[0037] Based on the pre-defined node health level standards for power distribution nodes, a central weighting process is performed on the physical location and power supply importance of each power distribution node in the situational awareness map. Node locations are categorized into core, intermediate, and terminal layers according to the building's power distribution network hierarchy, with each layer corresponding to a pre-defined location weighting coefficient. Power supply importance is categorized into primary, secondary, and tertiary load types according to the load type of the power distribution node, with each load type corresponding to a pre-defined importance weighting coefficient. The location weighting coefficient and power supply importance weighting coefficient corresponding to each electrical operating parameter are then fused together at a fixed ratio to obtain the fused value. All comprehensive weighted coefficients are arranged sequentially according to the preset parameter order to form an ordered set of coefficients, thus obtaining the node dimension weight vector of the power distribution node in the building.

[0038] The node dimension membership vector and the node dimension weight vector are matched with the same parameters and level order. The weight coefficient of each corresponding position is associated with the membership value. All the results after association processing are summarized and integrated according to the preset comprehensive evaluation rules. The basis for summarizing and integrating is to retain the original association characteristics of the values ​​and to avoid duplicate calculations. This is a preset scaling factor, set based on the actual application scenarios and scoring display requirements of the building's power distribution node operational health rating. It is used to convert the raw results of the comprehensive calculation formula to a standardized scoring range of 0-100. This is a fixed preset value verified by industry power distribution operation assessment experience. After scaling transformation, the results are standardized and organized to form a specific score, H, that can directly reflect the operating status of the power distribution node.

[0039] H, calculated by integrating all parameters using a formula, represents the standardized and quantitative result of the operational health status of the building's power distribution nodes. This formula is the specific numerical implementation of the operational health assessment module's weighted comprehensive evaluation of node dimension membership vectors and node dimension weight vectors. Its calculation logic perfectly aligns with the technical steps of weighted comprehensive evaluation, integrating and calculating the weight coefficients of electrical operating parameters, health level membership, and health level quantitative scores. This allows for the accurate integration of multi-dimensional operational status information of power distribution nodes. The formula's calculation process further ensures the objectivity and accuracy of the operational health score results for the building's power distribution nodes. The final operational health score can be directly used as the core numerical basis for the fault prediction module to conduct cascaded fault prediction, achieving technical data integration between the operational health assessment and fault prediction stages.

[0040] In the operational health assessment module, precise extraction of node operational parameters from the situational awareness map is achieved through clear extraction rules and verification standards, ensuring the integrity and validity of basic data. Fuzzy mapping based on preset multi-level node health standards allows the node dimension membership vector to accurately reflect the health level attribution characteristics of each parameter. Central weighting combined with node location and power supply importance ensures that the coefficient settings of the node dimension weight vector better align with the actual operational needs of building power distribution. Orderly matching and summarization rules for weighted comprehensive assessment guarantee the reproducibility of the processing. All formula parameters have clear sources and setting bases, and the calculation logic and technical steps are deeply integrated, further improving the objectivity and accuracy of the operational health score. This makes the entire technical process highly operable and reproducible. The final operational health score comprehensively and accurately reflects the actual operational health status of power distribution nodes within the building, providing accurate and reliable core numerical basis for subsequent cascading fault prediction and achieving efficient data connection between various technical links.

[0041] In the fault prediction module, the process of obtaining fault early warning information for power distribution nodes within the building is as follows: Based on the operational health score, change point detection is performed on the power distribution nodes within the building to identify potential risk nodes in the power distribution nodes within the building. Based on potential risk nodes, risk neighborhood mining is performed on the situational awareness map to obtain the risk node topology and electrical coupling degree of risk nodes in the building's power distribution nodes. Based on the topological relationship of risk nodes and potential risk nodes, load transfer tracing is performed to obtain the cascading fault propagation path of power distribution nodes within the building; Based on the node operating parameters, operational health scores, and electrical coupling degree of risk nodes in the cascaded fault propagation path, a multi-source information fusion assessment is conducted on the probability of fault occurrence and the scope of fault impact in the propagation path of the cascaded fault propagation path to obtain fault early warning information for power distribution nodes within the building.

