Energy storage power station industrial control instruction identification method and device and computer equipment

Abstract

The invention relates to an energy storage power station industrial control instruction identification method and device, computer equipment, a computer readable storage medium and a computer program product. The method comprises the following steps: acquiring network communication data of an energy storage power station in real time; under the condition that the network communication data meets a preset condition, updating the current industrial control instruction identification model according to the network communication data to obtain a new industrial control instruction identification model; obtaining current network communication data of the energy storage power station, and analyzing the current network communication data according to a preset industrial control protocol to obtain an industrial control instruction key field in the current network communication data; and inputting the industrial control instruction key field into the new industrial control instruction identification model as feature information to obtain an industrial control instruction identification result corresponding to the current network communication data. By adopting the method, the identification accuracy of the industrial control instruction can be improved.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a method, device, computer equipment, computer-readable storage medium and computer program product for identifying industrial control instructions of an energy storage power station. Background Art

[0002] During the operation of energy storage power stations, a large number of industrial control instructions are frequently transmitted in network communications. These instructions are used for operations such as equipment control, status monitoring and data collection.

[0003] However, with the increase in network scale and communication frequency, the accuracy and real-time performance of command recognition face great challenges. Traditional industrial control systems rely on static rules or fixed protocol parsing methods to identify commands, while the network traffic of energy storage power stations contains a large number of dynamic factors, such as changes in device access and command type updates. It is difficult to cope with the frequently changing communication environment and diverse command types, resulting in poor recognition accuracy for industrial control commands. Summary of the invention

[0004] Based on this, it is necessary to provide a method, device, computer equipment, computer-readable storage medium and computer program product for industrial control instruction recognition of energy storage power stations that can improve the accuracy of industrial control instruction recognition in order to address the above technical problems.

[0005] In a first aspect, the present application provides a method for identifying industrial control instructions of an energy storage power station, comprising:

[0006] Obtain network communication data of energy storage power stations in real time;

[0007] When the network communication data meets the preset conditions, the current industrial control instruction recognition model is updated according to the network communication data to obtain a new industrial control instruction recognition model;

[0008] Acquire current network communication data of the energy storage power station, parse and process the current network communication data according to a preset industrial control protocol, and obtain key fields of industrial control instructions in the current network communication data;

[0009] The key field of the industrial control instruction is used as feature information and input into the new industrial control instruction recognition model to obtain the industrial control instruction recognition result corresponding to the current network communication data.

[0010] In one embodiment, when the network communication data meets a preset condition, updating the current industrial control instruction recognition model according to the network communication data to obtain a new industrial control instruction recognition model includes:

[0011] Determine the frequency of sending industrial control instructions and the number of new devices connected within a preset time window according to the network communication data;

[0012] When the industrial control instruction sending frequency and / or the number of new devices connected exceeds the corresponding preset threshold, the model weight and input feature weight of the current industrial control instruction recognition model are updated according to the network communication data within the preset time window to obtain a new industrial control instruction recognition model.

[0013] In one embodiment, when the network communication data meets a preset condition, updating the current industrial control instruction recognition model according to the network communication data to obtain a new industrial control instruction recognition model includes:

[0014] Determining the amount of network communication data and the number of types of industrial control instructions sent within a preset time window according to the network communication data;

[0015] When the network communication data volume is less than the corresponding preset threshold, the model structure of the current industrial control instruction recognition model is trimmed according to the network communication data volume to obtain a new industrial control instruction recognition model;

[0016] and / or

[0017] When the number of the industrial control instruction sending types is less than the corresponding preset threshold, the input features of the current industrial control instruction recognition model are trimmed according to the number of the industrial control instruction sending types to obtain a new industrial control instruction recognition model.

[0018] In one embodiment, the key field of the industrial control instruction is input as feature information into the new industrial control instruction recognition model to obtain the industrial control instruction recognition result corresponding to the current network communication data, including:

[0019] Preprocessing the industrial control instruction key field to obtain the preprocessed industrial control instruction key field;

[0020] Performing standardization processing on the preprocessed industrial control instruction key fields to obtain standardized industrial control instruction key fields;

[0021] Converting the standardized key fields of the industrial control instructions into feature vectors;

[0022] The feature vector is input into the new industrial control instruction recognition model to obtain the industrial control instruction recognition result corresponding to the current network communication data.

[0023] In one of the embodiments, the new industrial control instruction recognition model includes a spatial feature extraction module, a short-time sequence feature extraction module, a long-time sequence feature extraction module and a recognition module;

[0024] The step of inputting the feature vector into the new industrial control instruction recognition model to obtain the industrial control instruction recognition result corresponding to the current network communication data includes:

[0025] Using the spatial feature extraction module to perform spatial feature extraction processing on the feature vector to obtain the spatial feature corresponding to the feature vector;

[0026] The short-time time series feature extraction module is used to perform short-time time series feature extraction processing on the spatial feature to obtain the short-time time series feature corresponding to the feature vector;

[0027] Using the long-term time series feature extraction module to perform long-term time series feature extraction processing on the short-term time series feature to obtain the long-term time series feature corresponding to the feature vector;

[0028] The long-term timing characteristics are input into the recognition module to obtain the industrial control instruction recognition result corresponding to the current network communication data.

[0029] In one embodiment, after obtaining the industrial control instruction recognition result corresponding to the current network communication data, the method further includes:

[0030] Monitoring the status information and execution result of a target industrial control instruction; the target industrial control instruction is an industrial control instruction corresponding to the industrial control instruction identification result;

[0031] Analyze the status information and the execution result to obtain an execution report of the target industrial control instruction, and upload the execution report to the blockchain network.

