A power distribution system digital twin update method, system, device, and medium

Abstract

The application relates to a power distribution system digital twin updating method, system, device and medium, and belongs to the technical field of power distribution monitoring. Each intelligent monitoring node in a power distribution network reports steady-state characteristic quantities to a regional coordinator and a digital twin master station in a low-frequency mode, and listens to broadcasts of neighbor nodes. When a certain intelligent monitoring node determines that a certain steady-state characteristic quantity of the node exceeds a dynamic event triggering threshold, it checks whether multiple associated characteristics are abnormal within the same time window. If yes, the node determines that an effective fault event has occurred, and marks the node as an event source node. The event source node switches a data reporting mode from a low-frequency mode to a high-frequency mode, and generates an event warning broadcast message, which is sent to all known neighbor nodes at maximum power. Neighbor nodes that receive the event warning broadcast message enter a cooperative state, and improve a data reporting frequency of the nodes to a cooperative medium-frequency mode. The digital twin master station fuses high-frequency data and medium-frequency data, and reconstructs an animation of a whole process of event occurrence and development.

Description

Technical Field

[0001] This application relates to the technical field of power distribution monitoring, and in particular to a method, system, device and medium for updating a digital twin of a power distribution system. Background Technology

[0002] With the development of power systems, the scale and complexity of distribution networks are constantly increasing, placing higher demands on the operation monitoring and fault handling of distribution networks. Digital twin technology, as an emerging technology, provides a new means for the management and operation and maintenance of distribution networks by mapping the operating status of physical systems to virtual space.

[0003] Traditional digital twins rely on a master station periodically polling and collecting data (such as voltage, current, and switch status) from various monitoring points (e.g., FTUs, DTUs, and smart meters) for updates, with update cycles typically on the order of seconds or even minutes. Due to the lag in twin data updates, when rapid transient events occur in the distribution network (such as faults or switching operations), maintenance personnel still see a static snapshot from several seconds ago on the master station screen. This results in high data update latency and loss of timing information, preventing the twin from accurately mirroring the dynamic processes of the physical system in real time. Summary of the Invention

[0004] In order to enable the digital twin to mirror the dynamic process of the physical system in real time and with high fidelity, this application provides a method, system, device and medium for updating the digital twin of a power distribution system.

[0005] Firstly, this application provides a method for updating a digital twin of a power distribution system, employing the following technical solution: A method for updating a digital twin of a power distribution system includes: Each smart monitoring node in the distribution network reports steady-state characteristic quantities to the regional coordinator and digital twin master station at low frequency, and listens to the broadcasts of neighboring nodes; When a smart monitoring node determines that a certain steady-state characteristic quantity of itself exceeds the dynamic event triggering threshold, it checks whether multiple related characteristic anomalies are met simultaneously within the same time window. If yes, it is determined to be a valid fault event, and the corresponding intelligent monitoring node is marked as the event source node; otherwise, it is determined to be interference. The event source node switches the data reporting mode from low-frequency mode to high-frequency mode and generates an event warning broadcast message, which is sent to all known neighbor nodes at maximum power. Neighboring nodes that receive the event warning broadcast message enter a collaborative state and increase their own data reporting frequency to a collaborative medium frequency mode; The digital twin master station merges high-frequency data from the event source node with mid-frequency data from the collaborating node to reconstruct the entire process of the event's occurrence and development.

[0006] By adopting the above technical solutions, from an overall operational perspective, the low-frequency reporting mode in normal operation reduces the data transmission pressure on the distribution network communication links, reduces the energy consumption of intelligent monitoring nodes, and also alleviates the data processing load on the regional coordinator and the digital twin master station, allowing the system to maintain a highly efficient and low-cost state during steady-state operation. When a fault occurs, the fault judgment logic based on multi-feature correlation verification can filter out interference false alarms caused by single-feature anomalies, improving the accuracy of fault event identification and avoiding the waste of resources caused by invalid responses to system operation. The high-frequency reporting switching of the event source node and the coordinated medium-frequency response of neighboring nodes enable the digital twin master station to obtain high-density, high-timeliness data of the fault area in the first instance, quickly complete the replication and dynamic update of the fault scenario, not only shortening the time for fault location and handling, but also allowing maintenance personnel to simulate fault handling plans in advance through the digital twin, reducing the risk of on-site operations. Based on the topological neighbor-based coordinated response mechanism, the monitoring nodes in the fault area form a local, efficient data acquisition network, ensuring the comprehensiveness of fault data without affecting the normal communication of non-faulty areas.

[0007] Optionally, the digital twin update method further includes: When the event source node or cooperating node sends a data packet, it listens to whether the communication channel is idle; If not, then the node calculates a dynamic priority score for the data packet to be sent; The node calculates the backoff time based on the dynamic priority score; The node with the shorter backoff time has priority to obtain channel access, while the other nodes wait for the backoff time before sending.