[0042] The operational health scores of each power distribution node within the building are organized into a time-series data sequence according to the collection time. The normal threshold range for the operational health score is preset to be 60-100 points out of 0-100 points. Scores below 60 points are judged as abnormal scores. Continuous status detection is performed on the time-series data sequence of each power distribution node. If the operational health score of a power distribution node is in the abnormal score range for three consecutive collection cycles, the node is judged as a health score mutation node and marked as a risk node. After completing the change point detection and risk marking of all power distribution nodes, all marked risk nodes are organized into a structured set to obtain the potential risk nodes of the power distribution nodes within the building.

[0043] In the situational awareness map, all potential risk nodes are accurately located based on topology identifiers. Taking each potential risk node as the core, the map is used to identify first-level adjacent nodes with direct physical connections and second-level adjacent nodes with indirect physical connections. The connection levels, connection methods, and topological positional relationships between each adjacent node and the potential risk node are sorted out. All the sorted relationships are structured and integrated according to node identifiers to obtain the risk node topology relationship of the power distribution nodes in the building. At the same time, all electrical operating parameters of the potential risk node and each adjacent node at the same timestamp are extracted, and the mutual influence characteristics between parameters are sorted out. According to the preset coupling degree classification standard, the influence characteristics are divided into three levels: strong, medium, and weak and labeled accordingly to form electrical coupling characteristic data between nodes, thus obtaining the electrical coupling degree of the risk nodes of the power distribution nodes in the building.

[0044] Based on the physical connection sequence and connection method of each node in the risk node topology, the transmission direction and process of power distribution load after a potential risk node fails are simulated. Starting from the potential risk node, each node to which the load transmission points is traced node by node, and each complete path of load transmission is recorded. At the same time, the rated load carrying capacity parameters of each node on each path are extracted. The actual transmitted load is compared with the node's rated load carrying capacity parameters. If the actual transmitted load exceeds the node's rated load carrying capacity parameters, the node is marked as a path interruption node. All complete load transmission paths and the corresponding interruption node information are structured and organized according to the topological order to obtain the cascading fault propagation path of the power distribution nodes in the building.

[0045] The system extracts the node operating parameters, operational health scores, and electrical coupling degrees of risk nodes for all nodes along the cascading fault propagation path. Using the unique topology identifier of the distribution node as the matching basis, the three types of data are matched one-to-one and integrated into a multi-source information dataset. Each propagation path in the dataset is analyzed one by one according to the preset evaluation rules. The probability of fault occurrence of a single node is determined by combining the node operational health score, the probability of fault transmission between nodes is determined by combining the electrical coupling degree of risk nodes, and the range of distribution nodes and distribution areas that may be affected by the fault is determined by combining the node operating parameters and topology relationships. The fault occurrence probability labeling results and the specific node and area information of the fault impact range corresponding to each propagation path are structured and encapsulated to obtain fault early warning information for distribution nodes in the building.

[0046] In the fault prediction module, the precise identification of potential risk nodes is achieved by setting clear scoring thresholds and detection cycles. The risk neighborhood mining stage accurately sorts out the topological association characteristics and electrical coupling characteristics of risk nodes, making the association information of risk nodes more complete. Load transfer tracking fully restores the propagation path and possible interruption of cascading faults by simulating load transfer and load carrying capacity verification. Multi-source information fusion evaluation conducts correlation analysis on node operating parameters, health scores and electrical coupling degree, realizing the accurate determination of fault occurrence probability and impact range. Each step has specific operating rules, judgment standards and quantitative thresholds, ensuring the reproducibility and operability of the technical solution. The final fault warning information includes complete content such as fault propagation path, probability of occurrence and impact range, which can accurately reflect the cascading fault risk status of power distribution nodes in the building, providing comprehensive and accurate risk data support for the optimization of subsequent monitoring and control instructions, and improving the foresight and pertinence of fault warnings in the building's power distribution system.

[0047] In the instruction optimization module, the process of obtaining remote monitoring and control instructions for the building's power distribution nodes is as follows: By performing reverse reasoning on fault early warning information and situational awareness maps, the operating constraints and key performance indicators of power distribution nodes within the building are obtained; The operational constraints and key performance indicators are transformed into target planning to obtain the optimization target set of the power distribution nodes in the building; Command strategies are constructed for the power distribution nodes within the building to obtain the basic monitoring and control commands for the power distribution nodes within the building. Adjustable parameters are identified for the basic monitoring and control commands to obtain the optimized variables for the basic monitoring and control commands. The optimization objective set, operational constraints, and optimization variables are encoded into candidate optimization instructions for the building's power distribution nodes; Based on the node operating status and key performance indicators of the situational awareness map, multi-attribute decision-making is performed on candidate optimization commands to obtain remote monitoring and control commands for the power distribution nodes in the building.