[0032] In a second aspect, the present application also provides an energy storage power station industrial control instruction recognition device, comprising:

[0033] A data acquisition module is used to acquire network communication data of the energy storage power station in real time;

[0034] A model updating module, used to update the current industrial control instruction recognition model according to the network communication data to obtain a new industrial control instruction recognition model when the network communication data meets the preset conditions;

[0035] A field parsing module, used to obtain the current network communication data of the energy storage power station, parse and process the current network communication data according to a preset industrial control protocol, and obtain the key fields of the industrial control instructions in the current network communication data;

[0036] The instruction recognition module is used to input the key field of the industrial control instruction as feature information into the new industrial control instruction recognition model to obtain the industrial control instruction recognition result corresponding to the current network communication data.

[0037] In a third aspect, the present application further provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0038] Obtain network communication data of energy storage power stations in real time;

[0039] When the network communication data meets the preset conditions, the current industrial control instruction recognition model is updated according to the network communication data to obtain a new industrial control instruction recognition model;

[0040] Acquire current network communication data of the energy storage power station, parse and process the current network communication data according to a preset industrial control protocol, and obtain key fields of industrial control instructions in the current network communication data;

[0041] The key field of the industrial control instruction is used as feature information and input into the new industrial control instruction recognition model to obtain the industrial control instruction recognition result corresponding to the current network communication data.

[0042] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the following steps are implemented:

[0043] Obtain network communication data of energy storage power stations in real time;

[0044] When the network communication data meets the preset conditions, the current industrial control instruction recognition model is updated according to the network communication data to obtain a new industrial control instruction recognition model;

[0045] Acquire current network communication data of the energy storage power station, parse and process the current network communication data according to a preset industrial control protocol, and obtain key fields of industrial control instructions in the current network communication data;

[0046] The key field of the industrial control instruction is used as feature information and input into the new industrial control instruction recognition model to obtain the industrial control instruction recognition result corresponding to the current network communication data.

[0047] In a fifth aspect, the present application further provides a computer program product, including a computer program, which implements the following steps when executed by a processor:

[0048] Obtain network communication data of energy storage power stations in real time;

[0049] When the network communication data meets the preset conditions, the current industrial control instruction recognition model is updated according to the network communication data to obtain a new industrial control instruction recognition model;

[0050] Acquire current network communication data of the energy storage power station, parse and process the current network communication data according to a preset industrial control protocol, and obtain key fields of industrial control instructions in the current network communication data;

[0051] The key field of the industrial control instruction is used as feature information and input into the new industrial control instruction recognition model to obtain the industrial control instruction recognition result corresponding to the current network communication data.

[0052] The above-mentioned energy storage power station industrial control instruction recognition method, device, computer equipment, computer-readable storage medium and computer program product first obtain the network communication data of the energy storage power station in real time. By obtaining the network communication data of the energy storage power station in real time, it is possible to dynamically perceive the latest instructions and status changes of each device in the energy storage power station, thereby ensuring the real-time nature of industrial control instruction recognition; then, when the network communication data meets the preset conditions, the current industrial control instruction recognition model is updated according to the network communication data to obtain a new industrial control instruction recognition model. By monitoring the network communication data and judging the conditions, the recognition model is automatically updated under specific conditions, so that the model can adjust the weight or structure according to the new communication data to adapt to the changing factors in the scene, thereby effectively improving the recognition accuracy and adaptability of the model in different scenes, and avoiding the recognition errors of the traditional static model caused by data changes; then, the current network communication of the energy storage power station is obtained. Data, according to the preset industrial control protocol, the current network communication data is parsed and processed to obtain the key fields of the industrial control instructions in the current network communication data. Through the preset parsing of the industrial control protocol, the key fields of the industrial control instructions can be extracted from the original data, such as the instruction type, device identification and operation parameters, etc., and the core information of the industrial control instructions can be accurately parsed, so that the model can recognize and understand the specific meaning of each instruction, improve the accuracy of data parsing, and ensure the accuracy of recognition; finally, the key fields of the industrial control instructions are used as feature information and input into the new industrial control instruction recognition model to obtain the industrial control instruction recognition result corresponding to the current network communication data. The extracted key fields of the industrial control instructions are input into the model as feature information, which can enhance the model's understanding and recognition ability of the instructions, help the model quickly identify the instruction type and the corresponding operation, and improve the recognition efficiency and accuracy of the model, so that it can more accurately reflect the real-time instruction situation. In the above method, by acquiring the network communication data of the energy storage power station in real time, the recognition model is dynamically updated under specific conditions to adapt to changes in the communication environment and improve the accuracy of recognition. By extracting key fields of instructions through protocol analysis, feature information can be effectively extracted, and instruction types and parameters can be quickly identified, thus achieving real-time monitoring and efficient identification of industrial control instructions. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the drawings required for use in the embodiments of the present application or related technical descriptions will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0054] Figure 1 A schematic diagram of a flow chart of a method for identifying industrial control instructions of an energy storage power station in one embodiment;

[0055] Figure 2 A flowchart of steps for obtaining an industrial control instruction recognition result in one embodiment;

[0056] Figure 3 A schematic diagram of a flow chart of a method for identifying industrial control instructions of an energy storage power station in another embodiment;

[0057] Figure 4 It is a structural block diagram of an energy storage power station industrial control instruction recognition device in one embodiment;

[0058] Figure 5 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0059] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0060] In one embodiment, Figure 1 As shown, a method for identifying industrial control instructions for energy storage power stations is provided. This embodiment uses the method applied to a terminal as an example for illustration. It is understandable that the method can also be applied to a server, and can also be applied to a system including a terminal and a server, and is implemented through the interaction between the terminal and the server. Among them, the terminal can be, but is not limited to, various personal computers, laptops, smart phones, and tablet computers. The server can be an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides cloud computing services. In this embodiment, the method includes the following steps:

[0061] Step S101, acquiring network communication data of the energy storage power station in real time.

