Electrical automation control system for intelligent operation and maintenance of power distribution cabinet

Abstract

The invention discloses an electrical automation control system for intelligent operation and maintenance of a power distribution cabinet, and relates to the technical field of intelligent operation and maintenance of power distribution cabinets. A single SoC heterogeneous multi-core integrated architecture is adopted, a protection core and a communication core are integrated on a single chip, and data interaction between the protection core and the communication core is achieved through an annular queue; the protection core exclusively occupies an AXI4-Lite slave port and directly mounts an annular queue, the communication core accesses the queue through an independent AXI4 master port, and the annular queue supports multi-priority data storage and lock-free synchronous read-write; according to the structural design, the data transmission efficiency and the real-time performance are greatly improved, the protection core collects data and then directly writes the data into the annular queue through an exclusive port, the communication core reads the data as needed, and serial port conversion and data copying are not needed; and secondly, the reliability and the stability of the protection action are guaranteed, the protection core exclusively occupies a port to realize that data reading and writing are not interfered by the communication core, a lock-free synchronization mechanism avoids time delay loss caused by bus competition and a software lock, and the pain point of protection action jitter in the traditional architecture is solved.

Description

Technical Field

[0001] This invention belongs to the field of intelligent operation and maintenance technology of power distribution cabinets, specifically an intelligent operation and maintenance electrical automation control system for power distribution cabinets. Background Technology

[0002] Existing intelligent operation and maintenance systems for power distribution cabinets generally adopt a dual-hardware separation architecture. This means that a separate MCU hardware module is responsible for collecting monitoring data and executing protection actions, while a separate communication hardware module handles data format conversion, encapsulation, and interaction with the backend decision-making module. Both types of hardware modules require independent power management circuits, signal conditioning circuits, and communication interface circuits, and rely on external communication methods such as UART serial ports for data exchange. Data transmission involves multiple stages: MCU acquisition, serial port transmission, communication module conversion, and backend reception.

[0003] The aforementioned existing technologies have significant technical defects: On the one hand, the dual hardware separation architecture not only results in large system hardware size and high failure rate, but also has weak anti-electromagnetic interference capability of serial port transmission, making it prone to packet loss and bit error problems in the strong electromagnetic environment of the power distribution cabinet. At the same time, the data conversion and transmission process of multiple stages will generate significant time delays, which cannot meet the rapid response requirements of fault protection. On the other hand, in the multi-core architecture of shared bus, the data transmission behavior of the communication module is prone to preempting bus resources, resulting in data read and write delays of the acquisition module, causing protection action jitter, and the lack of a priority mechanism makes it impossible for critical data such as faults to be processed first, further reducing the reliability and timeliness of protection actions.

[0004] To address the aforementioned technical problems, this invention provides an intelligent electrical automation control system for the operation and maintenance of power distribution cabinets. Summary of the Invention

[0005] The present invention aims to solve at least one of the technical problems existing in the prior art; to this end, the present invention proposes an intelligent operation and maintenance electrical automation control system for power distribution cabinets.

[0006] To achieve the above objectives, a first aspect of the present invention provides an intelligent operation and maintenance electrical automation control system for a power distribution cabinet, comprising a terminal operation and maintenance module and a decision control module, wherein the terminal operation and maintenance module is deployed in the power distribution cabinet; The terminal operation and maintenance module includes a protection core and a communication core integrated into a single SOC. The protection core and the communication core interact with each other through a circular queue. The protection core is used to collect operational monitoring data and execute protection actions according to control commands; the communication core is used to transmit data and obtain model inference results by performing fault inference on operational monitoring data. The decision control module generates control commands based on the detection data and sends the control commands to the terminal operation and maintenance module; the detection data includes operation monitoring data and model inference results.

[0007] In one possible implementation, the protection core and the communication core exchange data via a circular queue, including: The protection core reads and writes the circular queue through the exclusive AXI4-Lite port; the communication core reads and writes the circular queue through the AXI4 master port. The protection core or communication core reads and writes to the circular queue according to priority; where the circular queue refers to a random access memory that uses a circular data structure, and the circular queue supports several priorities.

[0008] In one possible implementation, the decision control module generates control commands based on the detection data, including: Receive detection data uploaded through a communication link group; wherein the communication link group includes three heterogeneous communication links; Control commands are generated based on the detection data and sent to the terminal maintenance module via the communication link group.