[0048] The process of obtaining the operational constraints and key performance indicators of the power distribution nodes within the building is as follows: Key elements of fault warning information are extracted to obtain the fault nodes and risk propagation paths of the fault warning information; Based on the faulty nodes and risk propagation paths, local graph focusing is performed on the situational awareness map to obtain the real-time operating parameters and topological connections of the faulty nodes. The real-time operating parameters and topological connections are then encapsulated in a structured manner to obtain the initial fact set of the faulty nodes. Based on the fault phenomena in the initial fact set, the perturbation propagation inversion of the risk propagation path is performed to obtain the candidate causal variables of the power distribution nodes in the building. Causal characteristic analysis was performed on the candidate causal variables to obtain the key constraint variables of the candidate causal variables; Based on the correlation between key constraint variables and fault nodes, implicit constraint inversion is performed on the situational awareness map to obtain the operational constraints and key performance indicators of the power distribution nodes within the building.

[0049] The core content of the fault warning information is filtered according to preset rules. Only the fault node identifier, the node sequence and direction information of the risk propagation path are retained. Redundant descriptions are eliminated, the topological identifiers of the selected fault nodes are organized into a structured set, and the risk propagation path is sorted into an ordered data unit to obtain the fault nodes and risk propagation paths of the fault warning information.

[0050] The fault phenomena, such as abnormal parameters and abnormal topology connections of fault nodes in the initial fact set, are analyzed. Based on this, the risk propagation path is traced back from the end to the starting node to simulate the transmission process of operational disturbances. Various factors that may directly cause faults are recorded. Irrelevant factors are eliminated according to the standard that "changes in factors directly lead to the occurrence or aggravation of faults" to obtain candidate causal variables of power distribution nodes in the building.

[0051] Causal characteristic analysis was carried out on each candidate causal variable to sort out the causal relationship between the variable and the fault phenomenon. Directly related variables were screened according to the standard that "variable change is the direct cause of fault". Redundant variables with overlapping causal relationships were eliminated, and the key constraint variables of the candidate causal variables were sorted out.

[0052] By analyzing the correspondence and impact of key constraint variables with fault nodes, and mining the electrical parameter ranges and load capacity limits of distribution nodes associated with fault nodes in the situational awareness map, the operational constraints are compiled into standardized constraint clauses. Based on risk prevention and control needs, specific operational standards are set from the dimensions of node stability, fault interruption, and distribution efficiency, and these are compiled into standardized indicator clauses to obtain key performance indicators.

[0053] The operational constraints are classified according to the node topology identifier, and the key performance indicators are divided into the dimensions of fault prevention, operation efficiency, and stability assurance. Specific optimization objectives are formulated based on the benchmark of "not violating constraints and meeting the minimum requirements of indicators". After eliminating conflicting objectives, they are integrated by dimension to obtain the optimization objective set of the power distribution nodes in the building.

[0054] Based on the needs of routine management and fault prevention of power distribution nodes, standardized control strategies are constructed for electrical parameter adjustment, load distribution control, and connection status adjustment. These strategies are converted into instruction text containing operation objects, types, and basic values. Basic monitoring and control instructions are then compiled. At the same time, the instructions are parsed and adjustable adjustment values, control thresholds, and other parameters are selected to obtain the optimized variables of the basic monitoring and control instructions.

[0055] According to the preset standardized coding rules, unique coding identifiers are assigned to optimization objectives, operational constraints, and optimization variables. The coding is combined according to the logical association between objectives, constraints, and variables, and integrated with the basic monitoring and control instructions to form standardized coding instructions. Each instruction corresponds to a unique combination scheme, and candidate optimization instructions for the power distribution nodes in the building are obtained.

[0056] The real-time electrical parameters, connection status, load bearing, and other operating statuses of nodes in the situational awareness map are extracted. Key performance indicators are used as multi-attribute decision evaluation dimensions. Evaluation criteria for "the degree to which node status indicators meet the standards after command execution" are set for each dimension. Candidate optimization commands are evaluated one by one, commands that do not meet the standards are eliminated, and commands with the best execution effect in each dimension are selected. The remote monitoring and control commands for power distribution nodes in the building are then compiled.