[0062] Among them, network communication data refers to the industrial control instructions, status information and sensor data transmitted through the network between the various devices in the energy storage power station and between the devices and the control center. These data reflect the real-time control and monitoring information in the operation of the energy storage power station, including key information such as equipment status, operating instructions, feedback information, etc., and have strong timeliness and dynamism.

[0063] Exemplarily, the terminal can obtain the network communication data of the energy storage power station in real time in a variety of ways. For example, by connecting to the network switching device, the terminal can continuously monitor the network communication traffic between the various devices in the energy storage power station, and ensure the integrity and real-time nature of the communication content by capturing the full amount of data. In a specific implementation, the terminal can capture all communication data packets in real time by connecting to the mirror port of the energy storage power station network switch. The mirror port copies the communication data of the entire energy storage power station to the terminal. The terminal uses high-performance network cards and big data technology to ensure that all data packet contents are fully recorded and stored in a high-speed environment. The terminal will also extract initial information (such as IP address, port number, data packet size, etc.) through basic protocol parsing to provide support for subsequent in-depth analysis and instruction recognition.

[0064] Step S102, when the network communication data meets the preset conditions, the current industrial control instruction recognition model is updated according to the network communication data to obtain a new industrial control instruction recognition model.

[0065] The preset conditions refer to specific trigger conditions used to determine changes in the network communication environment or business scenarios, such as instruction sending frequency, data traffic, number of device access, etc. When these factors in the network communication data exceed the preset threshold, it is determined that the current recognition model may not be able to adapt to the new business scenario, thereby triggering a model update.

[0066] Exemplarily, the terminal can determine whether the preset conditions are met by monitoring the characteristics of network communication data in real time. For example, the terminal can monitor the frequency of instruction transmission and the changing trend of data traffic. When these characteristics are detected to have changed significantly, the terminal can dynamically update the recognition model through incremental learning or online learning modules so that the model can adapt to the new data distribution and instruction pattern. In a specific implementation, the terminal can adjust the weight of the model, the weight of the input feature or the network structure according to the current scenario, thereby improving the recognition accuracy and adaptability of the model in the new scenario, and finally obtaining a new industrial control instruction recognition model suitable for the current business environment.

[0067] Step S103, obtaining the current network communication data of the energy storage power station, parsing and processing the current network communication data according to the preset industrial control protocol, and obtaining the key fields of the industrial control instructions in the current network communication data.

[0068] Among them, the industrial control protocol refers to a specific industrial control protocol used for communication between devices in energy storage power stations. The protocol defines the instruction format, field type and its structure so that key information in the instruction can be identified and extracted during the parsing process. Common industrial control protocols include Modbus, IEC 104, etc. These protocols cover key information such as instruction type, data content, timestamp, etc.

[0069] Exemplarily, the terminal can obtain key fields based on the deep parsing of industrial protocols. In the specific implementation, the terminal uses the parsing engine to parse the network messages in the energy storage power station layer by layer, from the IP layer, the transport layer (such as TCP / UDP) to the application layer protocol to ensure the integrity and validity of the data. The terminal first parses the IP layer and transport layer protocols of the communication data to ensure the correct transmission of the data, and then performs a deep analysis of the industrial control protocols (such as IEC 61850, IEC 104) at the application layer, focusing on extracting specific logical nodes, data objects, and remote control types in the protocol. For example, for the IEC 104 protocol, the terminal can identify the structure of telemetry and remote control data to ensure that the basic information of each instruction is accurately disassembled.

[0070] During the parsing process, the terminal gradually extracts key fields of industrial control behavior instructions, such as "invokeID", "confirmedServiceRequest", "itemId", "Data", "confirmedServiceResponse" and other fields. These fields accurately describe the type of instruction, target device and operating parameters, ensuring the correct transmission and execution of instructions in complex network environments. Based on long-term instruction data monitoring and parsing results, the terminal also builds an instruction feature library. This feature library covers common instruction patterns and types in energy storage power stations, serving as an important reference for the recognition model. When a new instruction arrives, the terminal determines whether it conforms to a known pattern through feature comparison to ensure the accuracy of recognition and the reliability of instruction execution.

[0071] Step S104: input the key fields of the industrial control instructions as feature information into a new industrial control instruction recognition model to obtain an industrial control instruction recognition result corresponding to the current network communication data.

[0072] Among them, the key fields of industrial control instructions refer to the important fields in the instructions extracted through protocol analysis. These fields can describe the core content and attributes of the instructions, including instruction type, target device, data content, etc. After characterization, these key fields will be used as input features of the recognition model for model recognition reasoning.

[0073] Exemplarily, after extracting the key fields of the industrial control instructions, the terminal can convert these fields into numerical feature vectors to meet the input requirements of the model. In a specific implementation, the terminal can map the field data to different dimensions of the feature vector according to the characteristics of different fields. For example, "itemId" can be mapped to the identifier of the target device, and the content of the "Data" field can be mapped to the actual operation data value or numerical sequence. At the same time, for time-related fields, the terminal can convert them into time series features so that the model can capture the dependencies in time series.

[0074] Next, the terminal inputs these feature vectors into the updated industrial control instruction recognition model. The recognition model usually includes a spatial feature extraction module, a short-term temporal feature extraction module, and a long-term temporal feature extraction module. The model first extracts the spatial distribution pattern in the features through a convolutional neural network (CNN), and then extracts the short-term and long-term dependency features of the instruction data through modules such as a recurrent neural network (RNN) and a long short-term memory network (LSTM). Finally, after these features are processed by the classification layer, the model outputs the recognition result and obtains the industrial control instruction type and parameter information corresponding to the current network communication data.

