A power multi-protocol adaptive conversion system and method based on open source honkong

Abstract

The application discloses a kind of power multi-protocol adaptive conversion systems and methods based on open source Hong Meng.It is disclosed that the system includes protocol analysis module, intelligent conversion engine module, online machine learning model module, genetic algorithm adaptation module, secure communication interface and machine learning optimization module, is deployed in the communication network of power system, is connected with various power equipment through network interface, simultaneously with dispatching master station, intelligent terminal etc. Communication;System uses microservice architecture, each module is independently deployed, and communicates and data interaction is carried out through distributed soft bus.The system of the application can collect and deeply analyze the communication data of heterogeneous devices in the power system in real time, realize the adaptive conversion of multiple protocols, improve the efficiency and accuracy of protocol conversion, while ensuring the security and integrity of data, to meet the needs of intelligent development of power system.

Description

Technical Field

[0001] The present invention belongs to the field of power system communication technology, and specifically relates to an open source Hongmeng-based power multi-protocol adaptive conversion system and method. Background Art

[0002] Power systems are populated by a vast array of equipment from diverse manufacturers and models, often using different communication protocols for data transmission, such as IEC60870-5-104, DNP3, and Modbus-TCP. This heterogeneity complicates communication between devices, hindering effective data integration and sharing, and hindering the intelligent development of power systems.

[0003] Existing power protocol conversion technology has several limitations. For one thing, some conversion systems lack the ability to deeply analyze multiple protocols, making them unable to accurately identify and process complex protocol messages. Furthermore, conversion strategies are inflexible in the face of dynamically changing power scenarios, making it difficult to adapt to meet diverse business needs. Furthermore, existing systems also suffer from deficiencies in data security and transmission efficiency, failing to effectively guarantee the integrity and confidentiality of power data.

[0004] Therefore, there is an urgent need for an efficient, intelligent, and safe power multi-protocol adaptive conversion system and method to solve the problem of protocol incompatibility in power systems. Summary of the Invention

[0005] In response to the shortcomings of the existing technology, one of the purposes of the present invention is to provide an open source Hongmeng-based power multi-protocol adaptive conversion system, which can collect and deeply analyze the communication data of heterogeneous devices in the power system in real time, realize adaptive conversion of multiple protocols, improve the efficiency and accuracy of protocol conversion, and ensure the security and integrity of data at the same time, so as to meet the needs of intelligent development of the power system.

[0006] The second object of the present invention is to provide an electric power protocol conversion method based on the open source Hongmeng electric power multi-protocol adaptive conversion system.

[0007] The present invention provides an open-source Hongmeng-based power multi-protocol adaptive conversion system, which includes a protocol parsing module, an intelligent conversion engine module, an online machine learning model module, a genetic algorithm adaptation module, a secure communication interface, and a machine learning optimization module.

[0008] The power multi-protocol adaptive conversion system based on open source Hongmeng is deployed in the communication network of the power system, connected to various power equipment through the network interface, and communicates with the dispatching master station, smart terminals, etc.

[0009] The system collects communication data through a protocol parsing module, extracts key information, and uploads the data to a genetic algorithm adaptation module. The genetic algorithm adaptation module uses a multi-objective optimization algorithm to search for the optimal solution of the protocol field mapping relationship based on the protocol feature vector, and generates a set of conversion strategies for different power scenarios. The online machine learning model module predicts the QoS index of each strategy in the conversion strategy set, and the TOPSIS multi-attribute decision algorithm selects the optimal conversion strategy. The intelligent conversion engine module performs protocol field mapping and data conversion through a dynamic protocol mapping unit based on the optimal conversion strategy. The secure communication interface encapsulates and encrypts the converted data for transmission, completing the adaptive conversion of power multi-protocols. The machine learning optimization module continuously analyzes historical protocol interaction characteristics, dynamically adjusts the protocol conversion delay and data integrity threshold, and updates the adaptive rule base and chromosome encoding rule base in real time according to network congestion status and business needs.