[0057] In the instruction optimization module, clear operating rules, judgment criteria, and structured processing requirements are set for each stage, ensuring the reproducibility and operability of the technical solution. Through multi-step information mining and analysis, reverse reasoning accurately obtains the operating constraints and key performance indicators that fit the actual fault state, laying a precise basis for optimization work. Subsequent stages are advanced layer by layer, making the candidate optimization instructions uniform in format and clear in logic. Multi-attribute decision-making combined with the real-time status and performance indicators of nodes achieves the optimal instruction selection. The final remote monitoring and control instructions match the fault prevention and control needs and the actual operating status of nodes, providing accurate and effective instruction support for edge collaborative control, and greatly improving the pertinence and adaptability of power distribution system monitoring and control instructions.

[0058] In the collaborative execution module, the process of obtaining the monitoring and operation optimization results of the building's power distribution nodes and sending these results to the cloud monitoring platform of the building's power distribution nodes is as follows: The remote monitoring and control commands are orchestrated and decomposed to obtain the target execution terminal and edge execution command set of the power distribution nodes in the building. Probe detection is performed on the target execution terminal to obtain the real-time operating status and communication response latency of the target execution terminal; Dynamic timing planning is performed on the real-time running status and communication response delay to obtain the collaborative control timing and action execution priority of the target execution terminal. Resource scheduling optimization is then performed on the collaborative control timing and action execution priority to obtain the edge collaborative scheduling strategy of the target execution terminal. Based on the edge collaborative scheduling strategy, the control actions of the edge execution instruction set are sent to the target execution terminal for distributed consensus recording, thereby obtaining the node control execution record of the target execution terminal; Based on the node control execution records, the control effectiveness of the real-time operating parameters of the nodes in the situational awareness map is evaluated to obtain the monitoring and operation optimization results of the power distribution nodes in the building, and the monitoring and operation optimization results are sent to the cloud monitoring platform of the power distribution nodes in the building.

[0059] The remote monitoring and control commands are broken down and parsed into operation objects and action content. Based on the topology identifier of the building's power distribution nodes marked in the command, the unique code of the execution terminal deployed on the corresponding power distribution node is matched. All execution terminals that need to perform control actions are selected and organized into a structured set according to the code to obtain the target execution terminals of the building's power distribution nodes. At the same time, the remote monitoring and control commands are broken down according to the operation requirements of each target execution terminal. Each terminal is configured with a unique control action command and organized into a structured command set to obtain the edge execution command set of the building's power distribution nodes.

[0060] A preset probe detection packet is sent to each target execution terminal. This packet contains standardized instructions for querying the terminal's operating status and testing the communication link. Feedback data from the target execution terminal is received. Core information such as the terminal's operating mode, hardware working status, and current load capacity is extracted from the feedback data to obtain the real-time operating status of the target execution terminal. The time difference between the time the probe detection packet is sent and the time the feedback data is received is calculated to obtain the communication response delay of the target execution terminal. For terminals that do not respond, the probe detection packet is sent three times. If there is still no response, the communication response delay is marked as timeout. The preset timeout threshold is 500 milliseconds. Timeout nodes are marked as abnormal terminals and recorded separately.

[0061] The real-time operating status of the target execution terminal is divided into three categories based on hardware load percentage: idle, lightly loaded, and heavily loaded. The classification criteria are: 0-30% load percentage for idle, 31-70% for lightly loaded, and 71-100% for heavily loaded. Communication response latency is divided into three categories based on time value: low latency, medium latency, and high latency. The classification criteria are: ≤100 milliseconds for low latency, 101-300 milliseconds for medium latency, and 301-500 milliseconds for high latency. The lower the operating status load and the shorter the communication response latency, the higher the priority of the control action. This is applied to each target execution... Each control action of the terminal is marked with a unique priority. Then, based on the priority and the topological relationship with the power distribution node, the execution time sequence of each terminal's control action is planned to obtain the collaborative control sequence and action execution priority of the target execution terminal. Combining the actual carrying capacity of the target execution terminal's hardware and communication resources, the number of execution actions within the same time window in the collaborative control sequence is limited. It is preset that a maximum of 5 control actions can be executed in a single time window to avoid resource conflicts. After sorting and optimizing the timing and priority information, the edge collaborative scheduling strategy of the target execution terminal is obtained.

