A global perception intelligent monitoring cloud control method for ring main unit

Abstract

The application relates to the technical field of intelligent monitoring, in particular to a ring main unit (RMU) global perception intelligent monitoring cloud control method, which comprises the following steps: acquiring electrical operation environment and mechanical state data of the RMU, forming fusion data through noise reduction extraction and space-time splicing, mapping to a digital twin model to generate a virtual state, obtaining an insulation aging index and a leakage probability through time sequence analysis, determining a response level according to a diagnosis result and generating a control instruction, and issuing the control instruction to execute setting value setting and fault isolation adjustment. In the application, multi-source data is fused and mapped in the same space-time scale, a continuous diagnosis chain is formed by the cabinet state operation environment and the risk trend, insulation aging and leakage hidden danger are converted from static limit judgment to dynamic evaluation, cloud control is converted from a single alarm record to a quantitative grading result, false positives and false negatives are reduced, abnormal prediction remote control matching and fault disposal closed loop capability and accuracy are improved.

Description

Technical Field

[0001] This invention relates to the field of intelligent monitoring technology, and in particular to a cloud-based intelligent monitoring and control method for ring main units with full-area perception. Background Technology

[0002] The field of intelligent monitoring technology mainly involves the continuous collection, comparison, recording, and display of the operating status, environmental status, location information, temperature changes, current and voltage changes, abnormal alarm records, communication transmission records, and remote command execution process of the monitored object. Its core aspects include the deployment of sensing elements, the setting of sampling period, the configuration of threshold criteria, the collection of data by field terminals, the determination of status at the edge, the transmission of alarm information, the management of cloud ledgers, and the remote viewing and control command issuance. It is typically used for the operation inspection and status management of dispersed objects in scenarios such as power, industry, municipal administration, and transportation.

[0003] The traditional ring main unit (RMU) all-area perception intelligent monitoring and cloud control method refers to the status monitoring and remote control method for the RMU cabinet, busbar compartment, cable compartment, circuit breaker compartment, mechanism compartment, and surrounding operating environment. The technical aspects it addresses include cabinet temperature and humidity, partial discharge, busbar temperature, cable joint temperature, three-phase current, three-phase voltage, grounding current, switch open / closed position, energy storage status, cabinet door open / closed status, smoke, water immersion, condensation, and environmental video. Traditional methods typically involve installing temperature and humidity probes, wireless temperature tags, ultrasonic partial discharge probes, transient ground voltage probes, current transformers, voltage sampling terminals, magnetic door switches, and smoke detectors inside the RMU. The system includes water immersion probes and cameras. On-site data acquisition terminals read various values ​​at fixed sampling intervals. The read values ​​are compared with preset temperature limits, humidity limits, partial discharge amplitude limits, current over-limits, access control status, and water immersion status. The data is then uploaded to the cloud platform via wired Ethernet, RS485, fiber optic, or 4G links. The cloud platform compares the status based on rated current, allowable temperature rise, ambient humidity limits, partial discharge amplitude limits, cabinet door status tables, and switch position tables. It also records events according to preset alarm levels, generates maintenance work orders, and issues remote control commands such as fan start / stop, heater start / stop, dehumidifier start / stop, audible and visual alarm reset, and monitoring parameter adjustment.

[0004] Traditional cloud control methods rely on fixed sampling intervals and single-item threshold comparisons. Multi-source data from the field lacks a unified and correlated representation before being uploaded. Cloud-based ledgers focus on event recording and work order dispatch, making it difficult to present the spatial coupling relationship between the internal state of the cabinet and the operating environment. Alarm judgment is greatly affected by instantaneous fluctuations and isolated indicators, and there is a tendency for discrepancies to arise between the anomaly level and the actual risk. Remote control is mostly limited to routine operations such as fan heating, dehumidification, and audible and visual resets. There is insufficient basis for fault evolution trends and protection action adjustments, and there is a significant lag in inspection and handling. Summary of the Invention

[0005] To address the technical problems existing in the prior art, this invention provides a method for intelligent monitoring and cloud control of a ring main unit with full-domain perception, comprising the following steps: S1: Obtain electrical operating parameters, cabinet environmental parameters, and mechanical characteristic parameters collected by the sensing terminal inside the ring network cabinet, use wavelet transform function to reduce noise, and perform feature extraction and spatiotemporal dimension splicing analysis through edge computing nodes to construct multimodal sensing fusion data; S2: Call the digital twin model of the ring main unit, input the multimodal sensing fusion data into the digital twin model of the ring main unit, adjust the virtual space nodes of the equipment, calculate the spatial mapping association attributes between the virtual space nodes of the equipment and the multimodal sensing fusion data, and construct the virtual mapping status data of the equipment; S3: Input the virtual mapping state data of the device into the long short-term memory network for time-series feature dependency analysis, calculate the insulation aging index and leakage probability, and compare them with the normal operating range threshold to construct quantitative evaluation and diagnostic data; S4: Calculate the control response level of the cloud monitoring node based on the quantitative assessment and diagnostic data, select the corresponding control execution strategy, and construct hierarchical intelligent control instructions; S5: The hierarchical intelligent control command is sent to the ring network cabinet execution terminal through a two-way communication link to adjust the setpoint parameters and the fault isolation action status of the protection control execution unit.

[0006] As a further aspect of the present invention, the specific steps of S1 are as follows: S101: Collect electrical operating parameters, cabinet internal environmental parameters, and mechanical characteristic parameters from the ring network cabinet sensing terminal. Set the number of wavelet transform function decomposition levels. Input the electrical operating parameters, cabinet internal environmental parameters, and mechanical characteristic parameters into the wavelet transform function to extract high-frequency and low-frequency coefficients. Set a Gaussian noise hard threshold. Filter out high-frequency coefficients whose values ​​are lower than the Gaussian noise hard threshold. Perform inverse transform reconstruction calculation on the unfiltered high-frequency and low-frequency coefficients to generate a pure sensing parameter sequence. S102: For the pure sensing parameter sequence, the pure sensing parameter sequence is divided into equal-step sliding windows through edge computing nodes. The peak extrema and root mean square value of the pure sensing parameter sequence within the sliding window are calculated. The frequency domain energy density and spectral phase parameter are calculated by applying fast Fourier transform. The peak extrema, root mean square value, frequency domain energy density and spectral phase parameter are aggregated to establish a time-frequency attribute grid and obtain the state sensing feature matrix. S103: Call the state-aware feature matrix, read the global timestamp reference, compare the acquisition time of the state-aware feature matrix with the global timestamp reference to calculate the time offset, perform time dimension alignment on the parameters of the state-aware feature matrix based on the time offset, read the corresponding three-dimensional coordinates, establish a topological mapping on the time-aligned parameters based on the three-dimensional coordinates and splice the attribute dimensions to construct multimodal perception fusion data.