[0075] In the above-mentioned method for identifying industrial control instructions of energy storage power stations, first, the network communication data of the energy storage power station is obtained in real time. By obtaining the network communication data of the energy storage power station in real time, the latest instructions and status changes of each device in the energy storage power station can be dynamically perceived, thereby ensuring the real-time nature of industrial control instruction recognition; then, when the network communication data meets the preset conditions, the current industrial control instruction recognition model is updated according to the network communication data to obtain a new industrial control instruction recognition model. Through monitoring and conditional judgment of the network communication data, the recognition model is automatically updated under specific conditions, so that the model can adjust the weight or structure according to the new communication data to adapt to the changing factors in the scene, thereby effectively improving the recognition accuracy and adaptability of the model in different scenarios, and avoiding the recognition errors of traditional static models caused by data changes; then, the current network communication data of the energy storage power station is obtained, and according to the preset industrial control protocol, the industrial control instruction recognition model is updated. The current network communication data is parsed and processed to obtain the key fields of the industrial control instructions in the current network communication data. Through the preset parsing of the industrial control protocol, the key fields of the industrial control instructions, such as instruction type, device identification and operation parameters, can be extracted from the original data, and the core information of the industrial control instructions can be accurately parsed, so that the model can recognize and understand the specific meaning of each instruction, improve the accuracy of data parsing, and ensure the accuracy of recognition; finally, the key fields of the industrial control instructions are used as feature information and input into the new industrial control instruction recognition model to obtain the industrial control instruction recognition result corresponding to the current network communication data. The extracted key fields of the industrial control instructions are input into the model as feature information, which can enhance the model's understanding and recognition ability of the instructions, help the model quickly identify the instruction type and the corresponding operation, and improve the recognition efficiency and accuracy of the model, so that it can more accurately reflect the real-time instruction situation. In the above method, by acquiring the network communication data of the energy storage power station in real time, the recognition model is dynamically updated under specific conditions to adapt to the changes in the communication environment and improve the accuracy of recognition. Extracting the key fields of the instructions through protocol parsing can effectively extract feature information, quickly identify the instruction type and parameters, and realize real-time monitoring and efficient recognition of industrial control instructions.

[0076] In an exemplary embodiment, the above-mentioned step S102 updates the current industrial control instruction recognition model according to the network communication data to obtain a new industrial control instruction recognition model when the network communication data meets the preset conditions, and also includes: determining the industrial control instruction sending frequency and the number of new device connections within a preset time window according to the network communication data; when the industrial control instruction sending frequency and / or the number of new device connections exceed the corresponding preset thresholds, the model weights and input feature weights of the current industrial control instruction recognition model are updated according to the network communication data within the preset time window to obtain a new industrial control instruction recognition model.

[0077] The preset time window refers to a fixed time period used to count changes in network communication data characteristics. The frequency of sending industrial control instructions and the number of new devices connected are calculated during this time period. The frequency of sending industrial control instructions refers to the rate at which instructions are sent within the time window, while the number of new devices connected refers to the number of new devices connected during the time period. Changes in these parameters can reflect the dynamics of the network communication environment. When they exceed the threshold, they trigger the update of the model.

[0078] Exemplarily, the terminal can monitor the sending frequency of industrial control instructions and the number of new devices connected in the energy storage power station by setting a preset time window. For example, the terminal can count the number of instruction sending times and the number of new devices added at regular intervals to form statistical values ​​of the sending frequency of industrial control instructions and the number of devices connected. When the sending frequency of industrial control instructions and / or the number of new devices connected exceeds the preset threshold, the terminal determines that the current communication environment may have changed significantly, thereby triggering an update operation of the recognition model. In a specific implementation, the terminal will update the model weight and input feature weight of the current industrial control instruction recognition model based on the network communication data within the time window. The model weight update ensures that the model can adapt to the new communication frequency and device access mode, thereby improving the accuracy of recognition; the input feature weight update dynamically adjusts the feature priority, such as increasing the feature weight of the newly added device, so that the model can better recognize the instructions of the new device.

[0079] In this embodiment, by setting a time window and dynamically monitoring the frequency of command sending and the number of device accesses, it is possible to effectively detect changes in the network environment and quickly adjust the recognition model, thereby significantly improving the adaptability and recognition accuracy of the model in complex communication environments.

[0080] In an exemplary embodiment, the above-mentioned step S102 updates the current industrial control instruction recognition model according to the network communication data to obtain a new industrial control instruction recognition model when the network communication data meets the preset conditions, and also includes: determining the network communication data volume and the number of industrial control instruction sending types within the preset time window according to the network communication data; when the network communication data volume is less than the corresponding preset threshold, according to the network communication data volume, the model structure of the current industrial control instruction recognition model is trimmed to obtain a new industrial control instruction recognition model; and / or when the number of industrial control instruction sending types is less than the corresponding preset threshold, according to the number of industrial control instruction sending types, the input features of the current industrial control instruction recognition model are trimmed to obtain a new industrial control instruction recognition model.

[0081] Among them, the network communication data volume refers to the total flow size of data packets in the time window. The number of industrial control instruction types sent refers to the total number of different types of industrial control instructions sent in the time window. By setting these thresholds, the dynamics and complexity of the current environment can be judged, thereby deciding whether to tailor the model structure and features.