[0009] In one possible implementation, control instructions can also be generated by the protection kernel, including: The communication core writes the model inference results into a circular queue; The protection core generates control commands based on model inference results and operational monitoring data.

[0010] In one possible implementation, the protection core performs protection actions according to control instructions, including: Extract control commands within a preset time period; the extraction order is: first, control commands issued by the decision control module, then control commands generated by the protection core; Match the protection action with the control command and execute the protection action.

[0011] In one possible implementation, a communication link group is constructed, including: Three heterogeneous communication links are constructed based on the principle of heterogeneous complementarity; whereby the principle of heterogeneous complementarity refers to selecting communication technologies with complementary advantages according to the operation and maintenance environment of the power distribution cabinet. A communication link group is formed by using three communication links.

[0012] In one possible implementation, the communication link group operates as follows: Three communication links transmit target data synchronously; the target data refers to the data exchanged between the decision control module and the terminal operation and maintenance platform. The target data transmission is considered successful when the target data transmitted through at least two communication links is consistent and the verification is passed; the verification includes anti-tampering verification and anti-replay verification.

[0013] In one possible implementation, during the operation of the communication link group, the communication quality of the three communication links is monitored in real time, including: Extract the communication quality of the primary link and the backup link in the communication link group; wherein the primary link includes two communication links and the backup link includes one communication link; If the communication quality of any primary link fails to meet the requirements, but the communication quality of the backup link meets the requirements, the data transmitted by the primary link will be diverted to the backup link.

[0014] In one possible implementation, the communication core performs fault reasoning through a fault reasoning model, which is updated via federated learning, including: The communication core trains and deploys a fault reasoning model based on the collected operation monitoring data, extracts the gradient of the trained model, and uploads it to the decision control module; the fault reasoning model is deployed after pre-training. The decision control module uses a weighted federated average algorithm to aggregate the model gradients sent by each communication core to generate a global optimization model; the global optimization model is then distributed to each communication core.

[0015] Compared with the prior art, the beneficial effects of the present invention are: 1. This invention adopts a single-SoC heterogeneous multi-core integrated architecture, integrating a protection core and a communication core on a single chip. The two interact through a ring queue. The protection core exclusively uses the AXI4-Lite slave port and directly attaches to the ring queue, while the communication core accesses the queue through an independent AXI4 master port. The ring queue supports multi-priority data storage and lock-free synchronous read / write. This structural design completely abandons the dual hardware separation architecture and serial port transmission mode in existing technologies, bringing multiple technical benefits: First, it significantly improves data transmission efficiency and real-time performance. After the protection core collects data, it directly writes it to the ring queue through its exclusive port, and the communication core reads it on demand, without the need for serial port conversion and data copying. Second, it ensures the reliability and stability of protection actions. The protection core's exclusive port enables data read / write without interference from the communication core, and the lock-free synchronization mechanism avoids the latency loss caused by bus contention and software locks, ensuring priority transmission of fault data and rapid execution of protection commands, thus solving the pain point of protection action jitter in traditional architectures.

[0016] 2. This invention uses three heterogeneous communication links to form a communication link group based on the principle of heterogeneous complementarity. The three links transmit target data synchronously. Successful transmission is determined by a voting mechanism that verifies tamper-proof verification, replay-proof verification, and consistency of data from at least two links. At the same time, link quality is monitored in real time and dynamic traffic splitting between primary and backup links is implemented. Compared with the existing dual-link cold standby architecture, this scheme has significant technical advantages: First, it significantly improves communication reliability and remote control success rate. The three heterogeneous links operate in parallel and complement each other, avoiding communication interruption caused by a single link failure and solving the problem of packet loss caused by strong electromagnetic interference and signal obstruction in power distribution rooms. Second, it reduces link switching latency and communication interruption risk. It abandons the multi-stage switching process of fault detection-link disconnection-wake-up and reconstruction in the traditional cold standby architecture. The faulty link does not need to interrupt the current communication. Seamless redundancy is achieved through dynamic traffic splitting, ensuring the continuity of control command issuance and maintenance data upload. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the 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.