[0007] As a further aspect of the present invention, the specific steps of S2 are as follows: S201: Call the digital twin model of the ring main unit, input the multimodal sensing fusion data into the digital twin model of the ring main unit, extract the virtual space nodes of the equipment in the digital twin model of the ring main unit, read the distribution coordinate vector and boundary size threshold of the virtual space nodes of the equipment, aggregate the distribution coordinate vector and the boundary size threshold, and establish the spatial geometric topology; S202: Call the spatial geometric topology, extract the electrical operation parameters and mechanical characteristic parameters of the multimodal perception fusion data, compare the electrical operation parameters with the rated reference value to calculate the operation state deviation rate, compare the mechanical characteristic parameters with the boundary size threshold to calculate the mechanical deformation displacement, adjust the three-dimensional coordinates of the virtual space nodes of the equipment according to the operation state deviation rate and the mechanical deformation displacement, and generate a dynamically updated node array; S203: Extract the relocation coordinates of the dynamically updated node array and the sensing coordinates of the multimodal perception fusion data, calculate the difference between the relocation coordinates and the sensing coordinates to obtain the Euclidean distance, calculate the spatial mapping association attribute between the device virtual space node and the multimodal perception fusion data based on the Euclidean distance and the operating state deviation rate, aggregate the spatial mapping association attribute, and construct the device virtual mapping state data.

[0008] As a further aspect of the present invention, the process of calculating the operating state deviation rate by comparing the electrical operating parameters with the rated reference value specifically involves: obtaining the stable state electrical parameters recorded by the ring main unit during the standard operating condition test phase; calculating the average value of the accumulated stable state electrical parameters and calibrating the average value as the rated reference value; extracting the absolute difference obtained by subtracting the rated reference value from the electrical operating parameters; dividing the absolute difference by the rated reference value to obtain the calculation ratio, and setting the calculation ratio as the operating state deviation rate.

[0009] As a further aspect of the present invention, the specific steps of S3 are as follows: S301: Input the virtual mapping state data of the device into the Long Short-Term Memory Network, set the activation parameters of the input gate and forget gate of the Long Short-Term Memory Network, use the Long Short-Term Memory Network to perform time-series feature dependency analysis on the virtual mapping state data of the device, extract the weight matrix of the hidden layer state nodes inside the dynamic time window, map the temporal node association relationship according to the weight matrix, and generate a temporal dependency feature vector. S302: Extract the partial discharge feature subsequence and current leakage feature subsequence within the time-dependent feature vector, extract the amplitude change gradient of the partial discharge feature subsequence to calculate the insulation aging index, calculate the leakage probability based on the mutation frequency distribution characteristics of the current leakage feature subsequence, aggregate the insulation aging index and leakage probability, and establish an operational anomaly characterization attribute set; S303: Read the normal operating range threshold, perform cross-comparison between the insulation aging index and leakage probability in the abnormal operation characterization attribute set and the normal operating range threshold, determine whether the insulation aging index and leakage probability exceed the normal operating range threshold, calculate the numerical difference of the part exceeding the normal operating range threshold, associate the numerical difference with the over-limit judgment status identifier, and construct quantitative evaluation and diagnostic data.

[0010] As a further aspect of the present invention, the process of extracting the amplitude change gradient of the partial discharge feature subsequence to calculate the insulation aging index specifically involves performing a difference operation on the amplitude of adjacent nodes of the partial discharge feature subsequence to obtain multiple sets of amplitude change gradients; extracting the global amplitude peak value of the partial discharge feature subsequence; and calculating the ratio of the sum of the amplitude change gradients to the global amplitude peak value as the insulation aging index. The process of calculating the leakage probability based on the mutation frequency distribution characteristics of the current leakage characteristic subsequence specifically involves: statistically analyzing the frequency of mutations in the current leakage characteristic subsequence within a time window as the mutation frequency distribution characteristics; multiplying the mutation frequency distribution characteristics by the leakage evaluation coefficient to obtain the leakage probability, wherein the leakage evaluation coefficient is set based on the correlation mapping relationship between the rated voltage and the insulation dielectric impedance.

[0011] As a further aspect of the present invention, the specific steps of S4 are as follows: S401: Read the severity interval threshold preset by the cloud monitoring node, compare the abnormal state parameter inside the quantitative assessment and diagnosis data with the severity interval threshold, calculate the abnormal range span of the abnormal state parameter exceeding the severity interval threshold, map the alarm state and intervention level according to the abnormal range span, and establish a control response level. S402: Invoke the control response level, read the pre-plan rule base, retrieve the candidate operation set associated with the control response level in the pre-plan rule base, extract the equipment constraint parameters corresponding to the candidate operation set, perform cross-comparison between the quantitative evaluation and diagnostic data and the equipment constraint parameters, filter the state switching rules that meet the equipment constraint parameters, and construct the regulation and execution control strategy; S403: Parse the internal node control logic and communication instruction code of the regulation execution control strategy, read the issued protocol template, fill the node control logic and the communication instruction code into the data payload segment associated with the issued protocol template, integrate the timestamp and the data payload segment to execute the communication protocol encapsulation, and generate hierarchical intelligent regulation instructions.

[0012] As a further aspect of the present invention, the process of calculating the abnormal range span of the abnormal state parameter exceeding the severity interval threshold specifically involves: reading a preset limit load rate; calculating the ratio of the preset limit load rate to the benchmark load rate; setting the ratio as a safety margin coefficient; multiplying the rated insulation withstand voltage index and the maximum leakage current with the safety margin coefficient to obtain a target tolerance parameter; dividing the target tolerance parameter into multiple numerical intervals; setting the boundary values ​​of the multiple numerical intervals as the severity interval threshold; extracting the insulation aging value and leakage probability value within the abnormal state parameter; comparing the insulation aging value and the leakage probability value with the multiple numerical intervals to locate the numerical interval to which the insulation aging value and the leakage probability value belong; calculating the absolute difference between the insulation aging value and the leakage probability value and the boundary value of the numerical interval; and summing all the absolute differences to generate the abnormal range span.

[0013] As a further aspect of the present invention, the specific steps of S5 are as follows: S501: The hierarchical intelligent control command is sent to the ring network cabinet execution terminal using a two-way communication link. The communication parsing function of the ring network cabinet execution terminal is called to disassemble the data payload of the hierarchical intelligent control command, extract the set value adjustment control code and action switching identifier, verify the data integrity of the set value adjustment control code and action switching identifier, and generate command parsing parameters. S502: For the instruction parsing parameters, read the operating set value of the internal protection control execution unit of the ring network cabinet execution terminal, overwrite the set value setting control code with the operating set value execution value, calculate the over-limit protection bias, adjust the set value setting parameters of the protection control execution unit according to the over-limit protection bias, reset the over-limit trigger threshold range, and establish a set value update sequence; S503: Invoke the setpoint update sequence, parse the fault isolation circuit interface address associated with the action switching identifier, apply a level drive signal to the fault isolation circuit interface address according to the setpoint update sequence, perform fault isolation action state adjustment, collect circuit breaker position parameters, aggregate the circuit breaker position parameters and the setpoint update sequence, and construct the equipment protection control state.

