CN-122026951-A - Power line communication self-healing method and system based on immune mechanism

CN122026951ACN 122026951 ACN122026951 ACN 122026951ACN-122026951-A

Abstract

The invention provides a power line communication self-healing method and a system based on an immune mechanism, wherein the method comprises the steps of obtaining power line communication data, constructing digital communication cells based on the power line communication data, obtaining cell state data of the digital communication cells after communication transmission, constructing and training a communication immune model comprising a channel detection sub-module and a decoupling analysis sub-module, carrying out damage detection and root cause analysis on the cell state data based on the communication immune model, generating a damage root cause analysis result, carrying out damage positioning according to the damage root cause analysis result, and carrying out damage healing based on a grading collaborative self-healing mechanism, thereby realizing accurate positioning of abnormal root causes and rapid restoration of communication abnormalities.

Inventors

YANG CHEN
CHEN JUNJIAN
ZHANG WEI
TAO WEI
YIN FURONG
LIU ZILONG

Assignees

南方电网数字电网研究院股份有限公司

Dates

Publication Date: 20260512
Application Date: 20260211

Claims (8)

1. The power line communication self-healing method based on the immune mechanism is characterized by comprising the following steps of: Acquiring power line communication data, constructing a digital communication cell based on the power line communication data, and acquiring cell state data of the digital communication cell after communication transmission; Constructing and training a communication immunity model comprising a channel detection sub-module and a decoupling analysis sub-module; performing damage detection and root cause analysis on the cell state data based on the communication immune model to generate a damage root cause analysis result; and (3) carrying out damage positioning according to the analysis result of the root cause of the damage, and carrying out damage healing based on a grading cooperative self-healing mechanism.
2. The method for self-healing power line communication based on an immune mechanism according to claim 1, wherein obtaining power line communication data, constructing digital communication cells based on the power line communication data, obtaining cell state data of the digital communication cells after communication transmission, comprises: collecting service data of a service to be transmitted and current channel state data, and injecting an immune repair gene for the transmission service based on the service data and the channel state data, wherein the immune repair gene is used for guiding the damage healing process of the service data; carrying out hash operation on the service data to generate a global integrity tag; logically dividing service data into a plurality of data blocks, and endowing each data block with a check value as a local positioning anchor point of the data block; packaging the service data, the channel state data, the immune repair gene, the global integrity tag and the local positioning anchor point sequence into a digital communication cell; After receiving the transmitted digital communication cells, the receiving end performs data verification according to the global integrity tag and the local positioning anchor point sequence to generate cell state data.
3. The method for self-healing power line communication based on immune mechanism according to claim 1, wherein constructing and training a communication immune model comprising a channel detection sub-module and a decoupling analysis sub-module comprises: constructing a channel detection submodule by adopting a space-time diagram convolution network and a multi-head attention mechanism, wherein the channel detection submodule is used for reasoning about the change of the communication network environment and the abnormal root cause of the cell state data; adopting historical communication operation data as a first training data set, and taking preset simulation data as a second training data set; Training the channel detection sub-module by taking the minimum joint loss as a target based on a preset multi-task joint loss function by adopting a progressive course learning strategy; the preset multi-task joint loss function comprises quantile loss between a predicted sequence and a true value, focus loss of node state classification and contrast loss based on a graph structure.
4. A self-healing method for power line communication based on an immune mechanism according to claim 3, wherein the method further comprises: Constructing a decoupling analysis sub-module by adopting a multi-relationship diagram attention network and a neural logic network, wherein the decoupling analysis sub-module is used for reasoning the association cooperative relationship among the digital communication cells; a step of pre-training the decoupling analysis submodule by adopting a step course learning strategy and synchronously combining the first training data set and the second training data set; the stage course learning strategy at least comprises atlas reconstruction training and utility fine adjustment; the graph mask self-encoder is adopted in the graph reconstruction training stage, and the graph reconstruction training stage is trained by taking the reconstruction loss of the minimized nodes and the prediction loss of the edges as loss functions; The utility fine tuning stage trains by taking the mean square error of the correlation and coordination relation between the minimized predictive digital communication cells and the actual true value as a loss function.
5. The immune mechanism-based power line communication self-healing method of claim 1, wherein the step of performing damage detection and root cause analysis on the cell state data based on the communication immune model to generate a damage root cause analysis result comprises the steps of: locating damaged digital communication cells based on the cell status data, generating a set of damaged digital communication cells; Inputting the damaged digital communication cell set into a decoupling analysis submodule for reasoning to obtain a damage associated cell map; the damage association cytogram is expressed as an edge connection structure formed by converting K candidate cell nodes with highest association degree with damaged digital communication cells and association relations thereof; Inputting the damaged digital communication cell set and the cell state data into a channel detection submodule to carry out cell damage root cause reasoning and obtain an original damage root cause analysis result; and carrying out reinforced reasoning based on the damage-associated cytogram and the original damage root cause analysis result to generate a damage root cause analysis result.
6. The immune mechanism-based power line communication self-healing method of claim 5, wherein the generating of the injury root cause analysis result based on the enhanced reasoning of the injury-associated cell map and the original injury root cause analysis result comprises: analyzing the influence of the damage reasoning set in the original damage root cause analysis result on the damage mode in the damage-associated cell map by adopting a Bayesian network, and generating a corresponding diagnosis confidence; Screening an original damage root cause analysis result based on the diagnosis confidence, and taking damage reasoning reaching a preset screening condition as a first damage root cause; generating an exploring digital communication cell based on the first injury root cause, and acquiring exploring cell state data; And inputting the explored cell state data as reinforced evidence into a Bayesian network for reinforced reasoning to generate a damage root cause analysis result.
7. The self-healing method of power line communication based on immune mechanism according to claim 1, wherein the method for performing the lesion localization according to the analysis result of the root cause of the lesion and performing the lesion healing based on the hierarchical collaborative self-healing mechanism comprises the following steps: Classifying the damage root cause analysis results by adopting a multi-layer perceptron to obtain corresponding damage positioning labels; The injury positioning label comprises cell self-healing, cooperative healing and virtual reconstruction; For the damage of which the label is cell self-healing, directly calling corresponding immune repair genes in the digital communication cells to carry out self-repair; For the damage with the label of cooperative healing, adopting a graph attention network and a preset resource game mechanism to perform cooperative healing; and for the damage of which the label is virtual reconstruction, discarding the data sequence corresponding to the damaged digital communication cell, and generating a corresponding reconstruction sequence to replace the discarding part.
8. A power line communication self-healing system based on an immune mechanism for implementing the method of any one of claims 1 to 7, comprising: the cell generation module is used for generating digital communication cells according to the power line communication data and synchronously acquiring cell state data of the digital communication cells after communication transmission; The model construction module is used for constructing and training a communication immunity model, and the communication immunity model comprises a channel detection sub-module and a decoupling analysis sub-module; The damage detection module is used for carrying out damage detection and root cause analysis on the digital communication cells by using the communication immune model to generate a damage root cause analysis result; The damage healing module is used for carrying out damage positioning according to the analysis result of the root cause of the damage and carrying out damage healing based on a hierarchical cooperative self-healing mechanism.

