A distributed energy heterogeneous protocol parsing method and system

Abstract

The application relates to a distributed energy heterogeneous protocol analysis method and system, which comprises the following steps: an edge terminal actively detects a new device, sends a cross-protocol detection message sequence, captures a response and extracts a communication characteristic fingerprint; the communication characteristic fingerprint is compared with a cloud protocol mother library, a candidate protocol list is generated through a fuzzy matching algorithm, a protocol with the highest probability in the candidate protocol list is selected as a basic syntax assumption, a standardized virtual point table is generated by combining a predefined semantic anchor point reverse mapping register address, the communication characteristic fingerprint and the standardized virtual point table are reported to a cloud server, verification is carried out, the protocol mother library is added, and the protocol is pushed to all edge terminals in the whole network. The application has the effects of plug-and-play, strong compatibility and expansibility, low operation and maintenance cost and high data standardization level.

Description

Technical Field

[0001] This invention relates to the field of smart grid technology, and in particular to a method and system for parsing heterogeneous protocols of distributed energy resources. Background Technology

[0002] With the advancement of the "carbon peaking and carbon neutrality" strategy, distributed energy devices, represented by distributed photovoltaics, energy storage, and electric vehicle charging piles, are being connected to low-voltage distribution networks at an unprecedented rate. However, the large-scale connection of these devices has brought severe challenges to the field deployment and subsequent operation and maintenance of the power grid. The core issue lies in the heterogeneity and non-standardization of communication protocols. Existing distributed energy devices are characterized by "diverse types and specifications," and "the communication protocols of the end devices are not unified with the communication protocols of the grid side." This means that engineers on site need to deal with devices from different manufacturers, following different protocols, or even the same protocol but with a large number of proprietary extensions. This greatly increases the complexity and workload of integration and commissioning, and seriously restricts the improvement of the plug-and-play and group control capabilities of distributed energy.

[0003] Taking the Modbus protocol, currently the most widely used protocol in industrial control and energy fields, as an example, although it is adopted by many equipment manufacturers as a "de facto standard," there is a serious discrepancy in its implementation. This difference manifests itself on several levels: (1) Differences in slave addresses: Although the Modbus protocol specifies the range of slave addresses, the default addresses of different equipment manufacturers are different. For example, the default address of mainstream inverter manufacturers such as Sungrow Power and Ginlong Technologies is 01H, while the default address of Aiswei is 03H, and that of GoodWe is F7H. This makes the traditional static address configuration scheme fail when the device is first connected, and automatic discovery cannot be achieved.

[0004] (2) Differences in function codes: The Modbus protocol allows users to define custom function codes, which facilitates manufacturers in implementing specific functions, but also undermines the uniformity of the protocol. Existing technical solutions typically only support standard public function codes. Once a device uses a custom function code for data interaction, traditional parsing methods become ineffective.

[0005] (3) Differences in register addresses and data formats: This is the core manifestation of protocol heterogeneity. Even data points with the same physical meaning, such as "total power generation" and "instantaneous active power," may have completely different register addresses, data types (such as 16-bit integers or 32-bit floating-point numbers), data lengths, byte order, and even access permissions in the communication protocols of multiple mainstream inverter manufacturers. For example, in the protocols of Sungrow Power and Ginlong Technologies, "total power generation" is located at address 5004-5005, while in other manufacturers' protocols it may be located in a completely different address range.

[0006] To address the aforementioned issues, existing technologies typically employ a static protocol library-based approach. The core of this approach is the pre-collection, organization, and hard-coding or configuration of a large number of known device communication protocol point tables. When a new device connects, its brand and model must be manually specified, and the system then retrieves the corresponding point table from the library for parsing. The fundamental flaw of this approach lies in its "passive matching" working mode, which presents the following technical challenges: Poor compatibility and weak scalability: It cannot handle unknown devices not included in the protocol library, new device models, or protocol variants of existing devices. Every time a new device appears, professional technicians must manually obtain the protocol documents, parse, encode, test, and update the protocol library, which is a cumbersome process with a long response time.