[0014] As a further aspect of the present invention, the process of resetting the over-limit trigger threshold interval specifically involves: parsing the setting control code to extract the target setting parameter; calculating the difference between the target setting parameter and the operating setting value to obtain the over-limit protection bias; reading the environmental state compensation coefficient; performing a multiplication operation between the environmental state compensation coefficient and the over-limit protection bias to generate a corrected protection bias; extracting the ratio of the historical temperature offset value of the ring main unit execution terminal to the preset reference temperature value to set the environmental state compensation coefficient; reading the equipment operating reference value; performing an addition operation between the corrected protection bias and the equipment operating reference value to generate a unidirectional trigger threshold; extracting the peak data of the current envelope of the protection control execution unit to set the equipment operating reference value; reading the initial lower limit boundary parameter; performing data splicing between the unidirectional trigger threshold and the initial lower limit boundary parameter to reset the over-limit trigger threshold interval; designating the unidirectional trigger threshold as positive boundary data and combining it with the initial lower limit boundary parameter to set the over-limit trigger threshold interval.

[0015] Compared with the prior art, the advantages and positive effects of the present invention are as follows: In this invention, noise reduction and extraction of electrical operating environment and mechanical characteristic data are synchronously performed and spliced ​​according to spatiotemporal relationships to form a multi-source fusion expression that can be uniformly analyzed. The fusion result is mapped to a digital twin space to establish a correlation between physical states and virtual nodes. Time-series dependency analysis is used to infer insulation aging and leakage risks around the mapped states, enabling anomaly identification to shift from isolated thresholds to continuous trend assessment. The cloud determines the response level and selects control strategies based on the quantitative diagnostic results. Control commands are applied to the setting and fault isolation states via bidirectional links, improving the accuracy of state presentation, risk prediction capability, and timeliness of closed-loop handling, while reducing the interference of false alarms and missed alarms on remote control. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a schematic diagram of the steps of the present invention; Figure 2 This is a detailed schematic diagram of S1 of the present invention; Figure 3 This is a detailed schematic diagram of S2 of the present invention; Figure 4 This is a detailed schematic diagram of S3 of the present invention; Figure 5 This is a detailed schematic diagram of S4 of the present invention; Figure 6 This is a detailed schematic diagram of S5 of the present invention. Detailed Implementation

[0018] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0019] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0020] This embodiment provides a method for intelligent monitoring and cloud control of ring main units with full-area perception. In practical applications, such as during the continuous operation of a substation ring main unit carrying out the access, segmentation, connection, and protection control of power distribution lines, the sensing terminal inside the unit collects the operating status along the busbar compartment, cable compartment, mechanism compartment, and secondary control area. The edge computing node completes local preprocessing and feature organization, the cloud monitoring node completes diagnostic evaluation and control strategy generation, and the ring main unit execution terminal receives control commands and feeds back the execution status, thereby forming a closed loop of end-to-cloud collaborative monitoring and cloud control, including the following steps: Please see Figure 1 and Figure 2S1: Acquire electrical operating parameters, cabinet environmental parameters, and mechanical characteristic parameters collected by the sensing terminal inside the ring main unit. Electrical operating parameters include data categories reflecting the primary circuit's operating status such as current, voltage, partial discharge, and leakage current. Cabinet environmental parameters include data categories reflecting the internal environment of the cabinet such as temperature, humidity, condensation, and gas state. Mechanical characteristic parameters include data categories reflecting the actuator's operational status such as circuit breaker position, mechanism stroke, energy storage status, and contact status. The sensing terminal is a data acquisition end installed inside the ring main unit and communicating with the edge computing node. Its output includes the data acquisition source, acquisition time, installation location, and data category identifier. In this method, the wavelet transform function is limited to a signal processing process that decomposes, identifies, and reconstructs the sensing sequence. The processing object is a continuous operating sequence from the sensing terminal, and the output is a clean sensing parameter sequence after interference removal. The edge computing node is a processing node deployed near the ring main unit. It receives data packets uploaded by the sensing terminal, performs denoising, feature extraction, time alignment, spatial stitching, and anomaly identification recording, and then passes the processed data object to the subsequent digital twin processing process. Multimodal perception fusion data is a unified data object after being processed by edge computing nodes. It carries electrical operation characteristics, cabinet environment characteristics, mechanical motion characteristics, time reference information, spatial location relationship and source identification, which are read by the digital twin model in S2. S101: The sensing terminal of the ring main unit monitors electrical operating parameters, cabinet environmental parameters, and mechanical characteristic parameters. Before data collection, the sensing terminal completes device identity verification and acquisition channel status confirmation. The collected data first enters the receiving buffer of the edge computing node. The receiving buffer is used to temporarily store data frames from different sensing ports and retains data source, acquisition time, port status, and integrity verification records. If there is missing data, duplicate upload, or inconsistent data category identifiers, the edge computing node adds an abnormal status identifier to the corresponding data frame. Identifiable data continues to be sent to the cleaning process, while unresolvable data is left in the abnormal record for traceability and does not directly enter the subsequent fusion process.

[0021] Electrical operating parameters, cabinet environmental parameters, and mechanical characteristic parameters are input into a wavelet transform function. The wavelet transform function divides the sensing sequence into high-frequency and low-frequency components according to a preset decomposition hierarchy. High-frequency components carry information about rapid changes caused by sudden changes, impacts, local interference, and switching actions, while low-frequency components carry information about operating trends, slow environmental changes, and continuous changes in mechanical state. A Gaussian noise hard-judgment boundary is established based on noise background in the acquisition channel, equipment operating baselines, and self-test results of the sensing port. This boundary is used to identify high-frequency components that do not belong to valid state changes. High-frequency components identified as noise do not participate in reconstruction. The retained high-frequency components, along with the low-frequency components, enter the reverse reconstruction process to form a clean sensing parameter sequence. The clean sensing parameter sequence continues to retain source identifiers and abnormal state identifiers, enabling subsequent feature extraction to distinguish between real state changes and acquisition link anomalies. S102: For the pure sensing parameter sequence, the edge computing node segments the sequence using a sliding window with equal step size. In this method, the sliding window is defined as a time segment container for observing the local state of the continuous sensing sequence. Its function is to ensure that electrical, environmental, and mechanical states enter the feature extraction process at the same processing granularity. Within each sliding window segment, the edge computing node identifies the peak extremum and root mean square (RMS) value. The peak extremum characterizes the impact state change within the window, and the RMS value characterizes the overall energy level within the window. For windows with abnormal state indicators in the acquisition link, the edge computing node first checks whether the abnormality affects the continuity of the window. If it does, a reliable state indicator is added to the window, and it is restricted from entering the direct weight source for subsequent diagnosis.