[0008] By adopting the above technical solutions, when a fault occurs, multiple monitoring nodes simultaneously trigger data reporting or broadcasting, which can easily lead to communication channel congestion, data packet conflicts, or even data loss. This mechanism, by monitoring the channel status in real time, fundamentally avoids the resource waste caused by blind transmission. Dynamic priority score calculation grants priority transmission rights to data packets containing core fault information, ensuring that the digital twin master station can obtain the most critical fault data immediately, buying valuable time for subsequent fault analysis and handling. Priority-based backoff time allocation makes the resolution of communication conflicts more efficient and orderly, preventing nodes from getting bogged down in meaningless repeated transmission attempts. This reduces the waste of communication resources and lowers node energy consumption, ensuring the stability of the entire communication network. This flexible channel scheduling method can also adapt to distribution network topologies of different sizes and complexities. Whether it's a simple linear network or a complex ring network, it can quickly adjust communication strategies to ensure the timeliness and integrity of fault data transmission, ultimately enabling the digital twin to more accurately and dynamically synchronize the actual fault status of the distribution network.

[0009] Optionally, the digital twin update method further includes: When the event source node is unable to establish a communication connection with the regional coordinator or digital twin master station, it uses the local peer-to-peer network to exchange the simplest state information with the directly adjacent neighbor node whose communication is interrupted. Each participating node determines the inference conclusion of the faulty section based on the exchanged simplest state information and preset physical rules; A representative node is determined from the participating nodes. When communication is restored, the representative node reports the final conclusion of the inferred fault section and the evidence data to the digital twin master station.

[0010] By adopting the above technical solutions, communication link interruptions are unavoidable emergencies in the actual operation of power distribution networks. Once the event source node loses connection with the upper-level regional coordinator or digital twin master station, the original fault data reporting path will be completely interrupted, and the digital twin will be unable to synchronize the real state of the fault area, thus affecting the accuracy of operation and maintenance decisions. The emergency mechanism of the local peer-to-peer network ensures that nodes in the network outage state are no longer isolated and helpless. By exchanging the simplest state information with neighboring nodes and combining it with preset physical rules for distributed fault segment inference, it can autonomously complete the preliminary location of the fault range without upper-level control. The setting of representative nodes allows the dispersed inference results to be integrated, and key data is reported all at once after communication is restored. This ensures the integrity and consistency of the fault information ultimately obtained by the digital twin master station, while avoiding network congestion caused by a large number of nodes reporting simultaneously after communication is restored. This distributed emergency response approach not only utilizes the edge computing capabilities of intelligent monitoring nodes, reducing reliance on upper-layer communication links, but also maintains minimal flow and analysis of fault information in extreme network outage scenarios. This ensures that the digital twin can quickly complete the fault scenario update after communication is restored, and makes the operation of the entire power distribution system's digital twin more closely resemble the complex and ever-changing actual working conditions of the power distribution network.

[0011] Optionally, the specific steps for determining the representative node from the participating nodes include: All participating nodes broadcast their own inferences and signal strength; Each parameter node compares the signal strength received from all other participating nodes; The participating node with the strongest signal strength is selected as the representative node.

[0012] By adopting the above technical solution, and by having all participating nodes broadcast their inferences and signal strengths, each node can obtain key status information from other nodes, providing a unified and objective basis for selecting the representative node. Using the node with the strongest signal strength as the representative node considers both the node's communication capabilities in the local peer-to-peer network (nodes with strong signals tend to establish connections with the digital twin master station faster and more stably after communication is restored, ensuring fault information is uploaded immediately) and the avoidance of selection biases caused by differences in node computing and storage capabilities. After all, in network outage emergency scenarios, communication reliability is a core element for ensuring information reporting. This signal strength-based selection method does not require additional complex negotiation mechanisms; each node only needs to independently compare signal strengths to determine the representative node. This reduces the communication interaction costs between nodes and minimizes potential conflicts and delays during the negotiation process, making the entire representative node determination process efficient and orderly.

[0013] Optionally, the digital twin update method further includes: When the event source node detects that all feature quantities are continuously lower than the corresponding dynamic recovery threshold, it broadcasts an event end message and switches the reporting mode from high-frequency mode back to low-frequency mode. The neighboring node that receives the event end message stops the cooperative intermediate frequency reporting mode.

[0014] By adopting the above technical solution, the broadcast of the event end message allows all participating neighbor nodes to synchronously perceive the end of the fault, promptly stop the collaborative intermediate frequency reporting mode, and jointly return to the daily low frequency reporting state. This not only quickly releases the communication and node computing resources occupied by the fault response, allowing the entire distribution network's communication network and node equipment to return to a highly efficient and energy-saving steady-state operation mode, but also avoids network congestion and resource waste caused by some nodes continuously maintaining high-frequency or intermediate frequency reporting, ensuring that system resources can be dynamically allocated as needed.

[0015] Optionally, the calculation steps for the steady-state characteristic quantities corresponding to the intelligent monitoring nodes include: The intelligent monitoring node samples the electrical quantities at its location at a fixed frequency. The electrical quantities include the instantaneous values ​​of the three-phase voltage and the instantaneous values ​​of the three-phase current. The intelligent monitoring node calculates the RMS current, RMS voltage, and zero-sequence current of the current power frequency cycle based on the sampled data; Calculate the rate of change of the effective value of the current relative to the previous power frequency cycle.

[0016] By adopting the above technical solution and sampling the instantaneous values ​​of three-phase voltage and three-phase current at a fixed frequency, the real-time electrical status of distribution network nodes can be comprehensively captured, avoiding characteristic quantity deviations caused by missing sampling dimensions. Based on this, the calculated RMS current, RMS voltage, and zero-sequence current at the current power frequency are core indicators reflecting the steady-state operation of the distribution network. They can intuitively demonstrate key information such as node voltage stability, current load, and the presence of grounding faults, providing core data dimensions for digital twins to reconstruct the true operating state of the distribution network. Furthermore, the calculation of the rate of change of the RMS current relative to the previous power frequency cycle further uncovers the dynamic trends of electrical quantities, enabling early detection of subtle signs of the distribution network transitioning from a steady-state to a fault state, making intelligent monitoring nodes more sensitive to abnormal conditions.