[0010] The protocol analysis module collects communication data of power equipment in real time, performs in-depth analysis according to the specifications of different protocols, uses the protocol fingerprint feature extraction algorithm to identify the communication protocol type, and constructs the protocol feature vector;

[0011] The intelligent conversion engine module is built on the open source Hongmeng distributed soft bus technology and includes a dynamic protocol mapping unit and an adaptive rule base;

[0012] The dynamic protocol mapping unit builds a protocol metadata database based on the open source Hongmeng distributed data management component; the protocol metadata database stores metadata of various protocols, including the definition, data type, value range, etc. of protocol fields, providing basic data support for protocol mapping;

[0013] The dynamic protocol mapping unit performs cross-protocol timing data alignment based on the DTW algorithm and introduces a pruning strategy and a fast search algorithm, specifically:

[0014] Suppose there are two time series of data and , build a The distance matrix D of It represents the cumulative distance between the first i elements in the time series X and the first j elements in Y, and is calculated using the following formula:

[0015] ;

[0016] in, is the local distance between elements xi and yj; cross-protocol timing data alignment is achieved based on the calculated cumulative distance;

[0017] The dynamic protocol mapping unit dynamically adjusts the mapping strength of the fields according to the uncertainty and importance of the protocol fields by a protocol entropy weight value analysis method, specifically:

[0018] For a protocol containing n fields to be mapped, the frequency of occurrence of each value of the i-th field is counted from historical communication data; assuming that the field has m different possible values, the probability estimate value of the j-th value is ; the number of times that the j-th possible value appears is divided by the total number of times that all values appear to calculate ;

[0019] Based on the calculated , the entropy value Ei of the i-th field is calculated, and the entropy value Ei is calculated using the following formula: ;

[0020] wherein k is a normalization coefficient, and , m is the number of different values that the protocol field can have;

[0021] The entropy weight value wi of the i-th field is calculated using the following formula:

[0022] ;

[0023] The obtained entropy weight value is used for subsequent field similarity matching calculation, and the field with a high entropy weight value has a higher priority in the mapping decision;

[0024] The adaptive rule base stores various protocol conversion rules, and dynamically adjusts the conversion strategy according to different power scenarios and business needs;

[0025] The machine learning optimization module integrates a time series prediction model based on an attention mechanism, which is used to analyze historical protocol interaction features and dynamically optimize protocol conversion latency and data integrity thresholds, specifically:

[0026] An LSTM-GAN (Long Short-Term Memory Generative Adversarial Network) hybrid neural network model is used to improve the generalization recognition ability of non-standard power protocol features through generative adversarial training; the LSTM network is used to process time series data and can capture the long-term dependence of data; the GAN network consists of a generator and a discriminator, which learn the distribution characteristics of data through adversarial training; the generator G attempts to generate samples similar to real data, and the discriminator D attempts to distinguish between generated samples and real samples, and both are continuously optimized in the adversarial process;

[0027] A protocol semantic graph is constructed, and a graph convolution network (GCN) is used to extract topological correlation features between protocols. The protocol semantic graph represents the protocols and their fields as nodes and edges of a graph. The GCN is used to perform convolution operations on the graph to extract feature representations of the nodes, thereby mining potential correlations between the protocols. The propagation rule of the GCN is as follows: ;

[0028] wherein, is an augmented adjacency matrix, which is the original adjacency matrix A plus the identity matrix I; is the degree matrix of ; is a node feature matrix of the first layer; is a to-be-learned weight matrix of the first layer; is an activation function;

[0029] The LSTM-GAN hybrid neural network model predicts the network congestion level and system load in a future preset time period, makes a decision based on the prediction result, dynamically adjusts the protocol conversion priority, and avoids high latency.

[0030] The dynamic adjustment of the protocol conversion priority is specifically as follows: a reinforcement learning strategy based on Q-learning is designed, the protocol conversion priority is dynamically adjusted according to the network congestion state, the state s represents the network congestion state, the action a represents the adjustment strategy of the protocol conversion priority, and the update formula of the Q value is as follows:

[0031]

[0032] wherein, is a learning rate; r is an immediate reward; is a discount factor; is the next state; is a candidate action of all possible actions taken by the agent in the next state The candidate action is one of all possible actions taken by the agent.

[0033] When the LSTM-GAN hybrid neural network model predicts that the network reliability decreases and the packet loss risk increases, the error correction intensity, the retransmission threshold, and the packet size are adjusted, the confirmation timeout time is dynamically set according to the predicted latency, and the data integrity threshold is optimized.