[0018] Figure 1 This is a schematic diagram of the principle of the intelligent operation and maintenance electrical automation control system for the power distribution cabinet in an embodiment of the present invention; Figure 2 This is a schematic diagram illustrating the principle of data interaction between the protection core and the communication core in an embodiment of the present invention; Figure 3 This is a schematic diagram illustrating the working principle of the communication link group in an embodiment of the present invention. Detailed Implementation

[0019] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] Please see Figure 1The first aspect of this invention provides an intelligent operation and maintenance electrical automation control system for a power distribution cabinet, including a terminal operation and maintenance module and a decision control module. The terminal operation and maintenance module is deployed in the power distribution cabinet. The terminal operation and maintenance module includes a protection core and a communication core integrated in a single system-on-a-chip (SOC). The protection core and the communication core interact with each other through a circular queue. The protection core is used to collect operation monitoring data and execute protection actions according to control commands. The communication core is used to transmit data and obtain model inference results by performing fault inference on the operation monitoring data. The decision control module generates control commands based on the detection data and sends the control commands to the terminal operation and maintenance module. The detection data includes operation monitoring data and model inference results.

[0021] The core objective of intelligent operation and maintenance electrical automation control for power distribution cabinets is to realize operation and maintenance functions such as power distribution cabinet operation status perception, fault early warning and handling, and remote control. It revolves around the workflow of data acquisition, transmission, decision-making and execution, and achieves full-process operation and maintenance management through the coordinated cooperation of various modules.

[0022] Existing intelligent operation and maintenance systems for power distribution cabinets generally have two independent hardware modules: a data acquisition module and a communication transmission module. During intelligent operation and maintenance of the power distribution cabinet, the data acquisition module collects operational monitoring data from the cabinet and transmits it to the communication transmission module via a serial port. The communication transmission module then performs format conversion and encapsulation processing on the operational monitoring data before uploading it to the decision control module.

[0023] Because two independent hardware modules are used, power management circuits, signal conditioning circuits, and communication interface circuits need to be designed for each module, which increases the deployment difficulty. Furthermore, the two independent hardware modules rely on serial port transmission for data exchange, but serial port transmission has weak electromagnetic interference resistance, and is prone to packet loss and bit errors in the strong electromagnetic environment of the power distribution cabinet, leading to communication failures. To solve this technical problem, this application adopts a single-SoC heterogeneous multi-core integrated architecture, integrating a protection core and a communication core on a single SOC. The protection core and the communication core exchange data through a circular queue.

[0024] In one example, a single-SoC heterogeneous multi-core integrated architecture is adopted, integrating an ARM Cortex-R52 protection core and a Cortex-A55 communication core. The protection core exclusively uses the AXI4-Lite slave port and directly attaches a ring queue, eliminating the need for additional serial communication circuitry. In the intelligent operation and maintenance of the power distribution cabinet, after the protection core collects the operation monitoring data, it directly writes it to the ring queue attached to the exclusive port. The communication core directly reads the data in the ring queue from the AXI4 master port and uploads it to the processing platform.

[0025] Operational monitoring data refers to data collected from distribution cabinets for intelligent operation and maintenance, including but not limited to: electrical quantity data (such as current, voltage, power, and energy), non-electrical quantity data (such as temperature and humidity), and equipment status data (such as circuit breaker status, disconnector status, protection device operation data, and communication status data). Operational monitoring data is collected through various types of sensors, which will not be listed here.

[0026] A circular queue is a random access memory that employs a circular data structure. It has a pre-defined temporary storage space, and its core characteristic is that data storage follows a first-in, first-out (FIFO) rule, with the storage spaces connected end-to-end to form a closed loop, allowing for the reuse of released storage units. The circular queue in this application is used to protect the temporary storage and exchange of data between the core and the communication core, and is an important component for improving data transmission efficiency and reliability.

[0027] Figure 2 This diagram illustrates the principle of data interaction between the protection core and the communication core. After collecting operational monitoring data from the distribution cabinet, the protection core directly writes the data to a circular queue via its dedicated AXI4-Lite port. The circular queue supports multiple priorities, allowing the protection core to write data to the corresponding priority queue based on its urgency, ensuring that high-priority operational monitoring data is processed first. Furthermore, the process of writing operational monitoring data to the circular queue does not require waiting for a response from the communication core, enabling asynchronous operations that are completed immediately after writing, minimizing the protection core's processing time.

[0028] The communication core monitors the status of the circular queue in real time from the AXI4 master port. When data is detected in the circular queue, it reads the data directly from the queue without the protection core needing to participate in data forwarding. The communication core reads data from the circular queue according to the first-in, first-out (FIFO) principle, and prioritizes reading data from higher-priority queues to ensure rapid uploading of critical data such as fault data. Because the communication core reads data directly from the circular queue, there is no need to copy data between the protection core and the communication core, avoiding the latency loss of traditional serial port transmission.