[0017] Optionally, the calculation steps for the dynamic event trigger threshold and the dynamic recovery threshold include: The intelligent monitoring node calculates and maintains the average and standard deviation of steady-state characteristic quantities within a sliding time window of length N; The dynamic event trigger threshold and the dynamic recovery threshold are calculated based on the average value and the standard deviation, respectively, and the dynamic recovery threshold is less than the dynamic event trigger threshold.

[0018] By adopting the above technical solution, and calculating the threshold based on the average value and standard deviation of steady-state characteristic quantities within a sliding time window, the dynamic changes in the operating state of the distribution network can be fully considered. This avoids the problem of false triggering or missed triggering that is prone to occur when fixed thresholds are used to deal with load fluctuations and environmental changes. The threshold calculation method based on the average value and adjusted according to the standard deviation can dynamically adjust the sensitivity of triggering and recovery according to the stability of system operation. When the system is operating relatively smoothly, the standard deviation is small, the threshold range is relatively narrow, and the perception of abnormal states is more sensitive. When the system is in a condition with large fluctuations, the standard deviation increases, and the threshold range is widened accordingly, avoiding frequent false triggering caused by normal load fluctuations. Setting the dynamic recovery threshold to be less than the dynamic event triggering threshold forms a threshold hysteresis interval, which can effectively avoid the system repeatedly switching operating modes near the threshold, reducing the waste of communication resources and system instability caused by frequent state switching.

[0019] Secondly, this application provides a digital twin update system for a power distribution system, which adopts the following technical solution: A digital twin update system for a power distribution system, comprising: The data reporting module allows each smart monitoring node in the distribution network to report steady-state characteristic quantities to the regional coordinator and the digital twin master station at low frequency through the corresponding data reporting module. The broadcast listening module is used to listen for broadcasts from neighboring nodes; The judgment module is used to check whether multiple related abnormal features are met simultaneously within the same time window when a certain steady-state feature quantity of an intelligent monitoring node determines that it exceeds the dynamic event trigger threshold. If so, it is determined to be a valid fault event and the corresponding intelligent monitoring node is marked as the event source node. If not, it is determined to be interference. The data reporting module corresponding to the event source node switches the reporting mode from low frequency mode to high frequency mode and generates an event warning broadcast message, which is sent to all known neighbor nodes at maximum power. The data receiving module allows neighboring nodes to enter a collaborative state after receiving the event warning broadcast message through the corresponding data receiving module, and to increase the reporting frequency of their own data to the collaborative medium-frequency mode. The digital twin master station merges the high-frequency data from the event source node with the medium-frequency data from the collaborative node to reconstruct the entire process of the event's occurrence and development.

[0020] Thirdly, this application provides a computer device that adopts the following technical solution: A computer device includes a memory, a processor, and a computer program stored in the memory, the processor executing the computer program to implement the digital twin update method for a power distribution system as described in the first aspect.

[0021] Fourthly, this application provides a computer-readable storage medium, which adopts the following technical solution: A computer-readable storage medium storing a computer program capable of being loaded by a processor and executing the digital twin update method for a power distribution system as described in the first aspect.

[0022] In summary, this application includes at least one of the following beneficial technical effects: The system adopts a low-frequency reporting mode in daily operations to reduce the data transmission pressure on the communication link, the energy consumption of intelligent monitoring nodes, and the data processing load of regional coordinators and digital twin master stations. This makes the system operate efficiently and at low cost in a steady state, ensuring that the digital twin can stably and in real time mirror the dynamic process of the distribution network physical system under normal conditions.

[0023] When a fault occurs, the fault is accurately identified through multi-feature association verification. The event source node reports frequently and the neighboring nodes respond in a coordinated manner, enabling the digital twin master station to quickly obtain high-density and high-timeliness fault data, accurately replicate the fault scenario and update it dynamically. This achieves a high-fidelity mirror image of the physical system fault dynamic process of the digital twin, shortening the fault location and handling time.

[0024] Flexible channel scheduling avoids communication congestion and data loss, ensuring priority transmission of critical fault data; the local peer-to-peer network emergency mechanism autonomously locates the fault range when communication is interrupted, represents the node to integrate information, and quickly completes the fault scenario after communication is restored, ensuring that the digital twin can continuously, in real time and with high fidelity mirror the dynamic process of the distribution network physical system under various communication conditions. Attached Figure Description

[0025] Figure 1 This is a first flowchart of an embodiment of the method of this application; Figure 2 This is a second flowchart of an embodiment of the method of this application; Figure 3 This is the third flowchart of an embodiment of the method of this application. Detailed Implementation

[0026] To make the purpose, technical solution, and advantages of this application clearer, the following description is provided in conjunction with the appendix. Figures 1-3 The present application will be further described in detail below with reference to embodiments. It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the scope of the application.

