Electrical cabinet remote monitoring and communication system based on Internet of Things

Abstract

The invention relates to the technical field of the Internet of Things, in particular to an electrical cabinet remote monitoring and communication system based on the Internet of Things. The system comprises a bounded rollback unit which is used for completing a collaborative protection decision with an adjacent electrical cabinet node in a preset longest decision time window by adopting a D-CPB mechanism, judging whether a collaborative protection decision instruction which is triggered oppositely is generated or not by adopting a multi-round negotiation protocol, and if the longest decision time window is ended or the maximum negotiation round number is reached, sending a collaborative protection decision instruction which is triggered oppositely to the adjacent electrical cabinet node. Generating a trigger signal if the cooperative protection decision instruction which is reversely triggered is not received and the local fault confidence coefficient is not lower than a preset fault confidence coefficient threshold value; the system also provides deterministic response time delay while ensuring the collaboration, and avoids infinite waiting or refusal operation caused by communication uncertainty.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of Internet of Things, in particular, to an electrical cabinet remote monitoring and communication system based on Internet of Things. BACKGROUND

[0002] With the promotion of power distribution automation and smart grid, electrical cabinets are gradually equipped with various sensors and communication modules and connected to the Internet of Things to realize real-time state collection, remote operation and maintenance and data-driven fault diagnosis. To meet the manageability and operation efficiency of large-scale deployment, the existing system usually adopts a data aggregation and decision architecture with a cloud platform as the core: Each electrical cabinet periodically uploads current, voltage, temperature and other operating parameters to the edge gateway or cloud through on-site sampling or low-frequency telemetry, and the cloud platform is responsible for historical storage, alarm discrimination and operation decision. When fast protection action is needed, it is usually dependent on the preset threshold of the local relay or the control instruction issued by the cloud, and the communication is transmitted on the conventional channel using a general protocol.

[0003] The main problems existing in the prior art are: 1. Insufficient local protection of single cabinet and multi-cabinet cooperation: Most systems only rely on independent local threshold or completely rely on cloud coordination, lacking a mechanism that can still achieve deterministic multi-cabinet cooperation segmentation under limited network, resulting in increased risk of misoperation or refusal to operate; 2. Lack of a unified time-limited-round-backout mechanism: There is no decision-making process that combines the longest waiting time, maximum negotiation rounds and bounded backout, so it is not possible to determine when the local safety will take over the protection action in the event of network anomalies, which cannot guarantee the time determinacy of fault handling, and is not conducive to post-audit and responsibility identification.

[0004] Therefore, an electrical cabinet remote monitoring and communication system based on Internet of Things is designed. SUMMARY

[0005] The present application aims to provide an electrical cabinet remote monitoring and communication system based on Internet of Things to solve the problems raised in the background.

[0006] To achieve the above-mentioned purpose, the present application aims to provide an electrical cabinet remote monitoring and communication system based on Internet of Things, comprising: a data sampling unit, the data sampling unit is used for high-speed sampling of the local electrical cabinet node to generate sampling data with a hardware timestamp; The bounded backoff unit is used to complete the collaborative protection decision with the adjacent electrical cabinet node based on the sampled data and using the D-CPB mechanism within the preset longest decision time window. It uses a multi-round negotiation protocol to determine whether to generate a collaborative protection decision instruction against triggering. If no collaborative protection decision instruction against triggering is received before the end of the longest decision time window or after the maximum number of negotiation rounds, and the local fault confidence is not lower than the preset fault confidence threshold, then a trigger signal is generated. The protection execution unit is used to respond to the trigger signal, thereby driving the local circuit breaker to trip and achieve fault isolation.

[0007] As a further improvement to this technical solution, the bounded rollback unit includes a fault fingerprint extraction module, a deterministic communication module, and a timeout monitoring module; The fault fingerprint extraction module is used to extract compressed fingerprints representing fault characteristics from sampled data when a preset trigger condition is detected, and to generate an emergency summary and local fault confidence. The deterministic communication module is used to multicast an emergency summary containing fault characteristics to adjacent electrical cabinet nodes in the form of an emergency message. At the same time, it receives and verifies the summary information of adjacent electrical cabinet nodes. It adopts a multi-round negotiation protocol and executes a weighted fusion algorithm to generate a global confidence score within the longest decision time window. Based on the global confidence score, it determines whether to generate a cooperative protection decision instruction against triggering. The timeout monitoring module is used to monitor the timeout status and negotiation rounds of the longest decision time window. If no cooperative protection decision instruction opposing triggering is received from the deterministic communication module before the longest decision time window expires or after the maximum number of negotiation rounds is reached, and the local fault confidence level is not lower than the preset fault confidence level threshold, then a trigger signal is generated.