[0034] The genetic algorithm adaptation module uses a multi-objective optimization algorithm to search for a Pareto optimal solution of the protocol field mapping relationship, generates a conversion strategy set for different power scenarios, and specifically as follows:

[0035] The protocol field matching degree, the conversion energy consumption, and the transmission latency are defined as optimization objective functions. The protocol field matching degree is f1, the conversion energy consumption is f2, and the transmission latency is f3. Therefore, the objective function vector is ;

[0036] The NSGA-III (Non-dominated Sorting Genetic Algorithm III) algorithm is used for non-dominated sorting under multi-dimensional constraints, and a group of non-dominated solutions are screened out as the Pareto front through fast non-dominated sorting and congestion distance calculation, so as to generate a Pareto front solution set.

[0037] The secure communication interface realizes encapsulation and encrypted transmission of data after protocol conversion based on an open-source microkernel architecture of Hongmeng, supports multi-modal communication with a dispatching master station and an intelligent terminal, and adopts a symmetric encryption algorithm (such as AES) to encrypt data, so as to ensure the confidentiality and integrity of data in the transmission process.

[0038] The application further provides a power protocol conversion method based on the system, comprising the following steps:

[0039] S1. A device communication link is established through open-source Hongmeng heterogeneous networking technology, power devices of different protocols are connected to the system, and communication data of the devices is collected in real time;

[0040] S2. A protocol fingerprint feature extraction algorithm is used to identify the type of communication protocol and construct a protocol feature vector;

[0041] S3. A genetic algorithm optimization module is called to generate a candidate conversion strategy set, and an online machine learning model is combined to predict an optimal strategy;

[0042] S4. Based on a micro-service architecture, a dynamic conversion rule is deployed in an intelligent conversion engine module, and a distributed transaction mechanism is used to ensure the consistency of cross-protocol transactions;

[0043] S5. A differential privacy technology is used to desensitize sensitive power data, and secure transmission after protocol conversion is completed.

[0044] In step S2, the protocol fingerprint feature extraction algorithm identifies the type of communication protocol by analyzing the format, field value and frequency of the message.

[0045] In a specific implementation, a machine learning algorithm (such as a support vector machine) is used to classify message features and construct a protocol feature vector.

[0046] Step S3 specifically comprises:

[0047] The genetic algorithm optimization module is used to generate a candidate conversion strategy set;

[0048] A protocol conversion quality evaluation index system is constructed; the quality evaluation indexes include data fidelity, time delay jitter rate and resource occupation rate; the data fidelity measures the similarity between the converted data and the original data, the time delay jitter rate reflects the time delay fluctuation in the data transmission process, and the resource occupation rate represents the computing resources and network resources consumed by the system in the protocol conversion process;

[0049] An XGBoost regression model is applied to predict the QoS indexes of each strategy according to the feature vectors of the strategies;

[0050] An optimal conversion strategy is selected from the strategies by a TOPSIS algorithm, specifically as follows:

[0051] Supposing that there are m candidate strategies, each strategy has n evaluation indexes, an initial decision matrix is constructed , and vector normalization processing is performed to obtain a normalized matrix .

[0052] According to the benefit type or cost type attributes of each evaluation index, a positive ideal solution and a negative ideal solution are determined, and are expressed by the following formulas:

[0053]

[0054]

[0055] wherein, is the value of the positive ideal solution on the jth index; is the value of the negative ideal solution on the jth index; is a set of benefit type indexes, is a set of cost type indexes;

[0056] The relative closeness of each strategy is calculated by the following formula:

[0057]

[0058] wherein, is the relative closeness of the ith strategy; is the distance from the ith strategy to the positive ideal solution, is the distance from the ith strategy to the negative ideal solution, which is calculated by the following formula:

[0059]

[0060] The strategies are sorted according to the relative closeness , and the strategy with the maximum relative closeness is selected as the optimal conversion strategy.

[0061] Step S4 is specifically as follows:​

[0062] The dynamic conversion rule is deployed through the intelligent conversion engine module, and protocol conversion is performed according to an optimal conversion strategy;

[0063] A two-phase commit protocol (2PC) is adopted, a timeout processing and a retry mechanism are added, and a blockchain smart contract is introduced for transaction state tracking; the blockchain smart contract is used for recording the state and execution result of the transaction, so as to ensure the non-tamperability and traceability of the transaction;

[0064] When an exception occurs in the protocol conversion process, a rollback compensation operation is triggered in an event-driven manner to restore the system state to the state before transaction execution.