[0062] According to the collaborative control timing and action execution priority set by the edge collaborative scheduling strategy, the control actions in the edge execution instruction set are sent one by one to the corresponding target execution terminal. After receiving the control actions, the target execution terminal synchronizes the control action information in the edge node cluster of the building's power distribution system. All participating edge nodes confirm the consistency of the received information and execution instructions of the control actions. After confirming that there is no deviation, a distributed consensus record is completed. The record includes the target execution terminal code, control action content, instruction reception time, and execution confirmation status. All the distributed consensus records of the target execution terminals are organized into a structured data set according to the execution timing to obtain the node control execution record of the target execution terminal.

[0063] The system extracts control action information that has been confirmed upon completion from the node control execution record. Based on the distribution node topology identifier, it matches the real-time operating parameters of the corresponding node in the situational awareness map before and after control. It compares the parameter changes and judges the control effectiveness according to preset standards. The judgment criteria are whether the distribution node operating parameters have returned to the normal range set by industry standards and whether the fault risk-related parameters have reached the risk elimination standard. The execution effectiveness of each control action is standardized and labeled. The system integrates the real-time operating parameters and control effectiveness labeling results of all distribution nodes in the building after control to form a structured evaluation data set, which yields the monitoring and operation optimization results of the distribution nodes in the building. Through a preset 32-bit check code and encrypted communication protocol, the complete data of the monitoring and operation optimization results is transmitted to the cloud monitoring platform of the distribution nodes in the building to ensure the integrity and accuracy of data transmission.

[0064] In the collaborative execution module, clear operating rules, quantitative division benchmarks, and preset thresholds are set for each link, ensuring the reproducibility and operability of the technical solution. Instruction orchestration and decomposition realize the precise matching of control instructions and execution terminals. Probe detection can comprehensively and accurately obtain the terminal's operation and communication status. Dynamic timing planning and resource scheduling optimization make the execution of control actions more in line with the actual carrying capacity of the terminal and avoid resource conflicts. Distributed consensus recording realizes the full traceability of the control action execution process. Control effectiveness evaluation accurately verifies the control effect through parameter comparison. The final monitoring operation optimization results can fully reflect the actual operating status of the power distribution system after control. Precise transmission to the cloud monitoring platform realizes the real-time update of remote monitoring data, providing practical support for the remote intelligent monitoring of the power distribution system in the building, and greatly improving the execution efficiency and accuracy of edge collaborative control.

[0065] Example 2: As Figure 2 As shown, a remote intelligent monitoring method for a building's power distribution system, based on the system in Embodiment 1, includes the following steps: S1. Perform multimodal data fusion on the electrical parameters and status information of the power distribution nodes in the building to obtain the standard monitoring dataset of the power distribution nodes in the building; S2. Based on the topology diagram of the power distribution nodes in the building, the standard monitoring dataset is projected into the state space to obtain the situational awareness map of the power distribution nodes in the building. S3. Perform fuzzy comprehensive scoring on the node operation parameters of the situational awareness map to obtain the operational health score of the power distribution nodes in the building. S4. Based on the operational health score, perform cascade fault prediction on the power distribution nodes within the building to obtain fault warning information for the power distribution nodes within the building. S5. Based on the fault warning information and situational awareness map, perform multi-objective optimization on the basic monitoring and control instructions of the power distribution nodes in the building to obtain the remote monitoring and control instructions of the power distribution nodes in the building. S6. Based on remote monitoring and control commands, perform edge collaborative control on the execution terminals of the power distribution nodes in the building to obtain the monitoring and operation optimization results of the power distribution nodes in the building, and send the monitoring and operation optimization results to the cloud monitoring platform of the power distribution nodes in the building.

[0066] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A remote intelligent monitoring system for building power distribution systems, characterized in that, include: The data fusion module is used to perform multimodal data fusion on the electrical parameters and status information of the power distribution nodes in the building to obtain a standard monitoring dataset of the power distribution nodes in the building. The perception map construction module is used to project the state space of a standard monitoring dataset based on the topology map of the power distribution nodes in the building, so as to obtain the situational perception map of the power distribution nodes in the building. The health assessment module is used to perform fuzzy comprehensive scoring on the node operation parameters of the situational awareness map to obtain the operation health score of the power distribution nodes in the building. The fault prediction module is used to predict cascaded faults of power distribution nodes within a building based on operational health scores, and to obtain fault warning information for power distribution nodes within the building. The instruction optimization module is used to perform multi-objective optimization of the basic monitoring and control instructions of the power distribution nodes in the building based on fault early warning information and situational awareness map, so as to obtain the remote monitoring and control instructions of the power distribution nodes in the building. The collaborative execution module is used to perform edge collaborative control of the execution terminals of the power distribution nodes in the building based on remote monitoring and control commands, obtain the monitoring and operation optimization results of the power distribution nodes in the building, and send the monitoring and operation optimization results to the cloud monitoring platform of the power distribution nodes in the building.