[0008] As a further improvement to this technical solution, the preset triggering conditions are amplitude threshold triggering, slope threshold triggering, and model scoring triggering; Among them, when there is a sampling point, the absolute value of the current amplitude corresponding to the sampling point is greater than or equal to the preset amplitude threshold, it is determined to be an amplitude threshold trigger. When the maximum value of the rate of change between adjacent sampling points is greater than or equal to the preset slope threshold, it is determined that the slope threshold is triggered. A set of features extracted from the sampling window is input into a pre-trained fault identification model, which can output a fault score. When the fault score is greater than or equal to the preset fault identification model score threshold, it is determined that the model score is triggered.

[0009] As a further improvement to this technical solution, the deterministic communication module adopts a priority scheduling strategy, using a high-priority channel to transmit emergency summaries and a low-priority channel to transmit routine monitoring data.

[0010] As a further improvement to this technical solution, the emergency summary carries a sequence number and a hardware timestamp during transmission.

[0011] As a further improvement to this technical solution, the deterministic communication module retransmits the emergency digest according to a predetermined number of repetitions and a retry interval after the initial transmission of the emergency digest.

[0012] As a further improvement to this technical solution, in the deterministic communication module, the longest decision time window is the maximum allowable time from the time the electrical cabinet first sends the emergency summary to the time it receives the summary of the adjacent electrical cabinet node and completes the preliminary fusion judgment.

[0013] As a further improvement to this technical solution, the deterministic communication module employs a multi-round negotiation protocol and executes a weighted fusion algorithm to generate a global confidence score within the longest decision time window. The specific steps for determining whether to generate a cooperative protection decision instruction against triggering based on the global confidence score are as follows: S21. In each round of negotiation window, receive and verify the digital signature of the emergency summary information sent from the adjacent electrical cabinet node, and only collect the verified information; The emergency summary information includes at least the fault confidence level of the sending electrical cabinet node and the identification identifier of the electrical cabinet node; S22. Based on the electrical cabinet node identification, assign dynamic weights to each electrical cabinet node; wherein, the dynamic weight assignment strategy is determined jointly by the node's real-time response status and historical reliability records; S23. Calculate the weighted global confidence score for the current round based on the confidence score and the assigned weights. ; S24. Define the collaborative abandonment threshold. and cooperative action threshold global confidence With the threshold of collaborative abandonment and cooperative action threshold The comparison is performed to determine whether to generate a collaborative protection decision instruction against triggering.

[0014] As a further improvement to this technical solution, the specific steps in S24 for determining whether to generate a cooperative protection decision instruction against triggering are as follows: If global confidence Less than or equal to the collaborative abandonment threshold If so, a collaborative decision instruction opposing the triggering is generated, and after being signed by this node, it is broadcast to neighboring nodes; If global confidence Higher than or equal to the cooperative action threshold If so, no opposition command will be generated in this round; If global confidence Above the collaborative abandonment threshold And global confidence Below the threshold for coordinated action If no clear consensus is reached in this round, then the next round of negotiations will begin; If the current round has reached the preset maximum negotiation round, the negotiation process will be terminated.

[0015] As a further improvement to this technical solution, the deterministic communication module is based on a Byzantine fault-tolerant decision-making mechanism, and only when it collects no less than [number missing] within the longest decision-making time window... Only when there are multiple valid and opposing collaborative decision instructions issued by different electrical cabinet nodes is the current collaborative decision deemed to be an opposing trigger. This is a pre-defined maximum number of malicious nodes that are allowed to be fault-tolerant.

[0016] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. In this IoT-based remote monitoring and communication system for electrical cabinets, a D-CPB mechanism is introduced. Within a predetermined time window, priority is given to negotiation. After timeout or reaching the maximum number of rounds, hardware priority rollback is immediately performed and a complete waveform is recorded when the local confidence threshold is met. This ensures coordination while providing deterministic response latency, avoiding infinite waiting or refusal to operate due to communication uncertainty, and comprehensively improving the real-time performance, reliability, security, and bandwidth efficiency of the protection system.