[0007] High operation and maintenance costs: With the massive access of distributed energy devices, the workload of maintaining the protocol library is growing exponentially, requiring a huge investment of human and material resources.

[0008] This model, which is highly dependent on human intervention, is not in line with the trend of intelligent development. It is a key technological bottleneck for achieving large-scale, low-cost, and automated access and management of power grid end-point equipment.

[0009] Therefore, there is an urgent need for a new method that can break free from the dependence on static protocol libraries and achieve proactive identification, dynamic parsing, and self-learning capabilities for unknown heterogeneous protocols. Summary of the Invention

[0010] This invention aims to solve the technical problems in the prior art, such as poor communication protocol compatibility, inability to automatically identify and parse unknown protocols, high dependence on manual labor for device access and system operation and maintenance, high cost and low efficiency when facing massive and heterogeneous distributed energy devices.

[0011] To achieve the above objectives, the present invention provides the following solution: A method for resolving heterogeneous protocols in distributed energy resources, comprising: Edge terminals actively probe new devices by sending cross-protocol probe message sequences, capturing responses, and extracting communication feature fingerprints. The communication feature fingerprint is compared with the cloud protocol database, and a candidate protocol list is generated by fuzzy matching algorithm. The protocol with the highest probability in the candidate protocol list is selected as the basic syntax hypothesis, and combined with the predefined semantic anchor reverse mapping register address to generate a standardized virtual point table. The communication feature fingerprint and the standardized virtual point table are reported to the cloud server for verification and addition to the protocol master library, and then pushed to all edge terminals in the network.

[0012] Optionally, the probe message sequence includes: Communication parameter polling: Automatically polls for different combinations of communication parameters; wherein, the communication parameter combinations include: baud rate, data bits, stop bits, and parity bits; Address and function code detection: Under successfully negotiated communication parameters, a detection message is sent; wherein, the detection message includes: detection messages for common protocol broadcast addresses and polling unicast addresses.

[0013] Optionally, the communication feature fingerprint is a multi-dimensional feature vector, including: frame header format, position and length of slave address field, position and length of function code field, description method of data length field, structural features of data field, verification algorithm type and byte order.

[0014] Optionally, the cloud protocol master library not only stores the complete protocol specifications of known devices, but also stores standard communication feature fingerprint vectors abstracted from massive protocols.

[0015] Optionally, generating a candidate protocol list using a fuzzy matching algorithm includes: A fuzzy matching algorithm is used to calculate the similarity score between the new device's communication feature fingerprint and all standard communication feature fingerprint vectors in the parent database, thereby obtaining a list of candidate protocols with probabilities.

[0016] Optionally, the predefined semantic anchor points refer to data points that have universal physical meaning and predictable data change patterns in devices of a preset type.

[0017] Optionally, by combining predefined semantic anchor reverse mapping register addresses, a standardized virtual point table is generated, including: Based on the aforementioned basic syntax assumptions, a read command is sent to the device, prioritizing the detection of high-probability public address ranges based on industry experience and the generalization of known protocols. The system continuously reads the preset high-probability public address range within a preset time period and analyzes the numerical patterns of the returned data. If the value returned by a certain address shows a monotonically increasing characteristic, it is bound to a predefined semantic anchor with high confidence. If the value returned by another address fluctuates within the rated power range and its trend is highly correlated with external reference data, it is also bound to a predefined semantic anchor. The standardized virtual point table is automatically generated by successfully back-mapping multiple predefined semantic anchors.

[0018] Optionally, the standardized virtual point table includes: the semantic name of the data point, the parsed register address, the data type, and the unit information.

[0019] Optionally, reporting the communication feature fingerprint and the standardized virtual point table to the cloud server includes: The communication feature fingerprint and the standardized virtual point table are reported as a complete new protocol sample to the cloud server; The cloud server aggregates new protocol samples reported from different edge terminals across the network, and uses machine learning algorithms to verify, deduplicate, and analyze the correlation of the new protocol samples. If multiple terminals report similar samples for the same unknown device model, the cloud will merge them and generalize them into a new standard protocol model, add it to the protocol master library, and update the feature vector of the master library at the same time. The validated and optimized protocol master library is pushed to all edge terminals across the network.