[0029] It should also be noted that the protection core executes protection actions based on control commands issued by the decision control module. The protection algorithms corresponding to these actions are directly deployed in the bare-metal program of the protection core, eliminating the need for complex system scheduling and enabling rapid response to fault triggers and execution of protection actions. The communication core can run a Linux system, possessing advantages such as strong computing power and high compatibility. It is responsible for non-real-time intelligent decision-making tasks, capable of parallel processing of data transmission and model inference, writing model inference results to a circular queue and uploading them to the decision control module.

[0030] When the occupancy rate of the circular queue exceeds a set threshold (e.g., 80%), the communication core priority promotion mechanism is automatically triggered to accelerate the data reading speed of the communication core and avoid queue overflow or data loss caused by the data writing speed exceeding the reading speed. The read / write pointer synchronization queue state can be preset between the protection core and the communication core without the need for software-level locking mechanisms. When the protection core writes data, it only updates the write pointer; when the communication core reads data, it only updates the read pointer, achieving lock-free synchronization and further reducing interaction latency.

[0031] AXI4 (Advanced eXtensible Interface 4) is a fourth-generation high-performance on-chip bus protocol designed for high-speed data transmission between modules within a System-on-a-Chip (SoC). In this application, the AXI4 bus handles data interaction between the protection core and the communication core, replacing the traditional dual-hardware UART serial port solution and reducing data transmission latency from an architectural perspective.

[0032] The AXI4 master port is a port capable of actively initiating data transmission requests. In this application, the communication core is equipped with an AXI4 master port, whose function is to actively initiate read data requests to the ring queue exclusively controlled by the protection core. The specific workflow is as follows: the communication core sends the read address and read control signal to the AXI4-Lite slave port of the ring queue through the AXI4 master port. After successful verification, it directly reads the operation monitoring data collected by the protection core from the ring queue, without the protection core needing to actively push the data, thus achieving efficient interaction through on-demand reading.

[0033] AXI4-Lite is a lightweight version of the AXI4 protocol, designed for low-speed, simple control interactions. The AXI4-Lite slave port is integrated into the passive response module, and its core function is to receive transmission requests from the master port and complete address decoding, data storage, or data output. In this application, the protection core exclusively uses the AXI4-Lite slave port, and this port is directly connected to a ring queue. This means that only the protection core can write data to the ring queue through this port; other modules cannot modify data through this port. The specific workflow is as follows: after the protection core collects operational monitoring data, it writes it directly to the ring queue through the AXI4-Lite slave port, without needing bus arbitration or other intermediate modules, ensuring the real-time performance and exclusivity of data writing. At the same time, this AXI4-Lite slave port can respond to read requests from the communication core's master port, cooperating to complete data interaction.

[0034] In the technical solution of this application, data transmission between the protection core and the communication core no longer relies on serial port transmission, thus increasing data transmission reliability. Simultaneously, the elimination of serial port conversion time reduces the latency from data acquisition to the decision control module, improving the operation and maintenance efficiency of the power distribution cabinet.

[0035] In traditional single-SoC multi-core architectures, multiple cores share the same bus resources, and each core's access to the bus requires resource allocation through an arbitration mechanism. If the protection core and communication core share the bus, the communication core may preempt bus resources when performing large-scale data transmission, causing data read / write delays for the protection core and consequently leading to protection action jitter. To address this issue, this application configures a dedicated AXI4-Lite slave port for the protection core, with a ring queue directly mounted on this dedicated port. The communication core accesses the queue through an independent AXI4 master port, eliminating bus resource contention between the two. Furthermore, the protection core's data read / write operations are completed directly through the dedicated port, unaffected by the communication core's data transmission behavior, ensuring the real-time performance and accuracy of protection actions.

[0036] The decision control module generates control commands based on the operational monitoring data and model inference results uploaded by the communication core, and then sends these commands to the protection core via the communication core. The protection core can then execute protection actions based on these commands. This solution has been disclosed in existing technologies, and the details of how to generate control commands based on operational monitoring data and model inference results will not be elaborated here.