[0017] 2. In this IoT-based remote monitoring and communication system for electrical cabinets, the introduction of multi-round negotiation and the fusion of confidence levels weighted by node historical reliability and sampling quality makes the accuracy of collaborative judgment among multiple electrical cabinets significantly better than that of a single threshold or simple majority voting. At the same time, the adoption of digital signature verification and Byzantine fault tolerance strategies resists malicious nodes or forged information, enhancing the traceability and anti-attack capability of the judgment. Attached Figure Description

[0018] Figure 1 This is an overall flowchart of the present invention; The meanings of the labels in the diagram are as follows: 1. Data sampling unit; 2. Bounded backoff unit; 21. Fault fingerprint extraction module; 22. Deterministic communication module; 23. Timeout monitoring module; 3. Protection execution unit. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. 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] Example: Please refer to Figure 1 As shown, an IoT-based remote monitoring and communication system for electrical cabinets is provided, including a data sampling unit 1, a bounded backoff unit 2, and a protection execution unit 3. The specific process is as follows: when an instantaneous abnormal event is detected at an electrical cabinet node, sampling data for a predetermined time window before and after the triggering time is collected and sampling data with hardware timestamps is generated; compressed fault fingerprints are extracted from the sampling data and confidence scores are calculated; the compressed fault fingerprint summary is broadcast to adjacent electrical cabinet nodes through the emergency summary channel; within a predetermined maximum period, the received summary is weighted and judged by rule to form a collaborative conclusion; if no collaborative conclusion opposing the local protection action is formed within the maximum period and the local confidence score is greater than the preset threshold, the local circuit breaker is triggered to trip; otherwise, the corresponding control is executed according to the collaborative conclusion.

[0021] Among them, the data sampling unit 1 is used to perform high-speed sampling of local electrical cabinet nodes and generate sampling data with hardware timestamps for multi-cabinet event alignment. The sampling frequency of the data sampling unit 1 is not less than 10kHz. Bounded backoff unit 2 is used to complete the collaborative protection decision with adjacent electrical cabinet nodes based on sampled data and using the D-CPB mechanism within the preset longest decision time window. It uses a multi-round negotiation protocol to determine whether to generate a collaborative protection decision instruction against triggering. If no collaborative protection decision instruction against triggering is received before the end of the longest decision time window or after the maximum number of negotiation rounds, and the local fault confidence is not lower than the preset fault confidence threshold, then a trigger signal is generated. Each electrical cabinet node has a bounded backoff unit 2. The specific name of the D-CPB mechanism is deterministic distributed cooperative protection and bounded backoff mechanism. This mechanism integrates fault fingerprint extraction, emergency summary broadcasting, deterministic short-time negotiation and weighted confidence fusion, and backoff triggering functions. Bounded rollback unit 2 includes a fault fingerprint extraction module 21, a deterministic communication module 22, and a timeout monitoring module 23; The fault fingerprint extraction module 21 is used to extract a compressed fingerprint representing fault characteristics from the sampled data when a preset trigger condition is detected, and to generate an emergency summary and a local fault confidence score. Specifically, it performs wavelet packet transform on the sampled data with hardware timestamps and extracts the energy of a specific frequency band as fault feature data. The length of the compressed fingerprint does not exceed 256 bytes, and the emergency summary is a subset of the compressed fingerprint. In this unit, the length of the summary message sent in the emergency summary channel is limited to no more than 64 bytes, and it includes an event identifier, hardware timestamp, confidence score, local suggestion, and message signature. The preset trigger conditions are amplitude threshold trigger, slope threshold trigger, and model scoring trigger; Among them, when there is a sampling point, the absolute value of the current amplitude corresponding to the sampling point is greater than or equal to the preset amplitude threshold, it is determined to be an amplitude threshold trigger; this condition is used to capture obvious transient amplitude anomalies such as overcurrent or short circuit, and is suitable for fault scenarios with obvious amplitude changes.

[0022] When the maximum value of the rate of change between adjacent sampling points is greater than or equal to the preset slope threshold, it is determined that the slope threshold is triggered. This condition focuses on detecting transient events that rise or fall rapidly (such as rapid current spikes or sudden changes in a very short time), and can be identified even if the amplitude does not exceed the amplitude threshold. A set of features extracted from the sampling window is input into a pre-trained fault identification model, which can output a fault score. When the fault score is greater than or equal to the preset fault identification model score threshold, it is determined that the model score is triggered. This condition is used to identify faults that may not have significant amplitude or slope but have characteristic waveform patterns (such as resonance, ground leakage, complex load change, etc.) to improve the coverage and accuracy of detection.