[0020] A distributed energy heterogeneous protocol parsing system, the system comprising: An edge terminal is used to deploy at the site where distributed energy equipment is accessed. The edge terminal includes: an active detection module, a communication feature fingerprint extraction module, a reverse mapping module, and a device model generation module. The active detection module is used to actively detect new devices and send cross-protocol detection message sequences. The communication feature fingerprint extraction module is used to capture the response and extract the communication feature fingerprint; The reverse mapping module is used to compare the communication feature fingerprint with the cloud protocol master library, generate a candidate protocol list through a fuzzy matching algorithm, select the protocol with the highest probability in the candidate protocol list as the basic syntax hypothesis, and generate a standardized virtual point table by combining the predefined semantic anchor reverse mapping register address. The device model generation module is used to report the communication feature fingerprint and the standardized virtual point table to the cloud server, verify them, add them to the protocol master library, and then push them to all edge terminals in the network. Cloud server: Used to configure the protocol master library, fuzzy matching engine and self-learning engine.

[0021] The beneficial effects of this invention are as follows: 1. Achieve plug-and-play functionality: Through active detection and self-learning, the system can automatically identify and parse unknown protocols, eliminating the reliance on static protocol libraries and manual configuration, and truly realizing "plug-and-play" functionality for distributed energy devices.

[0022] 2. Strong compatibility and scalability: Based on probability matching and semantic mapping, it has good fault tolerance for non-standard protocols and protocol variants, and can continuously learn new protocols through cloud-edge collaboration mechanisms, possessing unlimited scalability.

[0023] 3. Reduced operation and maintenance costs: Automated protocol parsing and a swarm intelligence self-learning system greatly reduce the manual intervention required for on-site debugging and subsequent protocol library maintenance, significantly reducing the operation and maintenance costs of distributed energy throughout its entire life cycle.

[0024] 4. Improve data standardization: By generating standardized "virtual point tables" and JSON device models, a unified and reliable data foundation is provided for upper-layer applications (such as power quality management and optimized scheduling). Attached Figure Description

[0025] Figure 1 This is a schematic diagram of the overall system architecture according to an embodiment of the present invention; Figure 2 This is a schematic diagram of a distributed energy heterogeneous protocol parsing method based on active detection and cloud-edge collaborative self-learning according to an embodiment of the present invention. Figure 3 This is a logical schematic diagram of the semantic anchor register reverse mapping in an embodiment of the present invention; Figure 4 This is a schematic diagram of a standardized JSON format device model file according to an embodiment of the present invention. Detailed Implementation

[0026] 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.

[0027] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0028] like Figure 2As shown, this embodiment proposes a method for parsing heterogeneous distributed energy protocols based on active detection and cloud-edge collaborative self-learning, including: Step 1: The edge terminal actively probes new devices by sending a cross-protocol probe message sequence, capturing responses, and extracting communication feature fingerprints; Step 2: Compare the communication feature fingerprint with the cloud protocol database, and generate a candidate protocol list using a fuzzy matching algorithm; Step 3: Select the protocol with the highest probability from the candidate protocol list as the basic syntax hypothesis, and combine it with the predefined semantic anchor reverse mapping register address to generate a standardized virtual point table; Step 4: Report the communication feature fingerprint and the standardized virtual point table to the cloud server, verify them, add them to the protocol master library, and then push them to all edge terminals in the network.

[0029] Specifically, in this embodiment, step 1, active detection and communication feature fingerprint extraction, includes: When an edge terminal (such as a secure access unit) detects a new, unknown distributed energy device connecting via a physical interface (such as RS485), it initiates an active probing process. This process does not simply attempt a single protocol; instead, it sends a carefully designed, cross-protocol "probe message sequence" to the device. This sequence must contain at least: Communication parameter polling: Automatically polls different combinations of communication parameters, such as baud rate (9600, 19200, etc.), data bits, stop bits, and parity bits (N, 8, 1; E, 8, 1, etc.).