[0082] Exemplarily, the terminal can set a preset time window and continuously monitor the amount of network communication data and the number of industrial control instruction sending types within the window. For example, the terminal can count the amount of communication data and the number of instruction types within the time window at regular intervals. If the amount of network communication data is lower than the preset threshold, the terminal can determine that the current environment has become simple, so the industrial control instruction recognition model can be structurally trimmed, and redundant network layers or nodes can be deleted to simplify the model structure. If the number of industrial control instruction sending types is lower than the preset threshold, the terminal can trim the input features of the model according to the number of instruction types, remove less used features, reduce computational overhead and improve model operation efficiency. In a specific implementation, when the amount of network communication data is less than the corresponding preset threshold, the terminal will trim the model structure of the recognition model, remove low-weight or redundant network layers or nodes, to simplify the model structure and improve resource utilization; when the number of industrial control instruction sending types is less than the corresponding preset threshold, the terminal will trim the input features of the recognition model, remove uncommon features, and make the model more focused on common instruction types in the current environment.

[0083] In this embodiment, by monitoring the amount of network communication data and the number of types of industrial control instructions sent, the structure and characteristics of the recognition model can be dynamically tailored according to changes in the communication environment, making the model more streamlined and efficient. This tailoring mechanism improves the operating efficiency of the model in a simplified environment and effectively reduces the consumption of computing resources.

[0084] In an exemplary embodiment, Figure 2 As shown, the above step S104 uses the key field of the industrial control instruction as feature information and inputs it into the new industrial control instruction recognition model to obtain the industrial control instruction recognition result corresponding to the current network communication data, which can be achieved by the following steps:

[0085] Step S201, preprocessing the key fields of the industrial control instructions to obtain the preprocessed key fields of the industrial control instructions;

[0086] Step S202, standardizing the preprocessed key fields of the industrial control instructions to obtain standardized key fields of the industrial control instructions;

[0087] Step S203, converting the standardized key fields of the industrial control instructions into feature vectors;

[0088] Step S204: input the feature vector into a new industrial control instruction recognition model to obtain an industrial control instruction recognition result corresponding to the current network communication data.

[0089] Among them, the key fields of industrial control instructions refer to the fields extracted from network communication data that can describe the core content such as instruction type, target device, parameter value, etc., such as instruction identification, data content, etc. These fields constitute the key information source for the recognition model to make instruction judgments.

[0090] Exemplarily, after obtaining the key fields of the industrial control instructions, the terminal can process them through the following steps to convert them into the input features of the model: Preprocessing: The terminal performs data cleaning and format adjustment on the key fields of the industrial control instructions, removes invalid characters, corrects the data format, and ensures the normalization of the field content. For example, the terminal can clean up the blank characters in the key fields, unify the field format, or perform data type conversion on the data content. Through preprocessing, impurities in the data can be eliminated to ensure the accuracy of the feature information. Standardization processing: After the preprocessing is completed, the terminal performs standardization processing on the key fields of the industrial control instructions. Standardization processing can adjust the data to a unified dimension or range, for example, by normalization or Z-score standardization, the data of each field is converted to the same scale, so that the model can reasonably allocate the weights of different features. Feature vector conversion: The terminal converts the standardized key fields of the industrial control instructions into feature vectors to adapt to the input format of the model. Specifically, each key field can be mapped to a dimension in the feature vector. For example, the instruction identification field is mapped to one dimension in the feature vector, and the data content is mapped to another dimension to form a complete feature vector, which is convenient for the model to process and analyze. Finally, the terminal inputs the generated feature vector into a new industrial control instruction recognition model. The recognition model can determine the industrial control instruction type and operation parameters corresponding to the current network communication data based on the information in the feature vector, and generate accurate recognition results.

[0091] In this embodiment, by preprocessing, standardizing and converting feature vectors of key fields of industrial control instructions, the standardization of data formats and the consistency of input features are achieved, ensuring the high quality of the recognition model input data, thereby improving the accuracy of industrial control instruction recognition and the stability of the model.

[0092] In an exemplary embodiment, the new industrial control instruction recognition model includes a spatial feature extraction module, a short-time sequence feature extraction module, a long-time sequence feature extraction module and a recognition module; the above step S204 inputs the feature vector into the new industrial control instruction recognition model to obtain the industrial control instruction recognition result corresponding to the current network communication data, and also includes: using the spatial feature extraction module to perform spatial feature extraction processing on the feature vector to obtain the spatial feature corresponding to the feature vector; using the short-time sequence feature extraction module to perform short-time sequence feature extraction processing on the spatial feature to obtain the short-time sequence feature corresponding to the feature vector; using the long-time sequence feature extraction module to perform long-time sequence feature extraction processing on the short-time sequence feature to obtain the long-time sequence feature corresponding to the feature vector; inputting the long-time sequence feature into the recognition module to obtain the industrial control instruction recognition result corresponding to the current network communication data.

[0093] Among them, the spatial feature extraction module refers to the component used to extract local spatial features from the input feature vector. It can identify the spatial pattern of the instruction content and structure in the feature vector, which is usually implemented through a convolutional neural network (CNN). The short-term temporal feature extraction module and the long-term temporal feature extraction module are responsible for identifying the sequence dependencies in the feature vector at a shorter time and a longer time, respectively, and are suitable for implementation through a recurrent neural network (RNN) and a long short-term memory network (LSTM). The recognition module is used for the final instruction type and parameter recognition, and outputs the industrial control instruction recognition results.