[0037] In addition to being generated by the decision control module, control commands can also be generated by the protection core of the terminal operation and maintenance module. Specifically, the communication core writes the model inference results into a circular queue for the protection core to read. The protection core matches control commands based on the model inference results and the corresponding operational monitoring data. If no control command is received from the decision control module within a preset time period, the protection core matches and executes protection actions according to its generated control command. This eliminates the need to wait for the decision control module to issue control commands in emergency situations, thus addressing urgent faults. The preset time period can be set according to actual conditions, such as 1 second, 10 seconds, etc.

[0038] In one example, assuming a preset time period of 1 second, the protection core checks whether it has received a control command. If no control command is received, it generates a control command immediately after receiving the model inference result. A timer starts after the control command is generated. If no control command is received from the decision control platform after 1 second, the protection core executes the generated control command.

[0039] In some other preferred embodiments, it can also be determined whether to execute the control command generated by the protection core based on the urgency of the situation. For example, if the arc intensity is ≥ the threshold and the inference failure probability is ≥ 95%, it is determined to be an emergency.

[0040] It should be noted that when the protection core generates control commands based on operational monitoring data and model inference results, it primarily matches these data with a pre-defined control command library before executing the corresponding protection action. The control commands generated by the protection core are directly matched, and its pre-defined control command library has limitations, only suitable for emergency situations. The decision control module can utilize a globally optimized fault inference model to obtain model inference results. Control commands generated based on these model inference results and operational monitoring data are more reliable.

[0041] The existing solution employs a dual-link cold standby architecture for data upload and command issuance, configuring two independent links: a primary link and a backup link. It also includes a link switching control circuit and a status detection module. The two links do not operate simultaneously. Under normal conditions, only the primary link operates, handling data upload and command reception, while the backup link remains in a dormant standby state. When the primary link fails, the status detection module triggers the switching control circuit, first disconnecting the primary link session, then waking up the backup link and re-establishing the communication session. After the switchover, the backup link takes over the communication tasks. This entire process involves multiple stages: fault detection, primary link disconnection, backup link wake-up, and session reconstruction.

[0042] Figure 3 This is a schematic diagram illustrating the working principle of a communication link group.

[0043] This application improves communication efficiency and remote control success rate through a three-to-two link voting method. It employs a three-link parallel monitoring structure, constructing three heterogeneous communication links using three different communication schemes, and includes a voting control module and a token verification module. All three communication links are simultaneously in a monitoring state. When the decision control module issues control commands, it includes command information, including a digest, timestamp, and signature token. After the three communication links synchronously receive the control commands, they are first verified by the token verification module to prevent tampering and replay attacks. Then, the voting control module compares the control commands received by the three communication links. If at least two heterogeneous links receive the same control commands and the verification passes, the operation is executed based on that control command without link switching. Even if one communication link fails, the normal communication process is not affected.

[0044] In one example, when constructing three heterogeneous communication links, considering the complex operation and maintenance environment of the power distribution cabinet, the three communication links are constructed based on the principle of heterogeneous complementarity. These three heterogeneous communication links are built based on 4G technology, industrial Ethernet, and LoRa technology, respectively, with the 4G link and industrial Ethernet link serving as the primary links and the LoRa link serving as the backup link. The three communication links simultaneously monitor and collect link status.

[0045] Because the intelligent operation and maintenance of power distribution cabinets is susceptible to strong electromagnetic interference and signal obstruction in some areas, three communication links with significantly different characteristics were selected and constructed based on the principle of heterogeneous complementarity. The 4G link can ensure wide-area coverage and high-speed data transmission, and can be adapted to remote control command issuance and large-volume data upload; the LoRa link has the characteristics of low power consumption and strong penetration, and can be adapted to communication gap filling in obstructed areas; the industrial Ethernet link has strong anti-electromagnetic interference capability and low transmission latency, and can be adapted to reliable short-range command transmission.

[0046] During data interaction, the communication quality of the three communication links is analyzed in real time by collecting link status data. When the communication quality of the primary link (4G link or industrial Ethernet link) does not meet the requirements (such as packet loss rate exceeding a preset threshold, or latency exceeding a preset threshold), the current operation is not interrupted, and the data is automatically diverted to the backup link. Moreover, the faulty link can initiate self-healing detection, and if it is detected that it has recovered to normal, it will automatically return to the link pool, achieving uninterrupted switching.

[0047] It should be noted that the three-to-two link voting method is mainly used for data transmission between the data processing platform and the communication core. Of course, its core application is to transmit control commands issued by the data processing platform to the communication core.