[0030] Address and Function Code Probe: Under successfully negotiated communication parameters, a probe message containing a common protocol broadcast address (such as 0x00 for Modbus) and a polling unicast address (such as 1-247) is sent. The function code portion of the message combines commonly used read commands for various mainstream protocols (such as Modbus function codes 0x03 and 0x04).

[0031] Edge terminals capture and analyze device response messages, focusing on their message structure rather than specific data content, to extract a unique "communication feature fingerprint" that identifies the protocol type. This fingerprint is a multi-dimensional feature vector that contains at least the following metadata: frame header format, location and length of the slave address field, location and length of the function code field, description method of the data length field (fixed length or variable length), structural features of the data field, check algorithm type (such as CRC16-Modbus, XOR check, etc.), and byte order (big-endian or little-endian).

[0032] Specifically, in this embodiment, step 2, protocol identification based on fuzzy matching and probabilistic assumptions, includes: Edge terminals report locally extracted "communication feature fingerprints" to the cloud server, or perform local matching using a lightweight protocol master library distributed from the cloud. The cloud protocol master library not only stores the complete protocol specifications of known devices, but more importantly, it stores standard "communication feature fingerprint vectors" abstracted from massive amounts of protocols. The system uses fuzzy matching algorithms (such as cosine similarity algorithms) to calculate the similarity score between the new device fingerprint and all standard fingerprints in the master library.

[0033] The matching result is not an absolute "yes" or "no," but rather a list of candidate protocols with probabilities. For example, the system might output: {"Modbus-RTU variant A": 0.85, "DLT645-2007": 0.10, "a certain manufacturer's proprietary protocol X": 0.05}. This probabilistic identification method greatly enhances the fault tolerance for identifying non-standard protocols, proprietary protocols, or existing protocol variants.

[0034] Specifically, in this embodiment, the logic of step 3, which involves the reverse mapping of register addresses based on data semantics, is as follows: Figure 3 As shown, it includes: This step is the key inventive step of the present invention, which aims to automatically parse the register address of the core data point without the need for a manufacturer's point table.

[0035] 1. Definition of "Semantic Anchor Point": This invention creatively proposes the concept of "semantic anchor point." A semantic anchor point refers to a data point in a specific type of equipment (such as a photovoltaic inverter) that possesses universal physical meaning and predictable data change patterns. For example, the "total power generation" (the value should monotonically increase over time), "instantaneous active power" (the value should fluctuate between 0 and the equipment's rated power, and be positively correlated with light intensity), and "internal temperature" (the value should change slowly within a reasonable range) of a photovoltaic inverter can all serve as semantic anchor points.

[0036] 2. Hypothesis-Verification-Locking Process: The system selects the protocol with the highest probability from the candidate list (e.g., Modbus-RTU variant A with an 85% probability) as the basic syntax hypothesis. Based on this syntax, the system sends a read command to the device, but its target address is not blindly tried. Instead, it prioritizes probing those "high-probability public address ranges" that, based on industry experience and induction of known protocols, are most likely to contain semantic anchor data, such as the 5000-5100 address range.

[0037] 3. Reverse verification through data patterns: The system continuously reads these high-probability addresses over a period of time (e.g., 15 minutes) and analyzes the numerical patterns of the returned data. If the value returned by a certain address (or combination of addresses) shows a monotonically increasing characteristic, the system binds it with high confidence to the semantic anchor point of "total power generation." If the value returned by another address fluctuates within the rated power range, and its trend is highly correlated with external reference data (such as the light intensity obtained from a weather API), it is bound to "instantaneous active power."

[0038] 4. Generate a “virtual point table”: Through successful reverse mapping of multiple key semantic anchor points, the system automatically constructs a core, standardized “virtual point table”, which contains key information such as the semantic name of the data point, the parsed register address, data type, and unit.

[0039] Specifically, in this embodiment, step 4, the protocol library self-learning and dynamic update for cloud-edge collaboration, includes: This embodiment constructs a closed-loop learning system capable of self-evolution and capability diffusion, which is underpinned by a powerful network effect, meaning that the learning outcomes of any node can rapidly empower the entire network.