[0065] In step S5, after protocol conversion, sensitive power data (such as user power consumption information, device operating parameters, etc.) is subjected to differential privacy processing, and then encrypted transmission is performed through a secure communication interface; the differential privacy technology protects the privacy of data by adding noise to the data.

[0066] The application discloses a power multi-protocol adaptive conversion system and method based on open source Hongmeng. BRIEF DESCRIPTION OF DRAWINGS

[0067] Figure 1 It is a structural schematic diagram of the system of the application;

[0068] Figure 2 It is a flowchart of the method of the application. DETAILED DESCRIPTION

[0069] The application provides a power multi-protocol adaptive conversion system based on open source Hongmeng. Figure 1 As shown in the structural schematic diagram, the system comprises a protocol analysis module, an intelligent conversion engine module, an online machine learning model module, a genetic algorithm adaptation module, a secure communication interface and a machine learning optimization module.

[0070] The power multi-protocol adaptive conversion system based on open source Hongmeng is deployed in the communication network of a power system, connected with various power devices through a network interface, and communicates with a dispatching master station, an intelligent terminal and the like.

[0071] The system collects communication data through a protocol analysis module, extracts key information, and uploads the data to a genetic algorithm adaptation module; the genetic algorithm adaptation module searches for the optimal solution of the protocol field mapping relationship using a multi-objective optimization algorithm based on the protocol feature vector, generates a conversion strategy set for different power scenarios; an online machine learning model module predicts the QoS indicators of each strategy in the conversion strategy set, and a TOPSIS multi-attribute decision algorithm selects the optimal conversion strategy; an intelligent conversion engine module maps protocol fields and converts data through a dynamic protocol mapping unit based on the optimal conversion strategy; a secure communication interface encapsulates and encrypts the converted data for transmission, completing power multi-protocol adaptive conversion; a machine learning optimization module continuously analyzes historical protocol interaction features, dynamically adjusts protocol conversion latency and data integrity thresholds, and simultaneously updates the adaptive rule base and chromosome encoding rule base in real time based on network congestion state and service demand.

[0072] The protocol analysis module collects communication data of power equipment in real time, and performs deep analysis according to the specifications of different protocols, identifies the communication protocol type using a protocol fingerprint feature extraction algorithm, and constructs a protocol feature vector;

[0073] The intelligent conversion engine module is based on the open source Hongmeng distributed soft bus technology, and includes a dynamic protocol mapping unit and an adaptive rule base;

[0074] The dynamic protocol mapping unit is based on the open source Hongmeng distributed data management component and constructs a protocol metadata database; the protocol metadata database stores the metadata of various protocols, including the definition of protocol fields, data types, value ranges, etc., providing basic data support for protocol mapping;

[0075] The dynamic protocol mapping unit performs cross-protocol time series data alignment based on the DTW algorithm, and introduces pruning strategies and fast search algorithms, specifically:

[0076] Let the time series of two data be X and Y and , construct a distance matrix D of ; where, represents the cumulative distance between the first i elements in the time series X and the first j elements in Y, which is calculated using the following formula:

[0077] ;

[0078] where, is the local distance between elements xi and yj; the cumulative distance is calculated according to

[0079] The dynamic protocol mapping unit dynamically adjusts the mapping strength of the fields according to the uncertainty and importance of the protocol fields by a protocol entropy weight value analysis method, specifically:

[0080] For a protocol containing n fields to be mapped, the frequency of occurrence of each value of the i-th field is counted from historical communication data; assuming that the field has m different possible values, the probability estimate value of the j-th value is ; the number of times that the j-th possible value appears is divided by the total number of times that all values appear to calculate ;

[0081] Based on the calculated , the entropy value Ei of the i-th field is calculated, and the entropy value Ei is calculated using the following formula: ;

[0082] wherein k is a normalization coefficient, and , m is the number of different values that can appear in the protocol field;

[0083] The entropy weight value wi of the i-th field is calculated using the following formula:

[0084] ;

[0085] The obtained entropy weight value is used for subsequent field similarity matching calculation, and the field with a high entropy weight value has a higher priority in the mapping decision;

[0086] The adaptive rule base stores various protocol conversion rules, and dynamically adjusts the conversion strategy according to different power scenarios and business needs;

[0087] The machine learning optimization module integrates a time series prediction model based on an attention mechanism, which is used to analyze historical protocol interaction features and dynamically optimize protocol conversion latency and data integrity thresholds, specifically:

[0088] An LSTM-GAN (Long Short-Term Memory Generative Adversarial Network) hybrid neural network model is used to improve the generalization recognition ability of non-standard power protocol features through generative adversarial training; the LSTM network is used to process time series data and can capture the long-term dependence of data; the GAN network consists of a generator and a discriminator, which learn the distribution characteristics of data through adversarial training; the generator G attempts to generate samples similar to real data, and the discriminator D attempts to distinguish between generated samples and real samples, and both are continuously optimized in the adversarial process;

[0089] A protocol semantic graph is constructed, and a graph convolution network (GCN) is used to extract topological correlation features between protocols. The protocol semantic graph represents the protocols and their fields as nodes and edges of a graph. The GCN is used to perform convolution operations on the graph to extract feature representations of the nodes, thereby mining potential correlations between the protocols. The propagation rule of the GCN is as follows: ;

[0090] wherein, is an augmented adjacency matrix, which is the original adjacency matrix A plus the identity matrix I; is a degree matrix of the original adjacency matrix A; is a node feature matrix of the first layer; is a to-be-learned weight matrix of the first layer; is an activation function;

[0091] The LSTM-GAN hybrid neural network model predicts the network congestion degree and system load in a future preset time period, makes a decision based on the prediction result, dynamically adjusts the protocol conversion priority, and avoids high latency.

[0092] The dynamic adjustment of the protocol conversion priority is specifically as follows: a reinforcement learning strategy based on Q-learning is designed, the protocol conversion priority is dynamically adjusted according to the network congestion state, the state s represents the network congestion state, the action a represents the adjustment strategy of the protocol conversion priority, and the update formula of the Q value is as follows:

[0093]

[0094] wherein, is a learning rate; r is an immediate reward; is a discount factor; is a next state; is a reward value in the next state A candidate action is selected from all possible actions of the agent;

[0095] When the LSTM-GAN hybrid neural network model predicts that the network reliability decreases and the packet loss risk increases, the error correction intensity, the retransmission threshold and the packet size are adjusted, the confirmation timeout time is dynamically set according to the predicted latency, and the data integrity threshold is optimized.

[0096] The genetic algorithm adaptation module uses a multi-objective optimization algorithm to search for a Pareto optimal solution of the protocol field mapping relationship, generates a conversion strategy set for different power scenarios, and specifically as follows:

[0097] The protocol field matching degree, the conversion energy consumption and the transmission latency are defined as optimization objective functions. The protocol field matching degree is f1, the conversion energy consumption is f2, and the transmission latency is f3. Therefore, the objective function vector is ;

[0098] The NSGA-III (Non-dominated Sorting Genetic Algorithm III) algorithm is used for non-dominated sorting under multi-dimensional constraints, and a group of non-dominated solutions are screened out as the Pareto front by fast non-dominated sorting and congestion distance calculation, so as to generate a Pareto front solution set.

[0099] The secure communication interface realizes encapsulation and encrypted transmission of data after protocol conversion based on an open-source microkernel architecture of Hongmeng, supports multi-modal communication with a scheduling master station and an intelligent terminal, and adopts a symmetric encryption algorithm (such as AES) to encrypt data, so as to ensure the confidentiality and integrity of data in the transmission process.

[0100] The application further provides a power protocol conversion method based on the system, a flowchart is shown in Figure 2 The method comprises the following steps:

[0101] S1. A device communication link is established by using open-source Hongmeng heterogeneous networking technology, power devices of different protocols are connected to the system, and communication data of the devices is collected in real time;

[0102] S2. A protocol fingerprint feature extraction algorithm is used to identify the type of communication protocol and construct a protocol feature vector;

[0103] In step S2, the protocol fingerprint feature extraction algorithm identifies the type of communication protocol by analyzing the format, field value, frequency and other characteristics of the message.

[0104] In a specific implementation, a machine learning algorithm (such as a support vector machine) is used to classify message features and construct a protocol feature vector.