[0022] Edge computing nodes apply Fast Fourier Transform (FFT) to obtain frequency domain energy density and spectral phase parameters. Frequency domain energy density reflects the energy distribution of state changes in the frequency domain, while spectral phase parameters reflect the phase relationship of state changes across different sensing channels. Peak extrema, root mean square (RMS) values, frequency domain energy density, and spectral phase parameters are aggregated into a time-frequency attribute grid. In this method, the time-frequency attribute grid is defined as a feature-carrying structure organized according to data category, time window, acquisition channel, and feature type. The state-aware feature matrix is ​​a feature object formed by channel alignment and field merging of the time-frequency attribute grid, and is output to S103 for time dimension alignment and spatial dimension stitching. S103: Invoke the state-aware feature matrix and read the global timestamp baseline. In this method, the global timestamp baseline is defined as a time reference maintained by the edge computing node and consistent with the cloud monitoring node, used to unify the data sequence of different sensing terminals. The edge computing node compares the acquisition time of the state-aware feature matrix with the global timestamp baseline, identifies the time offset status between each sensing channel, and performs time dimension alignment on the parameters of the state-aware feature matrix based on the offset status. If a sensing channel cannot complete time alignment, the edge computing node retains its source record and anomaly identifier, and treats the corresponding data of that channel as a pending confirmation state in subsequent mapping, rather than directly triggering control as a deterministic state.

[0023] The system reads the corresponding three-dimensional spatial coordinates. These coordinates represent the spatial positioning data between the installation location of the sensing terminal within the ring main unit and the location of the monitored component. This data originates from equipment installation and configuration, on-site debugging records, and the binding relationship between the sensing terminal and the device. Based on these three-dimensional coordinates, a topology mapping is established for time-aligned parameters. Electrical characteristics, environmental characteristics, mechanical characteristics, time stamps, spatial coordinates, source identifiers, and trusted status identifiers are then concatenated into a unified attribute dimension to construct multimodal sensing fusion data. This data is uploaded to the cloud monitoring node via an edge-cloud collaborative transmission mechanism, while traceable records are retained at the edge computing node. Subsequent readings of this data by the digital twin model directly reveal the status source, spatial location, and temporal relationship.

[0024] Please see Figure 1 and Figure 3 S2: Invoke the digital twin model of the ring main unit and input the multimodal sensing fusion data into the digital twin model of the ring main unit. In this method, the digital twin model of the ring main unit is defined as a virtual representation model corresponding to the primary conductive components, insulating components, mechanical transmission components, sensor placement positions, and execution terminal states of the ring main unit. The model contains virtual space nodes of the equipment, node connection relationships, boundary size determination rules, rated operating benchmarks, and historical state records. The virtual space nodes of the equipment are virtual node objects in the digital twin model corresponding to bus connection points, cable connection points, insulation isolation positions, switch mechanism positions, and protection control positions, carrying node coordinates, node types, associated components, and state fields. The virtual mapping state data of the equipment is a state object formed after the multimodal sensing fusion data is associated with the virtual space nodes of the equipment, which is used by S3 to perform time-series dependency analysis. S201: Invoke the digital twin model of the ring main unit. Input multimodal sensing fusion data into the digital twin model of the ring main unit. The input process first verifies the data source identifier, time alignment status, spatial coordinate field, and data category field. If the input data has missing spatial coordinates, the digital twin model completes the location identifier according to the binding relationship of the sensing terminal and the topological position in the cabinet, and writes the completion identifier into the mapping record. If the input data has source conflicts, the model retains the conflict state and waits for the edge computing node to return the verification information, without directly overwriting the existing node state with conflicting data.

[0025] The virtual space nodes of the ring main unit's digital twin model are extracted, and the distribution coordinate vectors and boundary size determination conditions of these virtual space nodes are read. In this method, the distribution coordinate vectors are limited to describing the position of the virtual nodes within the cabinet space, and the boundary size determination conditions are textual determination rules formed by equipment manufacturing and installation information, the relative positional relationships of components within the cabinet, and the allowable deformation range. The distribution coordinate vectors and boundary size determination conditions are aggregated to establish a spatial geometric topology. This spatial geometric topology is used to express the connections, proximity, isolation, and dependency relationships between virtual nodes. Subsequent steps adjust the node states based on this topology to avoid incorrectly mapping unconnected perceived data to unrelated components. S202: The spatial geometric topology is invoked to extract electrical operating parameters and mechanical characteristic parameters from the multimodal sensing fusion data. The electrical operating parameters are compared with rated reference values, which are derived from the steady-state electrical parameters in the standard operating condition record of the ring main unit and written into the digital twin model after equipment commissioning and configuration confirmation. In this method, the operating state deviation rate is defined as a state deviation characterization field formed by the electrical operating parameters relative to the rated reference values. This field indicates the degree of deviation between the current operating state and the steady-state state. It is formed by textual normalization of the difference relationship between the electrical operating parameters and the rated reference values.

[0026] Mechanical characteristic parameters are compared with boundary dimension judgment conditions to form mechanical deformation displacement. In this method, mechanical deformation displacement is defined as a relative change characterization field between mechanism stroke, contact position, switch state, and cabinet installation boundary, used to indicate whether the mechanical structure deviates from the acceptable action path. Based on the operating state deviation rate and mechanical deformation displacement, the three-dimensional coordinates of the equipment virtual space nodes are adjusted. The adjustment process does not change the component identity corresponding to the equipment virtual space node, but only updates its state position and associated state fields, generating a dynamically updated node array. The dynamically updated node array is a set of virtual node states organized according to spatial geometric topology, output to S203 for establishing a mapping association with the sensor coordinates; S203: Extract the relocation coordinates of the dynamically updated node array and the sensing coordinates of the multimodal sensing fusion data. The relocation coordinates are the coordinate fields after the virtual node state adjustment is completed in S202, and the sensing coordinates are the spatial coordinate fields of the sensing data formed in S103. The relocation coordinates and sensing coordinates are compared to obtain the spatial proximity state, and the spatial mapping association attributes between the virtual spatial nodes of the device and the multimodal sensing fusion data are identified by combining the operating state deviation rate. In this method, the spatial mapping association attributes are limited to association fields describing the virtual node, sensing source, state deviation, and spatial proximity relationship. Their contents include node identity, sensing channel source, spatial proximity state, operating deviation state, trusted state, and mapping record identifier.

[0027] Aggregate spatial mapping association attributes to construct device virtual mapping state data. This data is written to the current state cache via a digital twin model, and abnormal mappings, conflict mappings, and pending confirmation mappings are synchronously fed back to the edge computing nodes. If the same sensing data can be associated with adjacent virtual nodes, the digital twin model determines the primary mapping object based on the spatial geometric topology and sensing source priority, while retaining other associations as auxiliary mapping records. This process ensures that subsequent inputs received by the Long Short-Term Memory network have clear component orientation, state source, and spatial relationships.