[0040] 1. Edge Model Reporting: After successfully generating a "virtual point table", the edge terminal will report the device's "communication feature fingerprint" and the generated standardized device model (encapsulated in standardized formats such as JSON) as a complete "new protocol sample" to the cloud server.

[0041] 2. Cloud Aggregation and Generalization: The cloud backend aggregates "new protocol samples" reported from different edge terminals across the network. Machine learning algorithms (such as cluster analysis) are used to verify, deduplicate, and perform correlation analysis on these samples. If multiple terminals report similar samples for the same unknown device model, the cloud will merge them and generalize them into a new standard protocol model, which will be officially added to the protocol master library, while updating the feature vector of the master library.

[0042] 3. Dynamic updates and push of protocol library: The verified and optimized protocol master library (or its incremental update package) is dynamically and securely pushed to all edge terminals across the network.

[0043] 4. Forming a closed-loop evolutionary system: Through the process of "discovering new protocols at the edge → learning and generalizing in the cloud → upgrading capabilities across the entire network," the system achieves a leap from single-point breakthroughs to network-wide intelligence. The next time any edge terminal encounters a device of the same model, it can directly perform fast and deterministic parsing through the updated protocol library, eliminating the need for time-consuming active probing and reverse mapping, greatly improving efficiency and reducing long-term maintenance costs.

[0044] Corresponding to the method in this embodiment, this embodiment also provides a distributed energy heterogeneous protocol parsing system based on active detection and cloud-edge collaborative self-learning. The overall architecture of the system is as follows: Figure 1 As shown. The system includes: Edge terminal: Deployed at the site of distributed energy equipment access, including active detection module, communication feature fingerprint extraction module and reverse mapping module.

[0045] Cloud server: includes protocol master library, fuzzy matching engine and self-learning engine.

[0046] The edge terminal and the cloud server are connected through a communication network to collaboratively execute the above methods.

[0047] Specifically, in this embodiment, the system includes: An edge terminal is used to deploy at the site where distributed energy equipment is accessed. The edge terminal includes: an active detection module, a communication feature fingerprint extraction module, a reverse mapping module, and a device model generation module. The active detection module is used to actively detect new devices and send cross-protocol detection message sequences. The communication feature fingerprint extraction module is used to capture the response and extract the communication feature fingerprint; The reverse mapping module is used to compare the communication feature fingerprint with the cloud protocol master library, generate a candidate protocol list through a fuzzy matching algorithm, select the protocol with the highest probability in the candidate protocol list as the basic syntax hypothesis, and generate a standardized virtual point table by combining the predefined semantic anchor reverse mapping register address. The device model generation module is used to report the communication feature fingerprint and the standardized virtual point table to the cloud server, verify them, add them to the protocol master library, and then push them to all edge terminals in the network. Cloud server: Used to configure the protocol master library, fuzzy matching engine and self-learning engine.

[0048] The following embodiment uses a typical rural power distribution area as an application scenario to illustrate how the present invention can automatically parse the protocol of a newly connected photovoltaic inverter of an unknown brand.

[0049] Scene setting: Deployment environment: Deploy an edge intelligent terminal (secure access unit) integrating the method of this invention in a certain area.

[0050] Device to be connected: A 50kW three-phase photovoltaic inverter of unknown brand and model, physically connected to the edge terminal via an RS485 interface.

[0051] Step 1: Active Detection Once the edge terminal detects a new physical connection on the RS485 interface, it initiates an active detection process.

[0052] Communication parameter negotiation: The terminal attempts common communication parameter combinations such as (9600, N, 8, 1) and (19200, N, 8, 1) in sequence. Assume that when the parameter is "9600, N, 8, 1", the probe frame sent by the terminal receives a response from the device, thus locking in that communication parameter.

[0053] Address Probe: The terminal first sends a Modbus broadcast message (slave address 0x00), function code 0x03, requesting to read register address 0x0000, with a length of 1. If there is no response, it begins polling the unicast address, starting from 0x01, and sends the same read request sequentially.

[0054] Step 2: Feature Extraction and Protocol Assumptions Response capture: When polling the slave address 0x05, the edge terminal received a response message that conformed to the Modbus-RTU frame structure and had a correct CRC16 checksum. However, no correct responses were received at addresses 0x01 to 0x04.