[0023] The features extracted within the sampling window include root mean square, peak value, harmonic amplitude ratio, and short-time Fourier transform energy distribution; the fault identification model is a small-scale convolutional neural network suitable for edge devices; the scoring uses similarity based on distance metrics; and the threshold is set based on statistical results from experiments or validation sets. The deterministic communication module 22 is used to multicast an emergency summary containing fault characteristics to adjacent electrical cabinet nodes in the form of an emergency message. At the same time, it receives and verifies the summary information of adjacent electrical cabinet nodes. It adopts a multi-round negotiation protocol, executes a weighted fusion algorithm to generate a global confidence score within the longest decision time window, and determines whether to generate a cooperative protection decision instruction against triggering based on the global confidence score. The deterministic communication module 22 adopts a priority scheduling strategy, using a high-priority channel to transmit emergency summaries and a low-priority channel to transmit regular monitoring data; this improves the real-time performance and reliability of emergency summaries and avoids delays caused by regular data occupying bandwidth. The emergency summary carries a sequence number and a hardware timestamp during transmission. The sequence number is used to identify the data order and avoid duplication; the hardware timestamp is used to ensure time accuracy and consistency. The receiving end can accurately determine whether the message is up-to-date, lost, or duplicated, thereby ensuring the correctness and timeliness of emergency data. After the initial transmission of the emergency digest, the deterministic communication module 22 retransmits the emergency digest according to a predetermined number of repetitions and a retry interval; preferably, the number of repetitions is 3, the retry interval is 5 milliseconds, and the maximum period does not exceed 50 milliseconds. This redundant transmission mechanism reduces the probability of transmission failure due to channel interference or packet loss, significantly improving the success rate of emergency digest delivery.

[0024] In the deterministic communication module 22, the longest decision-making time window is the maximum allowable time from the first time the electrical cabinet sends an emergency summary to the time when it receives the summary of the adjacent electrical cabinet node and completes the preliminary fusion judgment; the time range of communication and fusion judgment is limited to avoid delays in protection decision-making due to excessive waiting time, and to ensure that the entire judgment process is completed within the defined time limit. Furthermore, the value of the longest decision time window is less than the difference between the maximum fault clearing time required by the power system relay protection and the inherent hardware action delay of the protection execution unit 3; by linking the decision time window with the relay protection action time constraint of the power system, it is ensured that there is still enough margin for the execution unit to act after the fusion judgment and communication are completed, so as to avoid the decision window being too long and occupying the protection action time, and to ensure that the decision and action are completed within the fault clearing time requirement.

[0025] In the deterministic communication module 22, a multi-round negotiation protocol is adopted. Within the longest decision time window, a weighted fusion algorithm is executed to generate a global confidence score. The specific steps for determining whether to generate a cooperative protection decision instruction against triggering based on the global confidence score are as follows: S21. In each round of negotiation window, receive and verify the digital signature of the emergency summary information sent from the adjacent electrical cabinet node, and only collect the verified information; verifying the digital signature and only collecting the verified information ensures that the basic data for subsequent fusion computing is tamper-proof and has a reliable source, directly providing the system with anti-attack capability; The emergency summary information should include at least the fault confidence level of the sending electrical cabinet node and the identification of the electrical cabinet node; S22. Based on the identification of electrical cabinet nodes, assign dynamic weights to each electrical cabinet node; wherein, the dynamic weight assignment strategy is determined by the node's real-time response status and historical reliability records. Real-time response status indicates whether a node has successfully responded to collaboration requests in the current and previous rounds, while historical reliability records indicate the accuracy of diagnostic conclusions for a node in past collaboration events. This means that if a node fails to respond multiple times in the current failure event (which may be due to a crash or communication interruption), its weight will be dynamically reduced, and nodes that are unreliable for a long time will be penalized, enabling the system to adapt to temporary failures.