[0028] Please see Figure 1 and Figure 4 S3: The virtual mapping state data of the device is input into a Long Short-Term Memory (LSTM) network for time-series feature dependency analysis. In this method, the LSM network is defined as a time-series analysis model consisting of an input gate, a forget gate, a memory state, an output gate, and a hidden state propagation structure. The input gate is used to select effective features from the current virtual mapping state data that enter the memory state. The forget gate is used to suppress historical information that is irrelevant to the current diagnosis or has insufficient credibility. The memory state is used to retain insulation changes, leakage changes, and mechanical state changes across time windows. The output gate is used to form the time-series dependency feature vector of the current time window. The insulation aging index is a diagnostic field characterizing the degree of insulation state change, and the leakage probability is a diagnostic field characterizing the leakage current state entering the leakage tendency category. The two are cross-compared with the normal operation judgment boundary to form quantitative evaluation diagnostic data. S301: The virtual mapping state data of the device is input into the Long Short-Term Memory (LSTM) network. Before input, it is organized into a time-series input sequence according to the virtual node identity, time order, state category, and trusted state identifier. The input gate activation parameters in this method are limited to gating rules that control the entry of the current input feature into the memory state, while the forget gate activation parameters are limited to gating rules that control the retention or decay of historical states. Both originate from parameter configurations and engineering operation verification records formed during the model training phase. During model training, historical operating states, manually labeled abnormal state records, protection action records, and maintenance confirmation records are read. Before the data enters the model, source verification, state labeling, abnormal record removal, and channel consistency processing are completed. During training, the input layer receives state data arranged chronologically. The gating structure adjusts the retention relationship between the current feature and historical states layer by layer. The output results are compared with the labeled states and preset judgment rules. Deviation feedback is used to adjust the direction of the gating parameters. After the model reaches a stable state for engineering judgment, it is used for online analysis.

[0029] A Long Short-Term Memory (LSTM) network is used to perform temporal feature dependency analysis on the virtual mapping state data of the device, extracting the weight matrix of hidden layer state nodes within the dynamic time window. In this method, the hidden layer state node weight matrix is ​​limited to the weight organization structure within the model used to express the time-transmission relationship of different virtual node states and is not output as an external numerical result. Based on the weight matrix, temporal node association relationships are mapped to generate temporal dependency feature vectors. These feature vectors carry information on partial discharge changes, leakage current changes, mechanical action continuity, environmental impact traces, and spatial node associations, and are output to S302 to form operational anomaly characterization attributes. S302: Extract partial discharge feature subsequences and current leakage feature subsequences from the time-dependent feature vector. The partial discharge feature subsequence is a time-continuous feature segment related to the state of insulating components, including amplitude variation trends, abrupt change states, and correlations with spatial nodes. The current leakage feature subsequence is a time-continuous feature segment related to the state of leakage channels, including leakage state changes, abrupt change density, and correlations with electrical operating states. If a feature subsequence has breaks, repetitions, or insufficiently reliable states, the Long Short-Term Memory (LSTM) network marks the corresponding segment as a state to be confirmed and synchronizes this state to the source field of the quantitative assessment and diagnostic data, preventing a single abnormal acquisition segment from directly triggering subsequent control.

[0030] The process of extracting the amplitude change gradient of partial discharge feature subsequences to calculate insulation aging indicators has been rewritten as a textual processing procedure. This involves differentially identifying the amplitude change relationship between adjacent state nodes in the partial discharge feature subsequence to form an amplitude change gradient set. This set is then combined with the global amplitude peak state of the subsequence, and the cumulative trend of gradient changes and peak reference relationship are merged into an insulation aging indicator. This indicator is not presented as a specific numerical value, but rather as a diagnostic field reflecting the transition of insulation state from stable to abnormal changes, and is included in the operational anomaly characterization attribute set.

[0031] The process of calculating leakage probability based on the frequency distribution characteristics of abrupt changes in the current leakage characteristic subsequence is rewritten into a textual processing procedure. This involves identifying the density of abrupt changes in the current leakage characteristic subsequence within a time window and combining this with leakage assessment coefficients to form the leakage probability. In this method, the leakage assessment coefficients are defined as assessment rules formed by the engineering mapping relationship between rated voltage state and insulation dielectric impedance state, sourced from ring main unit equipment configuration, insulation dielectric state records, and maintenance confirmation records. The leakage probability is included in the operational anomaly characterization attribute set and, together with insulation aging indicators, serves as the judgment object for S303. S303: Read the normal operation judgment boundary. In this method, the normal operation judgment boundary is limited to the engineering judgment rules formed by the cloud monitoring node based on the equipment's rated operating status, protection and control requirements, environmental compensation rules, and historical maintenance confirmation results. The insulation aging index and leakage probability within the abnormal operation characterization attribute set are cross-compared with the normal operation judgment boundary. The cross-comparison process first confirms the reliability of the diagnostic field's source, then confirms the operating status category to which the diagnostic field belongs, and finally identifies whether it has entered an out-of-limit state. The out-of-limit judgment state identifier is a status field that records whether the diagnostic field has entered the abnormal category, the pending confirmation category, or the normal category.

[0032] For portions exceeding the normal operation judgment boundary, the degree of abnormal deviation is expressed using a textual differential field, without outputting any specific numerical value. After associating the differential field with the out-of-limit judgment status identifier, quantitative assessment and diagnostic data is constructed. In this method, the quantitative assessment and diagnostic data is limited to diagnostic objects for cloud monitoring nodes to define response levels, including virtual node identity, anomaly category, deviation degree fields, data source, credibility status, temporal relationship, and traceability records. If the model output is abnormal, the gating result cannot be categorized, or the credibility status of the input sequence is insufficient, the quantitative assessment and diagnostic data is marked as pending review and fed back to the edge computing node to request supplementary sensing data or re-execution of the pre-processing steps S1 to S2.

[0033] Please see Figure 1 and Figure 5S4: The process of calculating the control response level of the cloud monitoring node based on the quantitative assessment and diagnostic data is completed using a textual assessment method. The cloud monitoring node is a cloud-side processing node that receives virtual mapping status data and quantitative assessment and diagnostic data from the device and generates control strategies. The control response level consists of status fields describing the alarm status, intervention level, and control constraint relationships. The control execution strategy is a set of control rules selected and verified by the cloud monitoring node from the pre-plan rule base. The hierarchical intelligent control instruction is an instruction object encapsulated according to the communication protocol and sent to the ring network cabinet execution terminal, carrying the node control logic, communication instruction code, execution constraints, time identifier, and verification information. S401: Read the severity interval judgment boundary preset by the cloud monitoring node, compare the abnormal state parameters within the quantitative assessment and diagnostic data with the severity interval judgment boundary, and form the abnormal range span. In this method, the severity interval judgment boundary is limited to the operational severity classification rules formed based on preset limit load rate, reference load rate, rated insulation withstand voltage state, and maximum leakage current state. The safety margin coefficient is a safety margin characterization field formed by the relationship between the limit load state and the reference load state. The target tolerance parameter is an acceptable state boundary field formed by combining the rated insulation withstand voltage state and the maximum leakage current state with the safety margin.