[0094] Exemplarily, the terminal first inputs the feature vector into the spatial feature extraction module, which identifies the spatial structure pattern in the feature vector through convolution operation and generates spatial features representing the instruction content and field layout. For example, the terminal can use convolutional neural network to extract data distribution, field position and other information in the feature vector, so as to establish a spatial representation of the instruction content. Next, the terminal inputs the spatial features into the short-term temporal feature extraction module. The short-term temporal feature extraction module uses methods such as recurrent neural network or gated recurrent unit (GRU) to capture the sequence dependency of instructions in the short term and generate short-term temporal features, which reflect the short-term association and operation continuity between instructions. This is very effective for capturing continuous operations, adjacent steps or short-term behavior patterns in industrial control instructions. Then, the terminal passes the short-term temporal features to the long-term temporal feature extraction module, which extracts the long-term dependency of instructions through long short-term memory network (LSTM) and generates long-term temporal features. This module can capture the correlation of instructions over a long time span, ensure that the model can identify the long-term dependency pattern in the industrial control process, and adapt to the structured recognition requirements of complex instruction sequences. Finally, the terminal inputs the long-term time series features into the recognition module, which makes a final comprehensive judgment on all the extracted features and generates the industrial control instruction recognition results corresponding to the current network communication data through the fully connected layer or Softmax method, that is, the type of the current instruction, the target device and parameter information.

[0095] In this embodiment, by extracting spatial features, short-term timing features and long-term timing features step by step, the local patterns and timing dependencies in the industrial control instructions can be fully captured, and accurate instruction recognition results can be output. The modular processing method effectively improves the recognition accuracy of the model, making it more adaptable and stable in recognition in a changeable and complex industrial control environment.

[0096] In an exemplary embodiment, after obtaining the industrial control instruction identification result corresponding to the current network communication data in the above step S104, it also includes: monitoring the status information and execution result of the target industrial control instruction; the target industrial control instruction is the industrial control instruction corresponding to the industrial control instruction identification result; analyzing the status information and execution result to obtain the execution report of the target industrial control instruction, and uploading the execution report to the blockchain network.

[0097] Among them, the target industrial control instruction refers to a specific industrial control instruction identified by the industrial control instruction recognition model, which requires further status monitoring and execution analysis to ensure the effectiveness and correct execution of the instruction. The status information includes various feedback data during the execution of the instruction, such as execution progress, device response status, etc. The execution result is the final feedback data after the instruction is completed, which is used to judge the execution effect of the instruction.

[0098] Exemplarily, after obtaining the industrial control instruction recognition result, the terminal will automatically monitor the status information and execution result of the target industrial control instruction. Specifically, the terminal collects the execution status information of the instruction in real time through interaction with the target device, such as whether the instruction is received normally, the execution progress, the real-time status of the device, etc. At the same time, the terminal records the execution result after the instruction is completed to determine whether the instruction is successfully executed or there is an exception. After obtaining the status information and execution result, the terminal analyzes these data and generates an execution report. The analysis process includes multiple evaluations of key nodes, response delays, instruction completion status, etc. in the execution process to ensure that the instruction execution meets expectations. For example, the terminal can judge whether the execution effect of the target instruction meets the requirements based on indicators such as execution success rate and response time. If there is an exception, detailed fault information will be recorded. Finally, the terminal uploads the generated execution report to the blockchain network. As a decentralized data storage platform, the blockchain network can ensure the integrity and immutability of the execution report, thereby providing a safe and reliable basis for subsequent operation audits and traceability. After the execution report is uploaded to the chain, all parties can query the execution status of the instruction through the blockchain network to improve the transparency and security of the industrial control system.

[0099] In this embodiment, by monitoring the status information of industrial control instructions, analyzing the execution results and uploading the reports to the chain, the traceability and data security of the entire process of industrial control instruction execution are achieved. The introduction of blockchain ensures the authenticity and verifiability of the execution report, providing a safer and more reliable management method for the execution of industrial control instructions in energy storage power stations.

[0100] In another exemplary embodiment, Figure 3 As shown, the present application provides a method for identifying industrial control instructions of an energy storage power station, the method comprising the following steps:

[0101] Step S301: initialization and traffic collection.

[0102] Step S302: deep protocol analysis and instruction feature extraction.

[0103] Step S303: intelligent feature recognition model.

[0104] Step S304: online reasoning and real-time recognition.

[0105] Step S305: industrial control behavior instruction audit.

[0106] The intelligent feature recognition model in step S303 can be obtained by the following steps:

[0107] Step S3031: data preprocessing and feature extraction.

[0108] Step S3032, model construction and deep learning algorithm.

[0109] Step S3033: model training and optimization.

[0110] Step S3034, adaptive adjustment and cropping mechanism.

[0111] Exemplarily, in step S301, a comprehensive configuration initialization is first performed to ensure that all network traffic of the energy storage power station can be completely and efficiently captured and analyzed and processed. The system is connected to the mirror port of the switch in a bypass manner to obtain the full amount of data packet traffic. High-performance network card drivers and big data technologies are used to collect and process all traffic in real time to provide basic data for subsequent protocol parsing and behavior identification. The traffic collection process captures all network communication data of the energy storage power station in real time through the mirror port, ensuring the comprehensiveness of the collection and high-precision transmission of the data. The traffic data undergoes preliminary protocol parsing to extract basic information such as IP address, port number, and data packet size to provide support for in-depth instruction parsing.

[0112] In step S302, in-depth analysis based on industrial protocols is the core part of the entire identification process, especially for IEC 61850, IEC 104 and other protocols widely used in energy storage power stations. Through the parsing engine, the network message is gradually disassembled and key information is extracted. Main steps: ① Protocol parsing: In-depth analysis of IP layer and transport layer protocols, such as TCP / UDP, to ensure the integrity and validity of communication data. Then, for industrial control protocols IEC 61850 and IEC 104, specific logical nodes, data objects, remote control types and other key information are parsed. ② Instruction feature extraction: During the parsing process, key fields related to industrial control behavior instructions are extracted, such as invokeID, confirmedServiceRequest, itemId, Data, confirmedServiceResponse, etc., to ensure the correct transmission and execution of instructions. ③ Feature library construction: Based on the results of long-term monitoring and parsing, a huge instruction feature library is constructed, covering various instruction modes and instruction types. As an important basis for instruction recognition, this library can ensure the accuracy and effectiveness of instruction execution by comparing the features of new instructions with known patterns.