[0026] S23. Calculate the weighted global confidence score for the current round based on the confidence score and the assigned weights. ;

[0027] In the formula, For the first Global confidence level of the wheel; For the current round; For the first The node at the th The weight of the wheel; For the i-th node in the... Wheel failure confidence level; This represents the total number of adjacent electrical cabinet nodes. ; S24. Define the collaborative abandonment threshold. and cooperative action threshold global confidence With the threshold of collaborative abandonment and cooperative action threshold The comparison is performed to determine whether to generate a collaborative protection decision instruction against triggering.

[0028] Among them, the collaborative action threshold It is a preset high-confidence threshold, the core consideration being to avoid false triggering, that is, to prevent the power from being mistakenly cut off when there is no real fault, causing unnecessary power outages. Therefore, It must be set high enough; in this embodiment, it is set to 0.9. Collaborative abandonment threshold This is a preset low confidence threshold, the core consideration being to avoid refusal to act, that is, to prevent the system from mistakenly classifying a real fault as normal and refusing to act, thus preventing the fault from escalating. Therefore, It must be set low enough; in this embodiment, it is set to 0.2.

[0029] The above thresholds are example preferred values, which can be determined through simulation, field tests or statistical verification data.

[0030] In S24, the specific steps for determining whether to generate a cooperative protection decision instruction against triggering are as follows: If global confidence Less than or equal to the collaborative abandonment threshold If so, a collaborative decision instruction opposing the triggering is generated, and after being signed by this node, it is broadcast to neighboring nodes; If global confidence Higher than or equal to the cooperative action threshold If so, no opposition command will be generated in this round; If global confidence Above the collaborative abandonment threshold And global confidence Below the threshold for coordinated action If no clear consensus is reached in this round, then the next round of negotiations will begin; If the current round has reached the preset maximum number of negotiation rounds, the negotiation process will be terminated. Traditional fault handling may be a one-off process: node detection, data transmission, waiting for a response, and then action. The multi-round negotiation protocol proposed in this embodiment introduces an iterative and rollback mechanism, allowing nodes to exchange information and self-correct multiple times before making a final decision. Each round of negotiation is equivalent to a rollback point. If the information in the current round is insufficient to make a high-confidence decision, the system does not freeze or act blindly, but rolls back to a state of waiting for more information and starts the next round of attempts.

[0031] Boundedness is reflected in the fact that the protocol pre-determines a maximum number of negotiation rounds. Regardless of whether an agreement is reached, the maximum number of negotiation rounds will be completed. After or exceeding the longest decision window time Afterwards, the process must terminate and provide a default conclusion (usually a timeout is considered as agreement to the local action), preventing the system from getting stuck in an infinite loop due to network disturbances or node failures, thus ensuring real-time performance.

[0032] This multi-round negotiation agreement is achieved through the following steps in a cycle: Round 0 (Initialization): Local triggering: Fault fingerprint extraction module 21 generates local fault feature data. and confidence level ; Broadcast proposal: Deterministic communication module 22 will This "proposal" is multicast to all neighboring nodes; Start timer: Simultaneously, start the current round's timer (duration is...). ) and total timeout timer (duration is ); Rounds 1 to K (cyclical negotiation): Receiving and summarizing: During the waiting time of each round, collect the fault data and confidence levels sent by other nodes; Weighted fusion calculation: Based on all collected information (including previous information), a weighted fusion algorithm is executed to calculate an updated global confidence score. ; Decision-making checks: such as ≥Action Threshold: This indicates that the collaborative conclusion is highly certain that a fault exists, generates an agreement trigger instruction, and immediately jumps to the final step; if ≤ Abandon Threshold: This indicates that the collaborative conclusion is highly confident that there is no fault or that the fault is minor. An anti-trigger instruction is generated, and the process immediately jumps to the final step.

[0033] If the threshold is abandoned < Action threshold: This indicates that the current information is still ambiguous and a clear consensus cannot be reached; Rollback and renegotiation: When the timer expires in the current round and the node is in the aforementioned ambiguous state, it rolls back to the next round. The node uses the current merged result (or its own updated result based on the new data) as a new proposal, broadcasts it again, and starts the next round of negotiation. Round check: If the current round number has reached the maximum preset round number K, the loop will be forcibly terminated and treated as a timeout. Final output decision: If an explicit agree or disagree instruction is received in the loop, then that instruction is output.

[0034] If due to reaching round K or total time If the loop exits due to a timeout, a timeout signal is output (this means that by default, cooperative decision-making does not oppose local actions).