[0034] The insulation aging value field and leakage probability field of the abnormal state parameter are extracted and compared with the severity interval judgment boundary to determine the state interval category into which they belong. In this method, the abnormal range is limited to the deviation characterization field of the abnormal state parameter relative to its assigned interval boundary. Based on the deviation characterization field, alarm status and intervention level are mapped to establish a control response hierarchy. If the quantitative assessment and diagnostic data includes a pending review status, the cloud monitoring node does not directly enter the high-intensity control path, but instead generates a review prompt and supplementary data collection request, and returns this request to the edge computing node via the end-cloud collaborative link. S402: Invoke the control response level and read the contingency plan rule base. In this method, the contingency plan rule base is limited to the set of control rules stored by the cloud monitoring node. The rule content originates from the ring main unit's protection configuration, operation and maintenance procedures, equipment constraints, communication protocols, and historical handling confirmation records. The candidate operation set is the set of executable operations associated with the current control response level in the contingency plan rule base, including control action categories such as setting adjustment, alarm transmission, remote interlocking, fault isolation preparation, and execution status verification. The equipment constraint parameters are constraint fields describing whether the ring main unit's execution terminal, protection control actuator, circuit breaker status, communication link status, and current operating mode support the candidate operations.

[0035] The quantitative assessment and diagnostic data are cross-compared with the equipment constraint parameters to filter state switching rules that meet the equipment constraint parameters, and to construct a control strategy for regulation and execution. The filtering process first excludes candidate operations that conflict with the current circuit breaker position, protection interlocking status, or communication confirmation status, and then retains state switching rules that are consistent with the control response level, anomaly category, and execution terminal permissions. If the set of candidate operations cannot meet the equipment constraints, the cloud monitoring node generates a prohibited execution status and a manual review prompt, and writes the reason for non-execution into the traceability record. When the control strategy for regulation and execution is output to S403, it already includes the node control logic that can be issued, execution conditions, rollback conditions, and confirmation requirements. S403: Parse the internal node control logic and communication command code of the control and regulation execution strategy. In this method, the node control logic is limited to logical fields describing the target virtual node, target execution terminal, control action category, execution sequence, and confirmation conditions; the communication command code is the command field matching the communication protocol of the ring network cabinet execution terminal. Read the issued protocol template and fill the node control logic and communication command code into the associated data payload segment of the issued protocol template. The issued protocol template is a data encapsulation format agreed upon between the cloud monitoring node and the ring network cabinet execution terminal, including command type, target address, control fields, verification fields, time stamp, and acknowledgment requirements.

[0036] The system integrates timestamps and data payload segments to encapsulate communication protocols and generate hierarchical intelligent control commands. These commands undergo permission and integrity checks before transmission; only those that pass the checks enter the bidirectional communication link. If the link fails, the execution terminal does not acknowledge the command, or the command parsing status is abnormal, the cloud monitoring node retains the command status and triggers a retransmission control or manual review path, not treating unacknowledged commands as executed. This stage utilizes end-cloud collaboration between cloud policy generation and edge status feedback to ensure consistency between control commands and the actual cabinet status.

[0037] Please see Figure 1 and Figure 6 S5: The hierarchical intelligent control commands are sent to the ring main unit execution terminal via a two-way communication link to adjust the setpoint parameters and fault isolation action status of the protection control actuator. In this method, the two-way communication link is limited to the communication path between the cloud monitoring node, the edge computing node, and the ring main unit execution terminal for command issuance, feedback, status verification, and anomaly alarms. The ring main unit execution terminal is the control end installed in the secondary control area of ​​the cabinet and connected to the protection control actuator. After receiving the hierarchical intelligent control commands, it completes parsing, verification, setpoint update, action driving, and status feedback. The equipment protection control status is the closed-loop status object output by S5, which is used by the cloud monitoring node to update diagnostic records and for subsequent maintenance traceability. S501: The hierarchical intelligent control command is sent to the ring main unit execution terminal via a two-way communication link. After receiving the command, the communication parsing function of the ring main unit execution terminal first checks the command source, protocol format, target address, time stamp, and integrity verification field. In this method, the communication parsing function is limited to the parsing process within the execution terminal used to disassemble the command data payload and extract the control fields. Its input is the hierarchical intelligent control command, and its output is the command parsing parameters. The setting control code is a control field indicating the adjustment direction and target field of the protection setting parameters, and the action switching identifier is a control field indicating the action status of the fault isolation circuit.

[0038] The hierarchical intelligent control command data payload is disassembled, the setpoint control code and action switching identifier are extracted, the integrity of the two data is verified, and command parsing parameters are generated. The command parsing parameters carry the target protection parameters, target execution address, action switching status, verification results, and acknowledgment requirements. If the command source fails verification, the data payload cannot be parsed, the target address is inconsistent with the execution terminal, or the action switching identifier conflicts with the current blocking status, the execution terminal generates a refusal to execute status and feeds back to the cloud monitoring node through a two-way communication link. Subsequently, S502 and S503 will not enter the action execution path. S502: For the instruction parsing parameters, read the operating setpoint of the protection control actuator inside the ring main unit's execution terminal. The operating setpoint is the current setting status field used for protection control by the execution terminal, derived from local configuration, the last confirmation receipt, and the current status record of the protection control actuator. The setpoint setting control code is overwritten and updated with the operating setpoint to form an over-limit protection bias. In this method, the over-limit protection bias is limited to a field representing the adjustment direction and adjustment magnitude between the target setting parameter and the existing operating setpoint, used to reset the protection trigger boundary, and is not output in specific numerical form.