2. The remote intelligent monitoring system for building power distribution system as described in claim 1, characterized in that, In the data fusion module, the process of obtaining the standard monitoring dataset for the building's power distribution nodes is as follows: Real-time status perception of power distribution nodes within the building is performed to obtain the electrical parameters and status information of the power distribution nodes within the building. The electrical parameters and status information are then aligned with time-series features to obtain synchronous monitoring data of the power distribution nodes within the building. The synchronous monitoring data is normalized to obtain standardized node monitoring data of the power distribution nodes in the building; Heterogeneous data correlation is performed on the electrical parameters and status information in the standardized node monitoring data to obtain multimodal data records of the power distribution nodes in the building; Based on the node number of the power distribution nodes in the building, semantic alignment and fusion are performed on the multimodal data records to obtain a standard monitoring dataset of the power distribution nodes in the building.

3. The remote intelligent monitoring system for building power distribution system as described in claim 1, characterized in that, In the perception map construction module, the process of obtaining the situational awareness map of the power distribution nodes within the building is as follows: The topology of the power distribution nodes within the building is encoded using graph structure encoding to obtain the node topology adjacency matrix of the power distribution nodes within the building. Based on the node topology adjacency matrix, spatial index binding is performed on the node monitoring data in the standard monitoring dataset to obtain the electrical status label node set of the power distribution nodes in the building. Based on the connection relationship of the node topology adjacency matrix, coupling feature extraction is performed on the electrical parameters of adjacent nodes in the electrical status label node set to obtain the electrical coupling relationship edge set of the power distribution nodes in the building. Knowledge graph embedding is performed on the electrical status label node set and the electrical coupling relationship edge set to obtain the situational awareness graph of the power distribution nodes in the building.

4. The remote intelligent monitoring system for building power distribution system as described in claim 1, characterized in that, In the operational health assessment module, the process of obtaining the operational health score of the building's power distribution nodes is as follows: Data mining is performed on the situational awareness map to obtain the node operation parameters of the power distribution nodes within the building; Based on the preset node health level standard, the node operation parameters of the power distribution nodes in the building are subjected to multi-level fuzzy mapping to obtain the node dimension membership vector of the power distribution nodes in the building. Based on the node health level standard, the node location and power supply importance of the situational awareness map are centrally weighted to obtain the node dimension weight vector of the power distribution node in the building. The operational health score of the power distribution nodes in the building is obtained by weighting and comprehensively evaluating the node membership vector and the node weight vector.

5. The remote intelligent monitoring system for building power distribution system as described in claim 4, characterized in that, The formula for calculating the health score is: ； In the formula, H represents the operational health score of the power distribution node within the building, m represents the total number of preset electrical operating parameter categories for the node health level standard, and n represents the total number of preset discrete state intervals for the node health level standard. For the i-th weight coefficient of the node dimension weight vector, Let represent the fuzzy membership degree of the i-th actual running parameter value in the node dimension membership vector, belonging to the j-th health level. Let be the quantitative score corresponding to the j-th health level in the node health level standard. This is the preset scaling factor.

6. The remote intelligent monitoring system for building power distribution system as described in claim 1, characterized in that, In the fault prediction module, the process of obtaining fault early warning information for power distribution nodes within the building is as follows: Based on the operational health score, change point detection is performed on the power distribution nodes within the building to identify potential risk nodes in the power distribution nodes within the building. Based on potential risk nodes, risk neighborhood mining is performed on the situational awareness map to obtain the risk node topology and electrical coupling degree of risk nodes in the building's power distribution nodes. Based on the topological relationship of risk nodes and potential risk nodes, load transfer tracing is performed to obtain the cascading fault propagation path of power distribution nodes within the building; Based on the node operating parameters, operational health scores, and electrical coupling degree of risk nodes in the cascaded fault propagation path, a multi-source information fusion assessment is conducted on the probability of fault occurrence and the scope of fault impact in the propagation path of the cascaded fault propagation path to obtain fault early warning information for power distribution nodes within the building.