[0035] The deterministic communication module 22 is based on the Byzantine fault-tolerant decision-making mechanism, and only when it collects no less than [a certain number of] data within the longest decision-making time window will it be considered valid. Only when there are multiple valid and opposing collaborative decision instructions issued by different electrical cabinet nodes is the current collaborative decision deemed to be an opposing trigger. The maximum number of malicious nodes that the system is allowed to tolerate is preset. Ordinary consensus protocols cannot cope with nodes sending malicious or forged error messages due to hacking or serious hardware failures (e.g., reporting "normal" when a real failure occurs, or inducing other nodes to act incorrectly). This solution introduces a Byzantine fault-tolerant decision-making mechanism, which enables the system to tolerate up to f such malicious nodes and still make correct decisions.

[0036] This means that an objection command cannot come from just one node; it must be valid only if a quorum of nodes collectively provide proof. This prevents a single malicious node from blocking necessary protective actions by forging objection commands.

[0037] During multi-round negotiation, when sending messages, the electrical cabinet node must digitally sign the messages (especially the objection command) using its private key; when receiving messages, the electrical cabinet node verifies the validity of the sender's signature. Messages with invalid signatures are discarded directly.

[0038] At the same time, the conditions for opposing the directive to take effect are strengthened: When an electrical cabinet node receives an objection to triggering, it cannot accept it immediately; it must check whether the objection is a signed set, and whether the signatures in the set come from at least f+1 different valid electrical cabinet nodes.

[0039] According to the Byzantine fault tolerance theory, for a system to tolerate f malicious electrical cabinet nodes, it needs at least 3f+1 total electrical cabinet nodes, and reaching consensus requires the agreement of 2f+1 electrical cabinet nodes. A valid objection instruction must contain at least f+1 signatures, ensuring that even if f malicious signatures are mixed in, at least one is still a genuine objection from an honest electrical cabinet node. This makes it impossible for a malicious electrical cabinet node to forge a valid objection instruction on its own.

[0040] Process Example: A fault was detected in electrical cabinet node A, triggering a broadcast proposal.

[0041] For a malicious electrical cabinet node B (controlled by a hacker) to prevent action, it must forge an objection command.

[0042] However, electrical cabinet node B requires the signatures of at least f+1 electrical cabinet nodes. It can only forge its own signature, not the signatures of other honest electrical cabinet nodes.

[0043] Therefore, its single objection command will be considered invalid and ignored by other electrical cabinet nodes.

[0044] Only when a genuine fault does not occur, and a sufficient number (≥f+1) of honest electrical cabinet nodes issue a signed objection instruction, will this instruction be accepted by the network, thus correctly preventing a malfunction.

[0045] The timeout monitoring module 23 is used to monitor the timeout status and negotiation rounds of the longest decision time window. If no cooperative protection decision instruction against triggering is received from the deterministic communication module 22 before the longest decision time window expires or after the maximum number of negotiation rounds is reached, and the local fault confidence level is not lower than the preset fault confidence level threshold, a trigger signal is generated. The protection execution unit 3 is used to respond to the trigger signal, thereby driving the local circuit breaker to trip and achieve fault isolation.

[0046] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely preferred examples and are not intended to limit the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the claimed invention.

Claims

1. An IoT-based remote monitoring and communication system for electrical cabinets, characterized in that, include: Data sampling unit (1), the data sampling unit (1) is used to perform high-speed sampling on local electrical cabinet nodes and generate sampling data with hardware timestamp; Bounded backoff unit (2), the bounded backoff unit (2) is used to complete the collaborative protection decision with the adjacent electrical cabinet node based on the sampled data and using the D-CPB mechanism within the preset longest decision time window, and use a multi-round negotiation protocol to determine whether to generate a collaborative protection decision instruction against triggering. If no collaborative protection decision instruction against triggering is received before the end of the longest decision time window or after the maximum number of negotiation rounds, and the local fault confidence is not lower than the preset fault confidence threshold, then a trigger signal is generated. The protection execution unit (3) is used to respond to the trigger signal, thereby driving the local circuit breaker to open and achieve fault isolation.