[0027] The first embodiment of this application discloses a method for updating a digital twin of a power distribution system. (Refer to...) Figure 1 The update method includes S110-S160: S110, each smart monitoring node in the distribution network reports steady-state characteristic quantities to the regional coordinator and digital twin master station at low frequency, and listens to the broadcasts of neighboring nodes; S120: When a smart monitoring node determines that a certain steady-state characteristic quantity of itself exceeds the dynamic event triggering threshold, it checks whether multiple related characteristic anomalies are met simultaneously within the same time window. S130, if yes, it is determined to be a valid fault event, and the corresponding intelligent monitoring node is marked as the event source node; if no, it is determined to be interference, and returns to S110. S140, the event source node switches the data reporting mode from low frequency mode to high frequency mode and generates an event warning broadcast message, which is sent to all known neighbor nodes at maximum power; S150, the neighboring node that receives the event warning broadcast message enters the cooperative state and increases its own data reporting frequency to the cooperative medium frequency mode; The S160 digital twin master station integrates high-frequency data from the event source node with mid-frequency data from the collaborating node to reconstruct the entire process of the event's occurrence and development.

[0028] In S110, the calculation steps for the steady-state characteristic quantities corresponding to the intelligent monitoring nodes include S111-S113 (not shown in the figure): S111, the intelligent monitoring node samples the electrical quantities at its location at a fixed frequency. The electrical quantities include the instantaneous values ​​of the three-phase voltage and the instantaneous values ​​of the three-phase current. S112, the intelligent monitoring node calculates the current RMS value, voltage RMS value and zero-sequence current of the current power frequency cycle based on the sampled data; S113 calculates the rate of change of the effective value of the current relative to the previous power frequency cycle.

[0029] Specifically, each smart monitoring node in the distribution network (such as a smart terminal installed on a distribution transformer, switching station, or critical line segment) will collect electrical quantities at its location in real time at a fixed sampling frequency. The sampling frequency can be set to 5kHz, that is, 100 points are collected within each standard 50Hz power frequency cycle to ensure that instantaneous details of voltage and current changes can be captured. The sampling objects are the instantaneous values ​​of three-phase voltage (Ua, Ub, Uc) and the instantaneous values ​​of three-phase current (Ia, Ib, Ic). The sampled data is temporarily stored in the node's local buffer, awaiting subsequent processing.

[0030] Based on the instantaneous sampling data collected by S111, the intelligent monitoring node calculates key electrical characteristic quantities locally. First, for each power frequency cycle (20ms cycle), the node calculates the effective values ​​(Irms) of the three-phase current and the effective values ​​(Urms) of the three-phase voltage. The calculation typically uses the root mean square (RMS) algorithm, applying it to all instantaneous value samples within the current cycle. Second, the node calculates the rate of change (ΔIrms) between the current effective value and the previous power frequency cycle's effective value. This rate of change reflects sudden changes in current and is an important indicator for fault detection. The formula is ΔIrms = (current Irms - previous Irms) / ΔT, where ΔT is 20ms. Furthermore, the node also calculates the zero-sequence current (I0), obtained by adding the instantaneous values ​​of the three-phase currents (I0 = Ia + Ib + Ic) and then performing an RMS calculation, or by directly calculating the effective value of the vector sum of the three-phase currents. This is used to detect asymmetric conditions such as grounding faults. These calculations are performed in real time on the node's local microprocessor, and the results are timestamped and stored.

[0031] In non-event states, intelligent monitoring nodes report their calculated steady-state characteristics to the regional coordinator (a regional aggregation node or edge computing gateway) and the digital twin master station at a lower frequency (i.e., low-frequency mode). These steady-state characteristics primarily refer to the RMS values ​​of current and voltage, with the reporting frequency set to once every 5 seconds or configured according to network bandwidth. The reported data packets typically include fields such as node ID, timestamp, Irms, and Urms. Simultaneously, the node continuously listens for broadcast information from other neighboring nodes within its communication range via wireless mesh networks or power line carrier communication. Neighboring nodes are those physically electrically adjacent or reachable by a single communication hop. The content being listened to primarily consists of status broadcasts or event broadcasts from neighboring nodes.

[0032] Each intelligent monitoring node continuously monitors its calculated associated characteristic quantities (i.e., Irms, Urms, ΔIrms, I0, etc. in S110) locally. Each characteristic quantity is associated with a dynamic event trigger threshold (The). When a node detects that one of its steady-state characteristic quantities (e.g., Irms) exceeds its corresponding dynamic event trigger threshold, it considers it a preliminary abnormal signal. Then, a checking mechanism is initiated: it checks whether other associated characteristic quantities have also become abnormal within the same short time window (e.g., the current cycle or 1-2 cycles before and after). For example, when the current Irms exceeds the limit, does the current change rate ΔIrms also exceed its threshold? Does the zero-sequence current I0 also rise abnormally? Does the voltage Urms drop sharply? The node will determine whether multiple associated characteristic anomalies are met simultaneously based on preset logical rules (e.g., "Irms exceeds the limit AND ΔIrms exceeds the limit AND I0 exceeds the limit").

[0033] Based on the inspection results of S140, the node makes a determination: if multiple preset associated abnormal characteristic conditions are indeed met simultaneously within the same time window (e.g., a sudden increase in Irms, a sharp rise in ΔIrms, and I0 being significantly non-zero), then a valid fault event (such as a short-circuit fault) is determined to have occurred, and the node is marked as the event source node. If only a single characteristic exceeds the limit, or multiple characteristic values ​​exceed the limit but are not strictly correlated in time, then it is determined as interference (e.g., the start-up of a large motor may only cause a sudden increase in Irms, but ΔIrms and I0 may be normal), and the node will not mark itself as the event source, but will continue its normal low-frequency reporting and listening state (returning to S110).