[0105] S3. A genetic algorithm optimization module is called to generate a candidate conversion strategy set, and an online machine learning model is used to predict an optimal strategy;

[0106] Step S3 specifically comprises:

[0107] The genetic algorithm optimization module is used to generate a candidate conversion strategy set;

[0108] A protocol conversion quality evaluation index system is constructed; the quality evaluation indexes include data fidelity, delay jitter rate and resource occupation rate; the data fidelity measures the similarity between the converted data and the original data, the delay jitter rate reflects the delay fluctuation in the data transmission process, and the resource occupation rate represents the computing resources and network resources consumed by the system in the protocol conversion process;

[0109] An XGBoost regression model is used to predict the QoS indexes of each strategy according to the feature vector of the strategy;

[0110] The optimal conversion strategy is selected from each strategy by the TOPSIS algorithm, specifically:

[0111] Suppose there are m candidate strategies, each with n evaluation indexes, an initial decision matrix is constructed , and vector normalization processing is performed to obtain the normalized matrix ;

[0112] According to the benefit type or cost type attributes of each evaluation index, the positive ideal solution and the negative ideal solution are determined, which are expressed by the following formulas:

[0113]

[0114]

[0115] wherein, is the value of the positive ideal solution on the jth index; is the value of the negative ideal solution on the jth index; is the set of benefit type indexes, is the set of cost type indexes;

[0116] The benefit type evaluation indexes include data fidelity, throughput, and success rate; the cost type evaluation indexes include conversion latency, latency jitter rate, and resource occupancy rate;

[0117] The relative closeness of each strategy is calculated using the following formula:

[0118]

[0119] wherein, is the relative closeness of the ith strategy; is the distance from the ith strategy to the positive ideal solution, is the distance from the ith strategy to the negative ideal solution, which is calculated using the following formula:

[0120]

[0121] According to the relative closeness , the strategies are sorted, and the strategy with the maximum value is selected as the optimal conversion strategy.

[0122] S4. Based on the microservice architecture, deploy dynamic conversion rules in the intelligent conversion engine module, and ensure cross-protocol transaction consistency through a distributed transaction mechanism;

[0123] Step S4 is specifically:

[0124] ​Deploy dynamic conversion rules through the intelligent conversion engine module, and perform protocol conversion according to the optimal conversion strategy;

[0125] A two-phase commit protocol (2PC) is adopted, a timeout processing and retry mechanism are added, and a blockchain smart contract is introduced for transaction state tracking; the blockchain smart contract is used to record the state and execution result of the transaction, ensuring the non-tamperability and traceability of the transaction;

[0126] When an exception occurs during protocol conversion, a rollback compensation operation is triggered through event-driven mode to restore the system state to the state before transaction execution.

[0127] S5. Adopting differential privacy technology to desensitize sensitive power data for secure transmission after protocol conversion.

[0128] In step S5, after protocol conversion, sensitive power data (such as user electricity consumption information, device operating parameters, etc.) is subjected to differential privacy processing, and then encrypted transmission is performed through a secure communication interface; the differential privacy technology protects the privacy of data by adding noise to the data.

Claims

1. A power multi-protocol adaptive conversion system based on open source Hongmeng, characterized in that, The system comprises a protocol analysis module, an intelligent conversion engine module, an online machine learning model module, a genetic algorithm adaptation module, a secure communication interface and a machine learning optimization module; The open source Hongmeng-based power multi-protocol adaptive conversion system is deployed in a communication network of a power system, connected with various power equipment through a network interface, and communicates with a dispatching master station and an intelligent terminal; The system collects communication data through the protocol analysis module, extracts key information, and uploads the data to the genetic algorithm adaptation module; the genetic algorithm adaptation module searches for optimal solutions for protocol field mapping relationships using a multi-objective optimization algorithm based on protocol feature vectors, generates a conversion strategy set for different power scenarios; the online machine learning model module predicts the QoS indicators of each strategy in the conversion strategy set, and the TOPSIS multi-attribute decision algorithm selects the optimal conversion strategy; the intelligent conversion engine module maps protocol fields and converts data through a dynamic protocol mapping unit based on the optimal conversion strategy; the secure communication interface encapsulates and encrypts the converted data for transmission, completing the power multi-protocol adaptive conversion; The machine learning optimization module continuously analyzes historical protocol interaction features, dynamically adjusts protocol conversion latency and data integrity thresholds, and updates adaptive rule libraries and chromosome encoding rule libraries in real time according to network congestion states and business demands; The intelligent conversion engine module is based on open source Hongmeng distributed soft bus technology and includes a dynamic protocol mapping unit and an adaptive rule library; The dynamic protocol mapping unit is based on an open source Hongmeng distributed data management component and constructs a protocol meta-database; The protocol meta-database stores meta-information of various protocols, including definitions, data types, and value ranges of protocol fields, providing basic data support for protocol mapping; The dynamic protocol mapping unit performs cross-protocol time series data alignment based on the DTW algorithm and introduces pruning strategies and fast search algorithms; The dynamic protocol mapping unit dynamically adjusts the mapping strength of fields based on protocol entropy weight analysis according to the uncertainty and importance of protocol fields; The adaptive rule library stores various protocol conversion rules and dynamically adjusts conversion strategies according to different power scenarios and business demands.