Description

Power line communication self-healing method and system based on immune mechanism Technical Field The invention relates to the technical field of communication, in particular to a power line communication self-healing method and system based on an immune mechanism. Background In the current increasingly complex heterogeneous converged network environment, the reliability of data transmission and the continuity of service face serious challenges. On one hand, the traditional method mainly depends on preset fixed redundancy or simple retry operation, and is often lack of joint analysis and deep mining according to the dynamic channel state and the semantic features of service data, so that communication repair efficiency is low or resources are hollow in a good channel under burst interference never caused, and on the other hand, the traditional fault detection and recovery method is often dependent on end-to-end timeout retransmission or centralized network management alarm, and is often difficult to distinguish whether link transmission errors, node processing faults or application layer data logic errors are caused, so that communication repair efficiency is low. Disclosure of Invention In view of the above-mentioned problems, in combination with the first aspect of the present invention, an embodiment of the present invention provides a self-healing method for power line communication based on an immune mechanism, the method including: Acquiring power line communication data, constructing a digital communication cell based on the power line communication data, and acquiring cell state data of the digital communication cell after communication transmission; Constructing and training a communication immunity model comprising a channel detection sub-module and a decoupling analysis sub-module; performing damage detection and root cause analysis on the cell state data based on the communication immune model to generate a damage root cause analysis result; and (3) carrying out damage positioning according to the analysis result of the root cause of the damage, and carrying out damage healing based on a grading cooperative self-healing mechanism. As a further aspect of the present invention, obtaining power line communication data, constructing a digital communication cell based on the power line communication data, obtaining cell state data of the digital communication cell after communication transmission, includes: collecting service data of a service to be transmitted and current channel state data, and injecting an immune repair gene for the transmission service based on the service data and the channel state data, wherein the immune repair gene is used for guiding the damage healing process of the service data; carrying out hash operation on the service data to generate a global integrity tag; logically dividing service data into a plurality of data blocks, and endowing each data block with a check value as a local positioning anchor point of the data block; packaging the service data, the channel state data, the immune repair gene, the global integrity tag and the local positioning anchor point sequence into a digital communication cell; After receiving the transmitted digital communication cells, the receiving end performs data verification according to the global integrity tag and the local positioning anchor point sequence to generate cell state data. As a further aspect of the present invention, constructing and training a communication immunity model including a channel detection sub-module and a decoupling analysis sub-module includes: constructing a channel detection submodule by adopting a space-time diagram convolution network and a multi-head attention mechanism, wherein the channel detection submodule is used for reasoning about the change of the communication network environment and the abnormal root cause of the cell state data; adopting historical communication operation data as a first training data set, and taking preset simulation data as a second training data set; Training the channel detection sub-module by taking the minimum joint loss as a target based on a preset multi-task joint loss function by adopting a progressive course learning strategy; the preset multi-task joint loss function comprises quantile loss between a predicted sequence and a true value, focus loss of node state classification and contrast loss based on a graph structure. As a further aspect of the present invention, the method further includes: Constructing a decoupling analysis sub-module by adopting a multi-relationship diagram attention network and a neural logic network, wherein the decoupling analysis sub-module is used for reasoning the association cooperative relationship among the digital communication cells; a step of pre-training the decoupling analysis submodule by adopting a step course learning strategy and synchronously combining the first training data set and the second training data set; the stag