[0034] Once a node is identified as the event source node, it immediately performs two key operations. First, it switches its data reporting mode from low-frequency to high-frequency. The high-frequency mode operates at a frequency significantly higher than steady-state reporting, reaching hundreds of times per second (e.g., 200Hz), ensuring the capture of details of the transient fault process. Second, the event source node generates an event warning broadcast message. This message includes the node's unique identifier (node ​​ID), the event level determined based on the degree of exceedance of associated characteristic quantities (e.g., current exceedance multiple, voltage drop depth) (e.g., Warning=1, Severe=2, Critical=3), a preliminary assessment of the event type (e.g., phase-to-phase short circuit, ground fault), and a timestamp of the event occurrence. After generating the event warning broadcast message, the event source node broadcasts it to all its known neighbor nodes at maximum communication power (e.g., adjusting the wireless module's transmit power to the highest level) to disseminate event information as quickly as possible and trigger a coordinated response. Broadcasting typically employs unacknowledged unicast or multicast methods to maximize speed.

[0035] Any neighboring node that receives an event warning broadcast message from the event source node will immediately enter a cooperative state, increasing its own data reporting frequency from low-frequency mode to cooperative mid-frequency mode. The mid-frequency mode has a frequency between low and high frequency (e.g., 20 times per second), providing more detailed data than the steady state to assist the master station in analysis, while avoiding the huge bandwidth pressure caused by high-frequency reporting like that of the event source node.

[0036] The digital twin master station continuously receives data from various nodes. When an event occurs, it simultaneously receives high-frequency data streams from the event source node and medium-frequency data streams from its cooperating nodes. The core task of the master station is to integrate this multi-source, multi-frequency data to reconstruct the entire process of the event's occurrence and development. In practice, the master station first aligns all received data onto a unified timeline based on timestamps. For data points of different frequencies, interpolation algorithms (such as linear interpolation or spline interpolation) can be used to generate continuous data at the same moment. Then, using this aligned data, combined with the distribution network topology model, the propagation path of fault current, voltage fluctuation range, and fault duration can be analyzed, thereby dynamically updating the state of the power grid in the digital twin and realistically recreating the spatiotemporal evolution of the fault event.

[0037] Reference Figure 2 The digital twin update method also includes S210-S240: S210, when the event source node or cooperating node sends a data packet, listen to whether the communication channel is idle; S220, if not, the node calculates a dynamic priority score for the data packet to be sent; S230, the node calculates the backoff time based on the dynamic priority score; S240: Nodes with shorter backoff times have priority to gain access to the channel, while other nodes wait for the backoff time before transmitting.

[0038] Specifically, when an event source node or a node in a cooperative state needs to send its data packets (whether high-frequency or medium-frequency), it must first listen to whether the communication channel it is using is idle before attempting to send. In practice, a carrier sense-based mechanism is typically used. The node activates its receiver to detect whether other nodes are transmitting signal energy on the channel. If the detected signal energy is below a certain silence threshold and remains below it for a fixed Distributed Inter-Frame Spacing (DIFS) time, the channel is considered idle. If the detected signal energy is above the threshold, the channel is considered busy.

[0039] If S210 detects that the channel is busy (i.e., the channel is not idle), the node will not immediately attempt to send, but will instead calculate a dynamic priority score P for the data packet to be sent. P = w1*S + w2*C + w3*(1-(Tcurrent-Tevent) / Ttimeout). Where S represents the severity level of the event associated with the data packet, usually mapped to a numerical value, such as Warning = 1, Severe = 2, Critical = 3; C represents the criticality of the node in the network topology, which is a static or semi-static attribute pre-assigned based on network topology analysis (such as adjacency matrix, power flow calculation), for example, end node = 1, backbone node = 2, power supply-side critical node (such as substation outgoing switch) = 3; (Tcurrent - Tevent) is the time difference between the current time (Tcurrent) and the time of the event (Tevent); Ttimeout is the maximum time (validity period) for the event data to be considered valid; w1, w2, w3 are weighting coefficients, satisfying w1 + w2 + w3 = 1, used to adjust the proportion of S, C, and time urgency in the priority calculation. The term (1 - (Tcurrent - Tevent) / Ttimeout) ensures that the newer the data after the event (smaller the time difference), the higher its priority; as time goes on (time difference approaches Ttimeout), this term approaches 0, and the priority decreases.

[0040] Based on the calculated dynamic priority score P, a node calculates a backoff time (Tbackoff) for its data packets. The calculation formula is: Tbackoff = Tmax / (P + 1). Here, Tmax is the preset maximum backoff time (e.g., 100ms). This ensures that data packets with higher priority scores P (more severe events, more critical nodes, more recent data) have shorter calculated backoff times (because the denominator P + 1 is larger). P + 1 is used to avoid a denominator of zero. Tbackoff is typically between 0 and Tmax.

[0041] After calculating the Tbackoff, a node starts a local timer to wait for the backoff period. During this waiting period, it continuously listens for the channel. When the Tbackoff expires, if the channel is idle, the node immediately gains access to the channel and sends its data packets. Because different nodes calculate different P values ​​and Tbackoff for different data packets, nodes with shorter Tbackoffs (i.e., higher priority nodes) will end their wait first and gain the opportunity to send. Nodes with longer Tbackoffs will have to wait longer before attempting to send.