2. The open source micro-mesh based power multi-protocol adaptive conversion system according to claim 1, wherein, The protocol analysis module collects communication data of power equipment in real time, performs in-depth analysis according to different protocol specifications, identifies communication protocol types using protocol fingerprint feature extraction algorithms, and constructs protocol feature vectors.

3. The open source micro-kernel based power multi-protocol adaptive conversion system according to claim 1, wherein, The genetic algorithm adaptation module uses a multi-objective optimization algorithm to search for Pareto optimal solutions for protocol field mapping relationships, generating a conversion strategy set for different power scenarios, specifically: The protocol field matching degree, conversion energy consumption and transmission time delay are defined as the optimization objective functions, the protocol field matching degree is f1, the conversion energy consumption is f2, and the transmission time delay is f3, and the objective function vector is ; The NSGA-III algorithm is used for non-dominated sorting under multi-dimensional constraints, and a group of non-dominated solutions are selected as the Pareto front by fast non-dominated sorting and congestion distance calculation, generating a Pareto front solution set.

4. The open source micro-kernel based power multi-protocol adaptive conversion system according to claim 1, wherein, The dynamic protocol mapping unit performs cross-protocol time series data alignment based on the DTW algorithm and introduces pruning strategies and fast search algorithms, specifically: Suppose there are two time series of data and , build a The distance matrix D of It represents the cumulative distance between the first i elements in the time series X and the first j elements in Y, and is calculated using the following formula: ； wherein, is the local distance between elements xi and yj; cumulative distance is computed to achieve cross-protocol timing data alignment; The dynamic protocol mapping unit dynamically adjusts the mapping strength of the fields according to the uncertainty and importance of the protocol fields by a protocol entropy weight value analysis method, specifically as follows: For a protocol including n fields to be mapped, from historical communication data, the frequency of each value of the i-th field is counted; let the i-th field have m different possible values, then the estimated probability of the j-th value is ; the number of times the j-th possible value appears is divided by the total number of times all values appear to calculate ; Based on the calculated , the entropy value Ei of the ith field is calculated, and the entropy value Ei is calculated using the following formula: ; where k is a normalization factor, and m is the number of different values that can occur for the protocol field; The entropy weight value wi of the ith field is calculated using the following formula: ； The obtained entropy weight value is used for subsequent field similarity matching calculation, and the field with high entropy weight value has higher priority in the mapping decision.

5. The open source micro-kernel based power multi-protocol adaptive conversion system according to claim 1, wherein, The secure communication interface realizes data encapsulation and encrypted transmission after protocol conversion based on the open source Hongmeng microkernel architecture, supports multi-modal communication with the scheduling master station and the intelligent terminal, and uses a symmetric encryption algorithm to encrypt the data to ensure the confidentiality and integrity of the data in the transmission process.