[0048] This application completely avoids the session reconstruction time during cold standby switching, significantly reducing link switching latency compared to existing solutions. Furthermore, token verification enhances command security, fundamentally addressing the core pain point of "link failure leading to operation failure and delayed fault handling" in power distribution remote control, thus ensuring the reliability of power grid operation and maintenance. Moreover, three heterogeneous communication links monitor in parallel, with each link being independently affected by environmental interference. Even if one link experiences performance degradation due to certain types of interference, the other two links can maintain stable communication. Consistent voting ensures that command and data transmission are unaffected, allowing for precise adaptation to complex environments such as electromagnetic interference and signal obstruction in distribution cabinets.

[0049] Existing solutions for fault reasoning based on operational monitoring data collected from distribution cabinets often employ pre-trained artificial intelligence (AI) models. These models are trained using a large number of pre-collected training samples and, once they meet the requirements, are deployed in the terminal maintenance module. Fault reasoning is then performed using this AI model. However, if more suitable sample data becomes available in the terminal maintenance module, only the AI ​​model deployed in that module undergoes incremental training, without external data supplementation or model iteration. This results in data silos between different terminal maintenance modules.

[0050] This application employs a terminal-server federated learning architecture. The terminal operation and maintenance module trains and deploys an artificial intelligence model based on locally collected operation monitoring data. Only the model gradient of the trained artificial intelligence model is extracted and uploaded to the server. The server uses a weighted federated averaging algorithm (e.g., dynamically allocating weights based on the amount and quality of client data) to aggregate the model gradients of multiple substation terminals, generate a globally optimized model, and then distribute it to each terminal operation and maintenance module. The terminal operation and maintenance module updates the model parameters of its local artificial intelligence model based on the global model.

[0051] This application addresses the problem of poor model generalization ability caused by insufficient data volume and data silos in single-terminal operation and maintenance modules, improves the fault reasoning accuracy of artificial intelligence models deployed in each terminal operation and maintenance module, significantly reduces invalid operation and maintenance work orders, and significantly improves the operation and maintenance efficiency of distribution networks.

[0052] Furthermore, existing solutions do not separate model updates from core operation and maintenance functions. Model deployment and intelligent operation and maintenance share hardware resources. When a model update is required, the core operation and maintenance functions must be interrupted first, and then restarted after the model update is completed. This application, based on a single-SoC heterogeneous multi-core integrated architecture, physically isolates the protection core and communication core, ensuring independent resources. The communication core receives update packets via a 3-to-2 link, completing model verification and deployment in parallel. During the update process, the protection core continues to run without interrupting data acquisition and protection action execution. After the model update is completed, the optimized fault judgment logic is synchronized through a circular queue. This application completely solves the problem of protection function interruption during model updates in existing technologies, ensuring that the model maintains optimal performance throughout the long-term operation of the terminal operation and maintenance module. It can dynamically adapt to changes in the power distribution cabinet's operating environment, providing a core guarantee for the long-term reliable operation of the terminal operation and maintenance module.

[0053] It is worth noting that the aforementioned terminal operation and maintenance modules, communication link groups, and protection actions do not function in isolation, but rather work synergistically. Specifically, the rapid data transmission of the single SoC integrated multi-core architecture provides real-time data support for federated learning model inference, while the three-to-two link voting ensures the reliable issuance of model update instructions and remote control instructions, significantly improving the intelligence level and operation and maintenance economy of the distribution terminal, and helping the power grid achieve its intelligent operation and maintenance goals.

[0054] It should be understood that the descriptions of technical features, technical solutions, beneficial effects, or similar language in this application do not imply that all features and advantages can be achieved in any single embodiment. Rather, it is understood that the description of a feature or beneficial effect means that a specific technical feature, technical solution, or beneficial effect is included in at least one embodiment. Therefore, the descriptions of technical features, technical solutions, or beneficial effects in this specification do not necessarily refer to the same embodiment. Furthermore, the technical features, technical solutions, and beneficial effects described in this embodiment can be combined in any suitable manner. Those skilled in the art will understand that embodiments can be implemented without one or more specific technical features, technical solutions, or beneficial effects of a particular embodiment. In other embodiments, additional technical features and beneficial effects may be identified in specific embodiments that do not embody all embodiments.