[0042] The regional coordinator receives data packets from event source nodes (high-frequency) and coordinating nodes (medium-frequency) within the region. The coordinator's primary task is to preprocess and aggregate this data. First, it aligns data packets from different nodes at the same (or very close) time according to their timestamps. Then, the coordinator packages these aligned, same-time data packets from multiple nodes into a structured regional event snapshot package. This snapshot package provides a snapshot of the state of multiple observation points in the region at a specific point in time. The coordinator then uploads this packaged regional event snapshot package to the digital twin master station. The master station receives pre-processed spatiotemporal correlation data, which is more conducive to fusion and reconstruction in S160.

[0043] Reference Figure 3 The digital twin update method also includes S310-S330: S310: When the event source node cannot establish a communication connection with the regional coordinator or digital twin master station, it uses the local peer network to exchange the simplest state information with the directly adjacent neighbor node whose communication is interrupted. S320, each participating node determines the inference conclusion of the fault section based on the exchanged simplest state information and preset physical rules; S330 determines a representative node from the participating nodes. When communication is restored, the representative node reports the final conclusion of the inferred fault section and the evidence data to the digital twin master station.

[0044] In S330, the specific steps for determining the representative node from the participating nodes include: All participating nodes broadcast their own inferences and signal strengths; each parameter node compares the signal strengths received from all other participating nodes; and the participating node with the largest signal strength is determined as the representative node.

[0045] Specifically, when a communication failure occurs (such as a regional coordinator outage, fiber optic cable interruption, or severe wireless interference), preventing the event source node from establishing a valid communication connection with the regional coordinator or digital twin master station, the node will initiate an emergency local distributed processing procedure. At this time, the node utilizes its still-available, lowest-level local communication capabilities (usually peer-to-peer network communication, such as point-to-point wireless links, power line carrier PLCs, or dedicated lines between adjacent nodes) to communicate with its physically adjacent neighboring nodes that are also experiencing communication outages. The information exchanged between them is simplified to the most critical states: the direction of the current detected by the node itself (typically defined as positive, represented by +1, and negative, represented by -1), the voltage status (a simple binary state: voltage present = 1, voltage absent = 0), and a flag indicating whether an overcurrent has been detected (overcurrent = 1, no overcurrent = 0). This information is extremely concise, consuming minimal bandwidth, and aims to maintain the most basic fault location capability under extreme communication constraints.

[0046] Each node participating in the S310 information exchange (called a participating node), upon receiving the simplest state information from its direct neighbor nodes, attempts to apply pre-defined rules based on the physical characteristics of the distribution network to infer the section where a fault might occur. A core example of a physical rule is: if this node (denoted as node A) detects an overcurrent flag as true (overcurrent = 1) and reports a positive current direction (+1, pointing downstream), and simultaneously, a direct downstream neighbor node of node A (denoted as node B) reports a voltage loss (voltage state = 0), then according to Kirchhoff's current law and fault characteristics, it can be inferred that the fault point is likely located on the line segment between node A and node B. This is because node B downstream of the fault point will experience a voltage loss, while node A upstream of the fault point will detect an overcurrent flowing towards the fault point. Based on its own and its neighbor's states, the node will apply similar rules to arrive at a local inference, such as "the fault is located between me and node X".

[0047] After completing their local inference, each participating node broadcasts its inference conclusion (i.e., the faulty section it identifies) along with an indicator representing the strength of its inferred signal. The signal strength indicator is a quantified value reflecting the node's confidence in its conclusion or the reliability of the information upon which it bases its assessment. Signal strength can be calculated in several ways: a simple method is to directly use the overcurrent value (Irms) detected by the node; a larger current indicates a stronger signal and a more pronounced fault. Another method is to calculate a score that combines factors such as current magnitude, voltage drop depth, and node location criticality. The broadcast is made to other participating nodes accessible through its local communication capabilities.

[0048] After broadcasting its own conclusion and signal strength, each participating node also receives conclusions and signal strengths broadcast by other participating nodes. Each node then compares the signal strength values ​​received from all other participating nodes locally.

[0049] Based on the comparison results, each participating node independently determines which participating node has the highest signal strength value. This node with the highest signal strength is recognized (by all nodes) as the representative node for this distributed fault diagnosis. The election process is distributed, requiring no central coordination, and each node arrives at the same result based on the same rule (highest signal strength).

[0050] The elected representative node will temporarily store its own inferences, signal strength values, and the simplest state information exchanged in S310 as evidence. Once the communication link is restored, the representative node is responsible for reporting this final, locally distributed consensus-based inference of the fault segment, along with supporting evidence, to the digital twin master station.

[0051] In addition, when the event source node detects that all feature quantities are continuously below the corresponding dynamic recovery threshold, it broadcasts an event end message and switches the reporting mode from high-frequency mode back to low-frequency mode; the neighboring nodes that receive the event end message stop cooperating with the intermediate frequency reporting mode.

[0052] The calculation steps for the dynamic event trigger threshold and the dynamic recovery threshold include: The intelligent monitoring node calculates the average value and standard deviation of steady-state characteristic quantities within a sliding time window of length N; based on the average value and standard deviation, it calculates the dynamic event trigger threshold and dynamic recovery threshold, respectively, with the dynamic recovery threshold being less than the dynamic event trigger threshold.