[0055] Fingerprint generation: The terminal analyzes the message structure of the successful response and extracts the communication feature fingerprint.

[0056] The protocol assumes that the terminal will report the fingerprint to the cloud. After performing fuzzy matching in the cloud's protocol master database, it returns a high-probability hypothesis: "90% probability of Modbus-RTU protocol, slave address is non-default 0x05". Based on this highest probability hypothesis, the system proceeds to the next step of the reverse mapping process.

[0057] Step 3: Reverse mapping: The system begins probing semantic anchors based on the syntax rules of Modbus-RTU.

[0058] 1. Detecting "Total Power Generation": The system prioritizes detecting the common address range previously used by Sungrow Power and Ginlong Technologies, such as 5000-5100. The system sends commands to read addresses 5004 and 5005 (two registers in total). The inverter returns two 16-bit words. The terminal combines them into a 32-bit unsigned integer in big-endian mode. Over the next 15 minutes, the terminal repeats this reading every minute, recording the numerical sequence: 34567, 34571, 34575,..., 34623. Analyzing this sequence reveals a strictly monotonically increasing trend with reasonable increments. Based on this, the system maps the address to the semantic anchor "Total Power Generation" with 98% confidence and infers its data type as U32, with a unit of 0.1 kWh (inferred from the magnitude of numerical changes and device power).

[0059] 2. Probing "Instantaneous Active Power": The system continues probing within the address range of 5000-5100. When reading address 5013 (one register), it returns a 16-bit integer. Within the same 15-minute observation period, this value fluctuates within this range, and its fluctuation trend shows a high positive correlation with the local total solar radiation (GHI) data obtained by the edge terminal from the public weather API. Based on this, the system maps address 5013 to the semantic anchor "Instantaneous Active Power" with a 95% confidence level, using the data type U16 and the unit W.

[0060] Step 4: Model Generation and Self-Learning 1. Model Generation: Based on the mapping results above, the edge terminal generates a standardized JSON format device model file locally, the specific content of which is as follows: Figure 4 As shown.

[0061] 2. Self-learning process: The edge terminal uploads this JSON file, along with its communication feature fingerprint, as a complete "new protocol sample" to the cloud. Upon receiving this sample, the cloud's self-learning engine uses clustering and feature comparison algorithms to find that its features are highly similar to a known "Ginlong Technologies model variant protocol" in the protocol master library. The cloud system determines this is a new instance and performs a merging operation, optimizing the feature vector of the protocol model (e.g., confirming that slave address 0x05 is also a possible factory configuration) and enhancing the confidence of related semantic anchor addresses.

[0062] 3. Capability Diffusion: Ultimately, the protocol master library update package containing this new knowledge is pushed to other edge terminals across the network. Subsequently, when any other terminal in the network connects to an inverter of the same model and configuration, it will be able to directly match the "Ginlong Technologies variant protocol" via protocol fingerprint and directly perform data parsing without needing to perform detection and mapping again. This fully demonstrates the entire closed-loop working process of this invention, from the unknown to the known, and then to swarm intelligence.

[0063] The embodiments described above are merely preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Various modifications and improvements made to the technical solutions of the present invention by those skilled in the art without departing from the spirit of the present invention should fall within the protection scope defined by the claims of the present invention.

Claims

1. A method for parsing heterogeneous protocols of distributed energy resources, characterized in that, include: Edge terminals actively probe new devices by sending cross-protocol probe message sequences, capturing responses, and extracting communication feature fingerprints. The communication feature fingerprint is compared with the cloud protocol database, and a candidate protocol list is generated by fuzzy matching algorithm. The protocol with the highest probability in the candidate protocol list is selected as the basic syntax hypothesis, and combined with the predefined semantic anchor reverse mapping register address to generate a standardized virtual point table. The communication feature fingerprint and the standardized virtual point table are reported to the cloud server for verification and addition to the protocol master library, and then pushed to all edge terminals in the network.