[0113] In step S3031, before executing instruction recognition, the collected energy storage power station network flow data is first preprocessed and feature data is extracted. The main goal of this step is to ensure the quality of the data and extract key fields for subsequent recognition.

[0114] Data cleaning: remove redundant and irrelevant data to ensure the validity of input data. Feature standardization: normalize the data to ensure that features of different ranges have consistent scales. Common methods include the Z-score normalization formula: . Where X is the original feature value, μ is the mean of the feature, and σ is the standard deviation. Feature extraction: Extract key field information in the network protocol, such as invokeID, confirmedServiceRequest, itemId, Data field, etc., to form a feature vector , used as input for the model.

[0115] In step S3032, a feature recognition model is constructed through a deep learning framework combining a convolutional neural network (CNN) and a recurrent neural network (RNN). The model can capture the spatiotemporal characteristics of industrial control instructions and adapt to different business scenarios.

[0116] ① Convolutional neural network (CNN) is used for spatial feature extraction:

[0117] Input: feature vector f of network data flow

[0118] Convolution layer: Use the convolution kernel W to extract local features. The formula is as follows:

[0119]

[0120] in, is the output of the convolutional layer, is the kth convolution kernel, is the input feature.

[0121] Pooling layer: downsamples the output of the convolutional layer to reduce the amount of data and enhance the translation invariance of the model. Maximum pooling is commonly used: .

[0122] ② Recurrent Neural Network (RNN) is used to capture time series features:

[0123] Input: Feature sequence processed by the convolution layer .

[0124] RNN unit: recursively processes time series features to capture the temporal dependencies in the instruction stream. The recursive formula of the classic RNN is: .in, is the hidden state at the current time step, is the current input, σ is the activation function, and is the weight matrix and b is the bias.

[0125] ③Long Short-Term Memory Network (LSTM):

[0126] In order to avoid the gradient vanishing problem in traditional RNN, the LSTM unit is introduced to better remember long time series information. The LSTM core structure includes a forget gate, an input gate, and an output gate. The calculation formula is:

[0127]

[0128]

[0129]

[0130]

[0131]

[0132] Among them, f t For the forget gate, i t is the input gate, o t is the output gate, c t is the unit state. W and b are the corresponding weight matrix and bias term. is the logistic activation function. t is the final hidden state output.

[0133] In step S3033, the collected network data is divided into a training set, a validation set and a test set to ensure the generalization ability of the model training.

[0134] Loss function: Use the cross-entropy loss function to evaluate the difference between the probability distribution of the model output and the true label:

[0135]

[0136] in, is the true label, The result predicted by the model.

[0137] Optimization algorithm: Use the Adam optimization algorithm to update the model parameters to accelerate convergence and avoid local optimal problems. The parameter update formula of Adam is:

[0138]

[0139]

[0140]

[0141] Among them, θ t is the model parameter, m t and v t are the momentum and square momentum of the gradient respectively, and α is the learning rate.

[0142] In step S3034, in order to ensure that the model can maintain efficient recognition when the business scenario changes, the model performance is actively optimized through adaptive adjustment and trimming mechanisms.

[0143] Adaptive adjustment: ① Scenario monitoring: Real-time monitoring of changes in traffic characteristics in energy storage power stations, capturing signals of business changes (such as new equipment, changes in instruction frequency, etc.). By setting the threshold, the adaptive adjustment module is automatically activated when an anomaly is detected. ② Model retraining: Through the online learning module, the model will be incrementally updated in small batches based on new data to ensure the adaptability of the model to new scenarios. The incremental learning method is used to prevent the model from forgetting its recognition capabilities in historical scenarios. ③ Dynamic weight update: The weight of the feature will be recalculated based on the feature distribution in the new scenario. For example, some features (such as itemId or Data fields) may have different importance in different business scenarios. The feature weights are dynamically adjusted through the following formula: Among them, w i The feature f i The weight, μ i is the feature mean, and α is the adjustment rate.

[0144] Pruning mechanism: ① Model complexity pruning: When the traffic scale or instruction type in the scene is reduced, the model complexity can be reduced by pruning unnecessary network layers or nodes, thereby improving computing efficiency. L1 regularization method is used to prune unimportant weights: . Among them, λ is the regularization coefficient, θ i is the network weight. ② Redundant feature pruning: By analyzing the importance of features in real time, pruning low-weight or redundant features can reduce the model calculation burden and improve real-time response capabilities.

[0145] In step S304, after the model completes the adaptive adjustment, it enters the online reasoning stage to process the collected new data in real time. Through the optimized model, the industrial control instructions in the energy storage power station can be quickly and accurately identified to ensure the efficient execution and safety audit of the instructions.

[0146] In step S305, after the intelligent feature recognition algorithm identifies the instruction, the execution process of each instruction is audited and recorded in detail. The audit function includes the following: Instruction execution time and response time: record the instruction issuance time, execution time, response time, etc. to ensure that the execution of each instruction meets expectations. Execution result analysis: by analyzing the instruction execution results, the reasons for the success or failure of the execution are fed back in real time, and the information is uploaded to the blockchain network to ensure the immutability and traceability of the data.