[0053] Specifically, during continuous high-frequency reporting, the event source node continuously monitors all its calculated associated characteristic quantities. When all these characteristic quantities are continuously (e.g., for several consecutive power frequency cycles) below their corresponding dynamic recovery threshold (Thr), it indicates that the fault event has ended. At this time, the event source node generates and broadcasts an event end message, which includes information such as the node ID, event end flag, and timestamp. After broadcasting, the event source node itself also switches its data reporting mode from high-frequency mode back to low-frequency mode (steady-state reporting mode in S110).

[0054] Any neighboring node that receives an event end message from the event source node (i.e., a node that was previously in a cooperative state) will perform a state switch after parsing the message: it will stop its cooperative mid-frequency reporting mode and restore its data reporting frequency to the normal low-frequency mode. This marks the end of the entire cooperative response process for the event, and the system returns to normal monitoring status.

[0055] The intelligent monitoring node maintains a fixed-length (e.g., N=1000, corresponding to approximately 20 seconds of data) sliding time window locally to store historical values ​​of its calculated steady-state characteristics. The window is updated as new data arrives, removing the oldest data and adding the newest. Periodically (e.g., every 5 seconds) or after each window update, the node calculates two key statistics for all steady-state characteristics stored within the sliding window: the mean and the standard deviation. The mean reflects the recent baseline level of the characteristic, while the standard deviation reflects its volatility.

[0056] Based on the calculated mean and standard deviation, the node dynamically calculates two important thresholds locally: the dynamic event trigger threshold (The) and the dynamic recovery threshold (Thr).

[0057] The = average value + k1 * standard deviation, Thr = average value + k2 * standard deviation, where k1 and k2 are configurable coefficients, with k1 being greater than k2. This calculation method allows the threshold to adapt to the current operating state of the power grid. When the load is stable, the average value is stable, the standard deviation is small, and the threshold is low, making it sensitive to minor anomalies. When the load fluctuates greatly, the average value may change, the standard deviation increases, and the threshold automatically rises to avoid false triggering due to normal fluctuations.

[0058] Based on the above method embodiments, the second embodiment of this application discloses a power distribution system digital twin update system. The power distribution system digital twin update system of this application embodiment can implement any of the above-described power distribution system digital twin update methods, and the specific working process of each module in the power distribution system digital twin update system can refer to the corresponding process in the above method embodiments.

[0059] For ease of understanding, an example is as follows: A digital twin update system for a power distribution system includes: The data reporting module allows each smart monitoring node in the distribution network to report steady-state characteristic quantities to the regional coordinator and the digital twin master station at low frequency. The broadcast listening module is used to listen for broadcasts from neighboring nodes; The judgment module is used to check whether multiple related abnormal features are met simultaneously within the same time window when a smart monitoring node determines that a certain steady-state feature quantity exceeds the dynamic event trigger threshold. If so, it is judged as a valid fault event and the corresponding smart monitoring node is marked as the event source node. If not, it is judged as interference. The data reporting module corresponding to the event source node switches the reporting mode from low frequency mode to high frequency mode and generates an event warning broadcast message, which is sent to all known neighbor nodes at maximum power. The data receiving module allows neighboring nodes to enter a collaborative state after receiving the event warning broadcast message through their corresponding data receiving module, and to increase their own data reporting frequency to the collaborative medium-frequency mode. The digital twin master station merges the high-frequency data from the event source node with the medium-frequency data from the collaborative nodes to reconstruct the entire process of the event's occurrence and development.

[0060] The third embodiment of this application provides a computer device, which may include a memory, a processor, and a computer program stored in the memory. The processor executes the computer program to implement a method for updating a digital twin of a power distribution system.

[0061] The memory can communicate with the processor via a communication bus, which can be an address bus, a data bus, a control bus, etc.

[0062] Additionally, the memory may include random access memory (RAM) or non-volatile memory (NVM), such as at least one disk storage device.

[0063] Furthermore, the processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc.

[0064] The fourth embodiment of this application provides a computer-readable storage medium storing a computer program that can be loaded by a processor and executed as a method for updating a digital twin of a power distribution system.

[0065] The computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in connection with an instruction execution system, apparatus, or device; the program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to wireless, wire, optical fiber, RF, etc., or any suitable combination thereof.

[0066] It should be noted that the computer device and storage medium in the embodiments of this application are respectively electronic devices and storage media for applying the above-described power distribution system digital twin update method. Therefore, all embodiments of the above-described power distribution system digital twin update method are applicable to the computer device and storage medium, and can achieve the same or similar beneficial effects. For the computer device / storage medium embodiments, since they are basically similar to the method embodiments, the description is relatively simple; relevant details can be found in the descriptions of the method embodiments.

[0067] Although this application has been described herein in conjunction with various embodiments, those skilled in the art, by reviewing the accompanying drawings, disclosure, and appended claims, will understand and implement other variations of the disclosed embodiments in carrying out the claimed application. In the claims, the word "comprising" does not exclude other components or steps, and "a" or "an" does not exclude a plurality. A single processor or other unit can implement several functions listed in the claims. While different dependent claims may recite certain measures, this does not mean that these measures cannot be combined to produce a good effect.