[0039] The process of resetting the over-limit trigger judgment interval is as follows: First, the setpoint control code is parsed to extract the target setpoint parameters. The difference between the target setpoint parameters and the operating setpoint is identified to form the over-limit protection bias. Then, the environmental state compensation coefficient is read and used to correct the over-limit protection bias, generating a corrected protection bias. The environmental state compensation coefficient is a compensation rule formed by the relationship between the historical temperature offset state of the ring main unit's execution terminal and the preset reference temperature state, derived from the cabinet's environmental perception records and operating configuration. Next, the equipment operating reference value is read, and the corrected protection bias is merged with the equipment operating reference value to form a unidirectional trigger boundary. The equipment operating reference value originates from the peak data of the current envelope of the protection control actuator. Finally, the initial lower limit boundary parameters are read, and the unidirectional trigger boundary is combined with the initial lower limit boundary parameters to reset the over-limit trigger judgment interval. After this process, the setpoint setting parameters of the protection control actuator are updated, forming a setpoint update sequence. The setpoint update sequence includes the update target, update source, compensation rule, boundary state, and receipt requirements, and is output to S503. S503: Invoke the setting update sequence and parse the fault isolation circuit interface address associated with the action switching identifier. In this method, the fault isolation circuit interface address is limited to the interface location information of the circuit breaker's opening and closing circuit, interlocking circuit, and status acquisition circuit within the ring main unit's execution terminal. Apply a level drive signal to the fault isolation circuit interface address according to the setting update sequence to perform fault isolation action status adjustment. In this method, the level drive signal is limited to a control signal output from the execution terminal to the fault isolation circuit. Before signal generation, it is necessary to confirm that the setting update sequence is valid, the action switching identifier is consistent with the current circuit breaker status, the protection interlocking condition has not been triggered, and the communication receipt requirements are met.

[0040] The system collects circuit breaker position change parameters, which reflect the position and circuit status information of the circuit breaker as it transitions from its original state to the target state. These parameters originate from the circuit breaker's auxiliary contacts, the mechanism status acquisition terminal, and the execution terminal's readback channel. The circuit breaker position change parameters and setting update sequences are aggregated to construct the equipment protection and control status. This status is transmitted back to the cloud monitoring node via a bidirectional communication link and synchronized to the edge computing node, forming an end-cloud collaborative closed-loop record. If the circuit breaker position change parameters do not match the action switching identifier, the execution terminal maintains the current protection status, records the unconfirmed status, and sends feedback to the cloud monitoring node for review. If the transmission link fails, the execution terminal retains the local receipt record and retransmits it after the link is restored, enabling continuous traceability of monitoring, diagnosis, control, and execution status within the same ring network cabinet's cloud control link.

[0041] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of protection of the described technical solutions.

Claims

1. A method for intelligent monitoring and cloud control of a ring main unit with full-area perception, characterized in that, Includes the following steps: S1: Obtain electrical operating parameters, cabinet environmental parameters, and mechanical characteristic parameters collected by the sensing terminal inside the ring network cabinet, use wavelet transform function to reduce noise, and perform feature extraction and spatiotemporal dimension splicing analysis through edge computing nodes to construct multimodal sensing fusion data; S2: Call the digital twin model of the ring main unit, input the multimodal sensing fusion data into the digital twin model of the ring main unit, adjust the virtual space nodes of the equipment, calculate the spatial mapping association attributes between the virtual space nodes of the equipment and the multimodal sensing fusion data, and construct the virtual mapping status data of the equipment; S3: Input the virtual mapping state data of the device into the long short-term memory network for time-series feature dependency analysis, calculate the insulation aging index and leakage probability, and compare them with the normal operating range threshold to construct quantitative evaluation and diagnostic data; S4: Calculate the control response level of the cloud monitoring node based on the quantitative assessment and diagnostic data, select the corresponding control execution strategy, and construct hierarchical intelligent control instructions; S5: The hierarchical intelligent control command is sent to the ring network cabinet execution terminal through a two-way communication link to adjust the setpoint parameters and the fault isolation action status of the protection control execution unit.

2. The ring main unit's all-domain perception intelligent monitoring and cloud control method according to claim 1, characterized in that, The specific steps of S1 are as follows: S101: Collect electrical operating parameters, cabinet internal environmental parameters, and mechanical characteristic parameters from the ring network cabinet sensing terminal. Set the number of wavelet transform function decomposition levels. Input the electrical operating parameters, cabinet internal environmental parameters, and mechanical characteristic parameters into the wavelet transform function to extract high-frequency and low-frequency coefficients. Set a Gaussian noise hard threshold. Filter out high-frequency coefficients whose values ​​are lower than the Gaussian noise hard threshold. Perform inverse transform reconstruction calculation on the unfiltered high-frequency and low-frequency coefficients to generate a pure sensing parameter sequence. S102: For the pure sensing parameter sequence, the pure sensing parameter sequence is divided into equal-step sliding windows through edge computing nodes. The peak extrema and root mean square value of the pure sensing parameter sequence within the sliding window are calculated. The frequency domain energy density and spectral phase parameter are calculated by applying fast Fourier transform. The peak extrema, root mean square value, frequency domain energy density and spectral phase parameter are aggregated to establish a time-frequency attribute grid and obtain the state sensing feature matrix. S103: Call the state-aware feature matrix, read the global timestamp reference, compare the acquisition time of the state-aware feature matrix with the global timestamp reference to calculate the time offset, perform time dimension alignment on the parameters of the state-aware feature matrix based on the time offset, read the corresponding three-dimensional coordinates, establish a topological mapping on the time-aligned parameters based on the three-dimensional coordinates and splice the attribute dimensions to construct multimodal perception fusion data.

3. The ring main unit's all-domain perception intelligent monitoring and cloud control method according to claim 1, characterized in that, The specific steps of S2 are as follows: S201: Call the digital twin model of the ring main unit, input the multimodal sensing fusion data into the digital twin model of the ring main unit, extract the virtual space nodes of the equipment in the digital twin model of the ring main unit, read the distribution coordinate vector and boundary size threshold of the virtual space nodes of the equipment, aggregate the distribution coordinate vector and the boundary size threshold, and establish the spatial geometric topology; S202: Call the spatial geometric topology, extract the electrical operation parameters and mechanical characteristic parameters of the multimodal perception fusion data, compare the electrical operation parameters with the rated reference value to calculate the operation state deviation rate, compare the mechanical characteristic parameters with the boundary size threshold to calculate the mechanical deformation displacement, adjust the three-dimensional coordinates of the virtual space nodes of the equipment according to the operation state deviation rate and the mechanical deformation displacement, and generate a dynamically updated node array; S203: Extract the relocation coordinates of the dynamically updated node array and the sensing coordinates of the multimodal perception fusion data, calculate the difference between the relocation coordinates and the sensing coordinates to obtain the Euclidean distance, calculate the spatial mapping association attribute between the device virtual space node and the multimodal perception fusion data based on the Euclidean distance and the operating state deviation rate, aggregate the spatial mapping association attribute, and construct the device virtual mapping state data.

4. The ring main unit's all-domain perception intelligent monitoring and cloud control method according to claim 3, characterized in that, The process of comparing the electrical operating parameters with the rated reference value to calculate the operating state deviation rate specifically involves: obtaining the stable state electrical parameters recorded by the ring main unit during the standard operating condition test phase; calculating the average value of the accumulated stable state electrical parameters and calibrating the average value as the rated reference value; extracting the absolute difference obtained by subtracting the rated reference value from the electrical operating parameters; dividing the absolute difference by the rated reference value to obtain the calculation ratio, and setting the calculation ratio as the operating state deviation rate.