2. The IoT-based remote monitoring and communication system for electrical cabinets according to claim 1, characterized in that: The bounded rollback unit (2) includes a fault fingerprint extraction module (21), a deterministic communication module (22), and a timeout monitoring module (23). Among them, the fault fingerprint extraction module (21) is used to extract a compressed fingerprint representing the fault characteristics from the sampled data when a preset trigger condition is detected, and to generate an emergency summary and a local fault confidence level. The deterministic communication module (22) is used to multicast an emergency summary containing fault characteristics to the adjacent electrical cabinet node in the form of an emergency message. At the same time, it receives and verifies the summary information of the adjacent electrical cabinet node. It adopts a multi-round negotiation protocol and executes a weighted fusion algorithm to generate a global confidence score within the longest decision time window. Based on the global confidence score, it determines whether to generate a cooperative protection decision instruction against triggering. The timeout monitoring module (23) is used to monitor the timeout status and negotiation rounds of the longest decision time window. If no cooperative protection decision instruction opposing triggering is received from the deterministic communication module (22) before the longest decision time window expires or after the maximum negotiation rounds are reached, and the local fault confidence level is not lower than the preset fault confidence level threshold, then a trigger signal is generated.

3. The IoT-based remote monitoring and communication system for electrical cabinets according to claim 2, characterized in that: The preset triggering conditions are amplitude threshold triggering, slope threshold triggering, and model scoring triggering.

4. The IoT-based remote monitoring and communication system for electrical cabinets according to claim 3, characterized in that: The deterministic communication module (22) adopts a priority scheduling strategy, using a high-priority channel to transmit emergency summaries and a low-priority channel to transmit regular monitoring data.

5. The IoT-based remote monitoring and communication system for electrical cabinets according to claim 4, characterized in that: The emergency summary carries a sequence number and a hardware timestamp during transmission.

6. The IoT-based remote monitoring and communication system for electrical cabinets according to claim 5, characterized in that: The deterministic communication module (22) retransmits the emergency summary after the first transmission of the emergency summary according to a predetermined number of repetitions and retry intervals.

7. The IoT-based remote monitoring and communication system for electrical cabinets according to claim 6, characterized in that: In the deterministic communication module (22), the longest decision time window is the maximum allowable time from the first time the electrical cabinet sends the emergency summary to the time when the summary of the adjacent electrical cabinet node is received and the preliminary fusion judgment is completed.

8. The IoT-based remote monitoring and communication system for electrical cabinets according to claim 7, characterized in that: In the deterministic communication module (22), a multi-round negotiation protocol is adopted. Within the longest decision time window, a weighted fusion algorithm is executed to generate a global confidence score. The specific steps for determining whether to generate a collaborative protection decision instruction against triggering based on the global confidence score are as follows: S21. In each round of negotiation window, receive and verify the digital signature of the emergency summary information sent from the adjacent electrical cabinet node, and only collect the verified information; The emergency summary information includes at least the fault confidence level of the sending electrical cabinet node and the identification identifier of the electrical cabinet node; S22. Based on the electrical cabinet node identification, assign dynamic weights to each electrical cabinet node; wherein, the dynamic weight assignment strategy is determined jointly by the node's real-time response status and historical reliability records; S23. Calculate the weighted global confidence score for the current round based on the confidence score and the assigned weights. ; S24. Define the collaborative abandonment threshold. and cooperative action threshold global confidence With the threshold of collaborative abandonment and cooperative action threshold The comparison is performed to determine whether to generate a collaborative protection decision instruction against triggering.

9. The IoT-based remote monitoring and communication system for electrical cabinets according to claim 8, characterized in that: In step S24, the specific steps for determining whether to generate a cooperative protection decision instruction against triggering are as follows: If global confidence Less than or equal to the collaborative abandonment threshold If so, a collaborative decision instruction opposing the triggering is generated, and after being signed by this node, it is broadcast to neighboring nodes; If global confidence Higher than or equal to the cooperative action threshold If so, no opposition command will be generated in this round; If global confidence Above the collaborative abandonment threshold And global confidence Below the threshold for coordinated action If no clear consensus is reached in this round, then the next round of negotiations will begin; If the current round has reached the preset maximum negotiation round, the negotiation process will be terminated.

10. The IoT-based remote monitoring and communication system for electrical cabinets according to claim 9, characterized in that: The deterministic communication module (22) is based on the Byzantine fault-tolerant decision-making mechanism and only collects no less than [number missing] within the longest decision time window. Only when there are multiple valid and opposing collaborative decision instructions issued by different electrical cabinet nodes is the current collaborative decision deemed to be an opposing trigger. This is a pre-defined maximum number of malicious nodes that are allowed to be fault-tolerant.

Images