6. The open source micro-kernel based power multi-protocol adaptive conversion system according to claim 1, wherein, The machine learning optimization module integrates a time series prediction model based on an attention mechanism to analyze historical protocol interaction features, dynamically optimize the protocol conversion delay and data integrity threshold, and dynamically adjust the protocol conversion priority according to the network congestion state, specifically as follows: An LSTM-GAN hybrid neural network model is used to improve the generalization recognition ability of non-standard power protocol features through generative adversarial training; the LSTM network is used to process time series data and can capture long-term dependencies in the data; The GAN network consists of a generator and a discriminator, which learn the distribution characteristics of the data through adversarial training; The generator G attempts to generate samples similar to the real data, and the discriminator D attempts to distinguish between the generated samples and the real samples, and both are continuously optimized in the adversarial process; A protocol semantic graph is constructed, and a graph convolution network is used to extract topological correlation features between protocols. The protocol semantic graph represents the protocols and fields as nodes and edges of a graph. The graph is convoluted by a GCN to extract feature representation of the nodes, thereby mining potential correlations between the protocols. The propagation rule of the GCN is: ; wherein, is the augmented adjacency matrix, which is the original adjacency matrix A plus the identity matrix I; is the degree matrix of A; is the degree matrix of A; is the node feature matrix of the 1st layer; is the weight matrix to be learned of the 1st layer; is the activation function; The LSTM-GAN hybrid neural network model predicts the network congestion level and system load in a preset time period in the future, makes decisions based on the prediction results, and dynamically adjusts the protocol conversion priority to avoid high latency; The dynamic adjustment of the protocol conversion priority is specifically as follows: a reinforcement learning strategy based on Q-learning is designed to dynamically adjust the protocol conversion priority according to the network congestion state, the state s represents the network congestion state, the action a represents the adjustment strategy of the protocol conversion priority, and the update formula of the Q value is: wherein, is a learning rate; r is an immediate reward; is a discount factor; is a next state; is a representation of the next state is a candidate action among all possible actions taken by the agent. When the LSTM-GAN hybrid neural network model predicts that the network reliability decreases and the packet loss risk increases, the error correction strength, retransmission threshold, and packet size are adjusted, and the confirmation timeout time is dynamically set according to the predicted latency to optimize the data integrity threshold.

7. A power protocol conversion method based on the system of any one of claims 1-6, characterized by, The method comprises the following steps: S1. Establish a device communication link through open source Hongmeng heterogeneous networking technology to connect power devices of different protocols to the system and collect real-time communication data of the devices; S2. Identify the communication protocol type using a protocol fingerprint feature extraction algorithm and construct a protocol feature vector; S3. Generate a candidate conversion strategy set using a genetic algorithm optimization module and predict the optimal strategy in combination with an online machine learning model; S4. Deploy dynamic conversion rules in the intelligent conversion engine module based on a microservice architecture and ensure cross-protocol transaction consistency through a distributed transaction mechanism; S5. Perform desensitization processing on sensitive power data using differential privacy technology to complete secure transmission after protocol conversion.

8. The power protocol conversion method of claim 7, wherein, Step S3 is specifically as follows: The genetic algorithm optimization module is used to generate a candidate conversion strategy set; A protocol conversion quality evaluation index system is constructed; the quality evaluation indexes include data fidelity, time delay jitter rate and resource occupation rate; the data fidelity measures the similarity between the converted data and the original data, the time delay jitter rate reflects the time delay fluctuation in the data transmission process, and the resource occupation rate represents the computing resources and network resources consumed by the system in the protocol conversion process; An XGBoost regression model is applied to predict the QoS indexes of each strategy according to the feature vectors of the strategies; An optimal conversion strategy is selected from the strategies through a TOPSIS algorithm, specifically as follows: There are m candidate strategies, each strategy has n evaluation indexes, and an initial decision matrix is constructed , and vector normalization processing is performed to obtain a normalized matrix ; The positive ideal solution and the negative ideal solution are determined according to the benefit type or cost type attribute of each evaluation index and negative ideal solution are expressed by the following equations: wherein, is the value of the positive ideal solution on the jth index; is the value of the negative ideal solution on the jth index; is the set of benefit-type indices, is the set of cost-type indices; The relative closeness of each strategy is calculated, and the following formula is used for calculation: wherein, is the relative closeness of the ith strategy; is the distance of the ith strategy to the positive ideal solution, is the distance of the ith strategy to the negative ideal solution, calculated using the following equation: According to relative closeness The strategies are ranked, and the The largest strategy is selected as the optimal conversion strategy.

9. The power protocol conversion method of claim 7, wherein, Step S4 is specifically as follows: A dynamic conversion rule is deployed through an intelligent conversion engine module, and protocol conversion is performed according to the optimal conversion strategy; A two-phase commit protocol is adopted, a timeout processing and retry mechanism are added, and a blockchain smart contract is introduced for transaction state tracking; The blockchain smart contract is used to record the state and execution result of the transaction, ensuring the non-tamperability and traceability of the transaction; When an exception occurs in the protocol conversion process, a rollback compensation operation is triggered in an event-driven manner to restore the system state to the state before the transaction execution.

10. The power protocol conversion method of claim 7, wherein, In step S5, after the protocol conversion, sensitive power data is subjected to differential privacy processing, and then is encrypted and transmitted through a secure communication interface; the differential privacy technology protects the privacy of data by adding noise to the data; the sensitive power data includes user power consumption information and device operating parameters.

Images