[0055] In the above embodiments, implementation can be achieved entirely or partially through software, hardware, firmware, or any other combination thereof. When implemented using a software program, it can be implemented entirely or partially in the form of a computer program product. This computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium accessible to a computer or a data storage device containing one or more servers, data centers, etc., that can be integrated with the medium. The available media can be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., DVDs), or semiconductor media (e.g., solid-state disks (SSDs)).

[0056] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.

Claims

1. A smart electrical automation control system for the operation and maintenance of a power distribution cabinet, characterized in that, It includes a terminal maintenance module and a decision control module, wherein the terminal maintenance module is deployed in the power distribution cabinet; the terminal maintenance module and the decision control module interact with each other through a communication link group. The terminal operation and maintenance module includes a protection core and a communication core integrated into a single SOC. The protection core and the communication core interact with each other through a circular queue. The protection core is used to collect operational monitoring data and execute protection actions according to control commands; the communication core is used to transmit data and obtain model inference results by performing fault inference on the operational monitoring data. The decision control module generates control commands based on the detection data and sends the control commands to the terminal operation and maintenance module; wherein, the detection data includes operation monitoring data and model inference results.

2. The intelligent operation and maintenance electrical automation control system for a power distribution cabinet according to claim 1, characterized in that, The protection core and the communication core interact with each other via a circular queue, including: The protection core reads and writes the ring queue through the exclusive AXI4-Lite port; the communication core reads and writes the ring queue through the AXI4 master port. The protection core or the communication core reads and writes the circular queue according to priority; wherein, the circular queue refers to a random access memory using a circular data structure, and the circular queue supports several priorities.

3. The intelligent operation and maintenance electrical automation control system for a power distribution cabinet according to claim 1, characterized in that, The decision control module generates control commands based on the detection data, including: Receive detection data uploaded through the communication link group; wherein the communication link group includes three heterogeneous communication links; Control commands are generated based on the detection data, and the control commands are sent to the terminal maintenance module through the communication link group.

4. The intelligent operation and maintenance electrical automation control system for a power distribution cabinet according to claim 1, characterized in that, The control commands can also be generated by the protection core, including: The communication core writes the model inference results into the circular queue; The protection core generates control commands based on the model inference results and the operation monitoring data.

5. The intelligent operation and maintenance electrical automation control system for a power distribution cabinet according to claim 4, characterized in that, The protection core executes protection actions according to control commands, including: Extract control commands within a preset time period; the extraction order is: first, control commands issued by the decision control module, then control commands generated by the protection core; The protection action is matched according to the control command and executed.

6. The intelligent operation and maintenance electrical automation control system for a power distribution cabinet according to claim 1, characterized in that, Constructing the communication link group includes: Three heterogeneous communication links are constructed based on the principle of heterogeneous complementarity; whereby the principle of heterogeneous complementarity refers to selecting communication technologies with complementary advantages according to the operation and maintenance environment of the power distribution cabinet. The three communication links are used to form the communication link group.

7. The intelligent operation and maintenance electrical automation control system for a power distribution cabinet according to claim 6, characterized in that, The communication link group operates in the following modes: The three communication links transmit the target data synchronously; wherein, the target data refers to the data exchanged between the decision control module and the terminal operation and maintenance platform; When the target data transmitted through at least two of the communication links is consistent and the verification is successful, the target data transmission is determined to be successful; wherein, the verification content includes anti-tampering verification and anti-replay verification.

8. The intelligent operation and maintenance electrical automation control system for a power distribution cabinet according to claim 7, characterized in that, During the operation of the communication link group, the communication quality of the three communication links is monitored in real time, including: Extract the communication quality of the primary link and the backup link in the communication link group; wherein, the primary link includes two communication links and the backup link includes one communication link; If the communication quality of any one of the primary links fails to meet the requirements, while the communication quality of the backup link meets the requirements, then the data transmitted by the primary link will be diverted to the backup link.

9. The intelligent operation and maintenance electrical automation control system for a power distribution cabinet according to claim 1, characterized in that, The communication core performs fault reasoning through a fault reasoning model, which is updated via federated learning, including: The communication core trains and deploys a fault reasoning model based on the collected operation monitoring data, and extracts the gradient of the trained model and uploads it to the decision control module; wherein, the fault reasoning model is deployed after pre-training; The decision control module uses a weighted federated average algorithm to aggregate the model gradients sent by each of the communication cores to generate a global optimization model; the global optimization model is then distributed to each of the communication cores.

Images