5. The ring main unit's full-domain perception intelligent monitoring and cloud control method according to claim 1, characterized in that, The specific steps for S3 are as follows: S301: Input the virtual mapping state data of the device into the Long Short-Term Memory Network, set the activation parameters of the input gate and forget gate of the Long Short-Term Memory Network, use the Long Short-Term Memory Network to perform time-series feature dependency analysis on the virtual mapping state data of the device, extract the weight matrix of the hidden layer state nodes inside the dynamic time window, map the temporal node association relationship according to the weight matrix, and generate a temporal dependency feature vector. S302: Extract the partial discharge feature subsequence and current leakage feature subsequence within the time-dependent feature vector, extract the amplitude change gradient of the partial discharge feature subsequence to calculate the insulation aging index, calculate the leakage probability based on the mutation frequency distribution characteristics of the current leakage feature subsequence, aggregate the insulation aging index and leakage probability, and establish an operational anomaly characterization attribute set; S303: Read the normal operating range threshold, perform cross-comparison between the insulation aging index and leakage probability in the abnormal operation characterization attribute set and the normal operating range threshold, determine whether the insulation aging index and leakage probability exceed the normal operating range threshold, calculate the numerical difference of the part exceeding the normal operating range threshold, associate the numerical difference with the over-limit judgment status identifier, and construct quantitative evaluation and diagnostic data.

6. The ring main unit's all-domain perception intelligent monitoring and cloud control method according to claim 5, characterized in that, The process of extracting the amplitude change gradient of the partial discharge feature subsequence to calculate the insulation aging index specifically involves performing a difference operation on the amplitude of adjacent nodes of the partial discharge feature subsequence to obtain multiple sets of amplitude change gradients; extracting the global amplitude peak of the partial discharge feature subsequence; and calculating the ratio of the sum of the amplitude change gradients to the global amplitude peak as the insulation aging index. The process of calculating the leakage probability based on the mutation frequency distribution characteristics of the current leakage characteristic subsequence specifically involves: statistically analyzing the frequency of mutations in the current leakage characteristic subsequence within a time window as the mutation frequency distribution characteristics; multiplying the mutation frequency distribution characteristics by the leakage evaluation coefficient to obtain the leakage probability, wherein the leakage evaluation coefficient is set based on the correlation mapping relationship between the rated voltage and the insulation dielectric impedance.

7. The ring main unit's full-domain perception intelligent monitoring and cloud control method according to claim 1, characterized in that, The specific steps of S4 are as follows: S401: Read the severity interval threshold preset by the cloud monitoring node, compare the abnormal state parameter inside the quantitative assessment and diagnosis data with the severity interval threshold, calculate the abnormal range span of the abnormal state parameter exceeding the severity interval threshold, map the alarm state and intervention level according to the abnormal range span, and establish a control response level. S402: Invoke the control response level, read the pre-plan rule base, retrieve the candidate operation set associated with the control response level in the pre-plan rule base, extract the equipment constraint parameters corresponding to the candidate operation set, perform cross-comparison between the quantitative evaluation and diagnostic data and the equipment constraint parameters, filter the state switching rules that meet the equipment constraint parameters, and construct the regulation and execution control strategy; S403: Parse the internal node control logic and communication instruction code of the regulation execution control strategy, read the issued protocol template, fill the node control logic and the communication instruction code into the data payload segment associated with the issued protocol template, integrate the timestamp and the data payload segment to execute the communication protocol encapsulation, and generate hierarchical intelligent regulation instructions.

8. The ring main unit's all-domain perception intelligent monitoring cloud control method according to claim 7, characterized in that, The process of calculating the abnormal range span of the abnormal state parameter exceeding the severity interval threshold specifically involves: reading a preset limit load rate; calculating the ratio of the preset limit load rate to the reference load rate; setting the ratio as a safety margin coefficient; multiplying the rated insulation withstand voltage index and the maximum leakage current with the safety margin coefficient to obtain a target tolerance parameter; dividing the target tolerance parameter into multiple numerical intervals; setting the boundary values ​​of the multiple numerical intervals as the severity interval threshold; extracting the insulation aging value and leakage probability value within the abnormal state parameter; comparing the insulation aging value and the leakage probability value with the multiple numerical intervals to locate the numerical interval to which the insulation aging value and the leakage probability value belong. Calculate the absolute difference between the insulation aging value and the leakage probability value and the boundary value of the value range, and perform a summation operation on all the absolute differences to generate the abnormal range span.

9. The ring main unit's all-domain perception intelligent monitoring and cloud control method according to claim 1, characterized in that, The specific steps of S5 are as follows: S501: The hierarchical intelligent control command is sent to the ring network cabinet execution terminal using a two-way communication link. The communication parsing function of the ring network cabinet execution terminal is called to disassemble the data payload of the hierarchical intelligent control command, extract the set value adjustment control code and action switching identifier, verify the data integrity of the set value adjustment control code and action switching identifier, and generate command parsing parameters. S502: For the instruction parsing parameters, read the operating set value of the internal protection control execution unit of the ring network cabinet execution terminal, overwrite the set value setting control code with the operating set value execution value, calculate the over-limit protection bias, adjust the set value setting parameters of the protection control execution unit according to the over-limit protection bias, reset the over-limit trigger threshold range, and establish a set value update sequence; S503: Invoke the setpoint update sequence, parse the fault isolation circuit interface address associated with the action switching identifier, apply a level drive signal to the fault isolation circuit interface address according to the setpoint update sequence, perform fault isolation action state adjustment, collect circuit breaker position parameters, aggregate the circuit breaker position parameters and the setpoint update sequence, and construct the equipment protection control state.

10. The ring main unit's all-domain perception intelligent monitoring cloud control method according to claim 9, characterized in that, The process of resetting the over-limit trigger threshold range specifically involves parsing the set value setting control code to extract the target setting parameter, calculating the difference between the target setting parameter and the running set value to obtain the over-limit protection bias; Read the environmental state compensation coefficient, use the environmental state compensation coefficient and the over-limit protection bias to perform a multiplication operation to generate a corrected protection bias, and extract the ratio of the historical temperature offset value of the ring network cabinet execution terminal to the preset reference temperature value to set the environmental state compensation coefficient. Read the device operating reference value, add the corrected protection bias to the device operating reference value to generate a unidirectional trigger threshold, extract the peak data of the current envelope of the protection control execution unit to set the device operating reference value; read the initial lower limit boundary parameter, concatenate the unidirectional trigger threshold with the initial lower limit boundary parameter to reset the over-limit trigger threshold range, designate the unidirectional trigger threshold as positive boundary data and combine it with the initial lower limit boundary parameter to set the over-limit trigger threshold range.

Images