[0068] The above are all preferred embodiments of this application and are not intended to limit the scope of protection of this application. Any feature disclosed in this specification (including the abstract and drawings) may be replaced by other equivalent or similar features unless specifically stated otherwise. That is, unless specifically stated otherwise, each feature is only one example of a series of equivalent or similar features.

Claims

1. A method for updating a digital twin of a power distribution system, characterized in that, include: Each smart monitoring node in the distribution network reports steady-state characteristic quantities to the regional coordinator and digital twin master station at low frequency, and listens to the broadcasts of neighboring nodes; When a smart monitoring node determines that a certain steady-state characteristic quantity of itself exceeds the dynamic event triggering threshold, it checks whether multiple related characteristic anomalies are met simultaneously within the same time window. If yes, it is determined to be a valid fault event, and the corresponding intelligent monitoring node is marked as the event source node; otherwise, it is determined to be interference. The event source node switches the data reporting mode from low-frequency mode to high-frequency mode and generates an event warning broadcast message, which is sent to all known neighbor nodes at maximum power. Neighboring nodes that receive the event warning broadcast message enter a collaborative state and increase their own data reporting frequency to a collaborative medium frequency mode; The digital twin master station merges high-frequency data from the event source node with mid-frequency data from the collaborating node to reconstruct the entire process of the event's occurrence and development.

2. The method for updating a digital twin of a power distribution system according to claim 1, characterized in that, The digital twin update method further includes: When the event source node or cooperating node sends a data packet, it listens to whether the communication channel is idle; If not, the node calculates a dynamic priority score for the data packet to be sent; The node calculates the backoff time based on the dynamic priority score; The node with the shorter backoff time has priority to obtain channel access, while the other nodes wait for the backoff time before sending.

3. The method for updating a digital twin of a power distribution system according to claim 2, characterized in that, The digital twin update method further includes: When the event source node is unable to establish a communication connection with the regional coordinator or digital twin master station, it uses the local peer-to-peer network to exchange the simplest state information with the directly adjacent neighbor node whose communication is interrupted. Each participating node determines the inference conclusion of the faulty section based on the exchanged simplest state information and preset physical rules; A representative node is determined from the participating nodes. When communication is restored, the representative node reports the final conclusion of the inferred fault section and the evidence data to the digital twin master station.

4. The method for updating a digital twin of a power distribution system according to claim 3, characterized in that, The specific steps for determining the representative node from the participating nodes include: All participating nodes broadcast their own inferences and signal strength; Each parameter node compares the signal strength received from all other participating nodes; The participating node with the strongest signal strength is selected as the representative node.

5. The method for updating a digital twin of a power distribution system according to claim 1, characterized in that, The digital twin update method further includes: When the event source node detects that all feature quantities are continuously lower than the corresponding dynamic recovery threshold, it broadcasts an event end message and switches the reporting mode from high-frequency mode back to low-frequency mode. The neighboring node that receives the event end message stops the cooperative intermediate frequency reporting mode.

6. The method for updating a digital twin of a power distribution system according to claim 1, characterized in that, The calculation steps for the steady-state characteristic quantities corresponding to the intelligent monitoring node include: The intelligent monitoring node samples the electrical quantities at its location at a fixed frequency. The electrical quantities include the instantaneous values ​​of the three-phase voltage and the instantaneous values ​​of the three-phase current. The intelligent monitoring node calculates the RMS current, RMS voltage, and zero-sequence current of the current power frequency cycle based on the sampled data; Calculate the rate of change of the effective value of the current relative to the previous power frequency cycle.

7. The method for updating a digital twin of a power distribution system according to claim 5, characterized in that, The calculation steps for the dynamic event trigger threshold and the dynamic recovery threshold include: The intelligent monitoring node calculates and maintains the average and standard deviation of steady-state characteristic quantities within a sliding time window of length N; The dynamic event trigger threshold and the dynamic recovery threshold are calculated based on the average value and the standard deviation, respectively, and the dynamic recovery threshold is less than the dynamic event trigger threshold.

8. A digital twin update system for a power distribution system, characterized in that, Performing the digital twin update method for a power distribution system as described in any one of claims 1 to 7, comprising: The data reporting module allows each smart monitoring node in the distribution network to report steady-state characteristic quantities to the regional coordinator and the digital twin master station at low frequency through the corresponding data reporting module. The broadcast listening module is used to listen for broadcasts from neighboring nodes; The judgment module is used to check whether multiple related abnormal features are met simultaneously within the same time window when a certain steady-state feature quantity of a smart monitoring node determines that it exceeds the dynamic event trigger threshold. If so, it is determined to be a valid fault event and the corresponding smart monitoring node is marked as the event source node. If not, it is determined to be interference. The data reporting module corresponding to the event source node switches the reporting mode from low frequency mode to high frequency mode and generates an event warning broadcast message, which is sent to all known neighbor nodes at maximum power. The data receiving module allows neighboring nodes to enter a collaborative state after receiving the event warning broadcast message through the corresponding data receiving module, and to increase the reporting frequency of their own data to the collaborative medium-frequency mode. The digital twin master station merges the high-frequency data from the event source node with the medium-frequency data from the collaborative node to reconstruct the entire process of the event's occurrence and development.

9. A computer device, characterized in that, It includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the program, implements the digital twin update method for a power distribution system as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The system stores a computer program capable of being loaded by a processor and executing the digital twin update method for a power distribution system as described in any one of claims 1 to 7.

Images