7. The remote intelligent monitoring system for building power distribution system as described in claim 1, characterized in that, In the instruction optimization module, the process of obtaining remote monitoring and control instructions for the building's power distribution nodes is as follows: By performing reverse reasoning on fault early warning information and situational awareness maps, the operating constraints and key performance indicators of power distribution nodes within the building are obtained; The operational constraints and key performance indicators are transformed into target planning to obtain the optimization target set of the power distribution nodes in the building; Command strategies are constructed for the power distribution nodes within the building to obtain the basic monitoring and control commands for the power distribution nodes within the building. Adjustable parameters are identified for the basic monitoring and control commands to obtain the optimized variables for the basic monitoring and control commands. The optimization objective set, operational constraints, and optimization variables are encoded into candidate optimization instructions for the building's power distribution nodes; Based on the node operating status and key performance indicators of the situational awareness map, multi-attribute decision-making is performed on candidate optimization commands to obtain remote monitoring and control commands for the power distribution nodes in the building.

8. The remote intelligent monitoring system for building power distribution system as described in claim 7, characterized in that, The process of obtaining the operational constraints and key performance indicators of the power distribution nodes within a building is as follows: Key elements of fault warning information are extracted to obtain the fault nodes and risk propagation paths of the fault warning information; Based on the faulty nodes and risk propagation paths, local graph focusing is performed on the situational awareness map to obtain the real-time operating parameters and topological connections of the faulty nodes. The real-time operating parameters and topological connections are then encapsulated in a structured manner to obtain the initial fact set of the faulty nodes. Based on the fault phenomena in the initial fact set, the perturbation propagation inversion of the risk propagation path is performed to obtain the candidate causal variables of the power distribution nodes in the building. Causal characteristic analysis was performed on the candidate causal variables to obtain the key constraint variables of the candidate causal variables; Based on the correlation between key constraint variables and fault nodes, implicit constraint inversion is performed on the situational awareness map to obtain the operational constraints and key performance indicators of the power distribution nodes within the building.

9. The remote intelligent monitoring system for building power distribution system as described in claim 1, characterized in that, In the collaborative execution module, the process of obtaining the monitoring and operation optimization results of the building's power distribution nodes and sending these results to the cloud monitoring platform of the building's power distribution nodes is as follows: The remote monitoring and control commands are orchestrated and decomposed to obtain the target execution terminal and edge execution command set of the power distribution nodes in the building. Probe detection is performed on the target execution terminal to obtain the real-time operating status and communication response latency of the target execution terminal; Dynamic timing planning is performed on the real-time running status and communication response delay to obtain the collaborative control timing and action execution priority of the target execution terminal. Resource scheduling optimization is then performed on the collaborative control timing and action execution priority to obtain the edge collaborative scheduling strategy of the target execution terminal. Based on the edge collaborative scheduling strategy, the control actions of the edge execution instruction set are sent to the target execution terminal for distributed consensus recording, thereby obtaining the node control execution record of the target execution terminal; Based on the node control execution records, the control effectiveness of the real-time operating parameters of the nodes in the situational awareness map is evaluated to obtain the monitoring and operation optimization results of the power distribution nodes in the building, and the monitoring and operation optimization results are sent to the cloud monitoring platform of the power distribution nodes in the building.

10. A remote intelligent monitoring method for a building's power distribution system, applicable to the claims. The system according to any one of the claims is characterized in that the steps include: S1. Perform multimodal data fusion on the electrical parameters and status information of the power distribution nodes in the building to obtain the standard monitoring dataset of the power distribution nodes in the building; S2. Based on the topology diagram of the power distribution nodes in the building, the standard monitoring dataset is projected into the state space to obtain the situational awareness map of the power distribution nodes in the building. S3. Perform fuzzy comprehensive scoring on the node operation parameters of the situational awareness map to obtain the operational health score of the power distribution nodes in the building. S4. Based on the operational health score, perform cascade fault prediction on the power distribution nodes within the building to obtain fault warning information for the power distribution nodes within the building. S5. Based on the fault warning information and situational awareness map, perform multi-objective optimization on the basic monitoring and control instructions of the power distribution nodes in the building to obtain the remote monitoring and control instructions of the power distribution nodes in the building. S6. Based on remote monitoring and control commands, perform edge collaborative control on the execution terminals of the power distribution nodes in the building to obtain the monitoring and operation optimization results of the power distribution nodes in the building, and send the monitoring and operation optimization results to the cloud monitoring platform of the power distribution nodes in the building.

Images