2. The distributed energy heterogeneous protocol parsing method according to claim 1, characterized in that, The probe message sequence includes: Communication parameter polling: Automatically polls for different combinations of communication parameters; wherein, the communication parameter combinations include: baud rate, data bits, stop bits, and parity bits; Address and function code detection: Under successfully negotiated communication parameters, a detection message is sent; wherein, the detection message includes: detection messages for common protocol broadcast addresses and polling unicast addresses.

3. The distributed energy heterogeneous protocol parsing method according to claim 1, characterized in that, The communication feature fingerprint is a multi-dimensional feature vector, including: frame header format, position and length of slave address field, position and length of function code field, description method of data length field, structural features of data field, verification algorithm type, and byte order.

4. The distributed energy heterogeneous protocol parsing method according to claim 1, characterized in that, The cloud-based protocol library not only stores the complete protocol specifications of known devices, but also stores standard communication feature fingerprint vectors abstracted from massive protocols.

5. The distributed energy heterogeneous protocol parsing method according to claim 1, characterized in that, The candidate protocol list generated by the fuzzy matching algorithm includes: A fuzzy matching algorithm is used to calculate the similarity score between the new device's communication feature fingerprint and all standard communication feature fingerprint vectors in the parent database, thereby obtaining a list of candidate protocols with probabilities.

6. The distributed energy heterogeneous protocol parsing method according to claim 1, characterized in that, The predefined semantic anchor points refer to data points that have universal physical meaning and predictable data change patterns in devices of a preset type.

7. The distributed energy heterogeneous protocol parsing method according to claim 1, characterized in that, By combining predefined semantic anchor reverse mapping register addresses, a standardized virtual point table is generated, including: Based on the aforementioned basic syntax assumptions, a read command is sent to the device, prioritizing the detection of high-probability public address ranges based on industry experience and the generalization of known protocols. The system continuously reads the preset high-probability public address range within a preset time period and analyzes the numerical patterns of the returned data. If the value returned by a certain address shows a monotonically increasing characteristic, it is bound to a predefined semantic anchor with high confidence. If the value returned by another address fluctuates within the rated power range and its trend is highly correlated with external reference data, it is also bound to a predefined semantic anchor. The standardized virtual point table is automatically generated by successfully back-mapping multiple predefined semantic anchors.

8. The distributed energy heterogeneous protocol parsing method according to claim 1, characterized in that, The standardized virtual point table includes: the semantic name of the data point, the parsed register address, the data type, and the unit information.

9. The distributed energy heterogeneous protocol parsing method according to claim 1, characterized in that, Reporting the communication feature fingerprint and the standardized virtual point table to the cloud server includes: The communication feature fingerprint and the standardized virtual point table are reported as a complete new protocol sample to the cloud server; The cloud server aggregates new protocol samples reported from different edge terminals across the network, and uses machine learning algorithms to verify, deduplicate, and analyze the correlation of the new protocol samples. If multiple terminals report similar samples for the same unknown device model, the cloud will merge them and generalize them into a new standard protocol model, add it to the protocol master library, and update the feature vector of the master library at the same time. The validated and optimized protocol master library is pushed to all edge terminals across the network.

10. A distributed energy heterogeneous protocol parsing system, characterized in that, For implementing the method as described in any one of claims 1-9, the system comprises: An edge terminal is used to deploy at the site where distributed energy equipment is accessed. The edge terminal includes: an active detection module, a communication feature fingerprint extraction module, a reverse mapping module, and a device model generation module. The active detection module is used to actively detect new devices and send cross-protocol detection message sequences. The communication feature fingerprint extraction module is used to capture the response and extract the communication feature fingerprint; The reverse mapping module is used to compare the communication feature fingerprint with the cloud protocol master library, generate a candidate protocol list through a fuzzy matching algorithm, select the protocol with the highest probability in the candidate protocol list as the basic syntax hypothesis, and generate a standardized virtual point table by combining the predefined semantic anchor reverse mapping register address. The device model generation module is used to report the communication feature fingerprint and the standardized virtual point table to the cloud server, verify them, add them to the protocol master library, and then push them to all edge terminals in the network. Cloud server: Used to configure the protocol master library, fuzzy matching engine and self-learning engine.

Images