[0147] In this embodiment, through the intelligent feature recognition algorithm, various industrial control instructions in the energy storage power station can be efficiently and accurately identified. Compared with the traditional fixed rule recognition method, the intelligent algorithm has dynamic adjustment and self-learning capabilities, and can automatically update the recognition logic according to the changes in the actual business scenario, which greatly improves the accuracy and efficiency of instruction recognition. This performance improvement ensures that the energy storage power station can accurately identify and execute each instruction in a complex and changeable network environment, reducing the risk of misoperation. By integrating efficient traffic monitoring and deep protocol parsing technology, it can analyze the data flow in the energy storage power station network in real time, and perform immediate processing and auditing of key industrial control instructions. When abnormal instructions or unauthorized operations are detected, it can respond quickly and take corrective measures. Compared with the lag of traditional monitoring technology, the delay of instruction recognition and response is greatly shortened, and the safety and stability of the energy storage power station are improved. The adaptive learning algorithm is adopted, which can autonomously adapt to the access of new equipment or changes in business scenarios without frequent manual intervention or readjustment of recognition rules, thereby reducing maintenance costs. In addition, this method has good scalability and can support the network scale and instruction types that may be expanded in the future of energy storage power stations. Through the self-optimization mechanism of the model, it can be continuously updated as the business develops to ensure its long-term and efficient operation. The blockchain technology ensures the security and immutability of all instruction audit data, greatly improving the operational transparency and audit reliability of the energy storage power station. The execution process, time and results of each instruction are recorded in detail and stored on the blockchain to ensure the integrity and traceability of the data. It provides strong protection for the safe operation of the energy storage power station, reduces potential operational risks, and enhances the safety and stability of the power grid operation.

[0148] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0149] Based on the same inventive concept, the embodiment of the present application also provides an energy storage power station industrial control instruction recognition device for implementing the energy storage power station industrial control instruction recognition method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more energy storage power station industrial control instruction recognition device embodiments provided below can refer to the limitations of the energy storage power station industrial control instruction recognition method above, and will not be repeated here.

[0150] In an exemplary embodiment, Figure 4 As shown, a device for identifying industrial control instructions for energy storage power stations is provided, including: a data acquisition module 401, a model update module 402, a field parsing module 403 and an instruction identification module 404, wherein:

[0151] Data acquisition module 401, used to acquire network communication data of energy storage power station in real time;

[0152] The model updating module 402 is used to update the current industrial control instruction recognition model according to the network communication data to obtain a new industrial control instruction recognition model when the network communication data meets the preset conditions;

[0153] The field parsing module 403 is used to obtain the current network communication data of the energy storage power station, parse and process the current network communication data according to the preset industrial control protocol, and obtain the key fields of the industrial control instructions in the current network communication data;

[0154] The instruction recognition module 404 is used to input the key fields of the industrial control instruction as feature information into a new industrial control instruction recognition model to obtain the industrial control instruction recognition result corresponding to the current network communication data.

[0155] In one of the embodiments, the above-mentioned model update module 402 is also used to determine the frequency of industrial control instruction sending and the number of new device connections within a preset time window based on network communication data; when the frequency of industrial control instruction sending and / or the number of new device connections exceed the corresponding preset thresholds, the model weight and input feature weight of the current industrial control instruction recognition model are updated according to the network communication data within the preset time window to obtain a new industrial control instruction recognition model.

[0156] In one of the embodiments, the above-mentioned model updating module 402 is also used to determine the amount of network communication data and the number of industrial control instruction sending types within a preset time window based on the network communication data; when the amount of network communication data is less than the corresponding preset threshold, the model structure of the current industrial control instruction recognition model is trimmed according to the amount of network communication data to obtain a new industrial control instruction recognition model; and / or when the number of industrial control instruction sending types is less than the corresponding preset threshold, the input features of the current industrial control instruction recognition model are trimmed according to the number of industrial control instruction sending types to obtain a new industrial control instruction recognition model.

[0157] In one of the embodiments, the above-mentioned instruction recognition module 404 is also used to preprocess the key fields of the industrial control instruction to obtain the preprocessed key fields of the industrial control instruction; standardize the preprocessed key fields of the industrial control instruction to obtain the standardized key fields of the industrial control instruction; convert the standardized key fields of the industrial control instruction into a feature vector; input the feature vector into a new industrial control instruction recognition model to obtain the industrial control instruction recognition result corresponding to the current network communication data.

[0158] In one embodiment, the new industrial control instruction recognition model includes a spatial feature extraction module, a short-time sequence feature extraction module, a long-time sequence feature extraction module and an identification module; the above-mentioned instruction recognition module 404 is also used to use the spatial feature extraction module to perform spatial feature extraction processing on the feature vector to obtain the spatial feature corresponding to the feature vector; use the short-time sequence feature extraction module to perform short-time sequence feature extraction processing on the spatial feature to obtain the short-time sequence feature corresponding to the feature vector; use the long-time sequence feature extraction module to perform long-time sequence feature extraction processing on the short-time sequence feature to obtain the long-time sequence feature corresponding to the feature vector; input the long-time sequence feature into the identification module to obtain the industrial control instruction recognition result corresponding to the current network communication data.

[0159] In one of the embodiments, the above-mentioned energy storage power station industrial control instruction identification device also includes an instruction monitoring module for monitoring the status information and execution results of the target industrial control instruction; the target industrial control instruction is the industrial control instruction corresponding to the industrial control instruction identification result; analyze the status information and execution results, obtain the execution report of the target industrial control instruction, and upload the execution report to the blockchain network.

[0160] Each module in the above energy storage power station industrial control instruction recognition device can be implemented in whole or in part by software, hardware and their combination. Each of the above modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to each of the above modules.

[0161] In an exemplary embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as shown in FIG. Figure 5 As shown. The computer device includes a processor, a memory, an input / output interface, a communication interface, a display unit and an input device. The processor, the memory and the input / output interface are connected through a system bus, and the communication interface, the display unit and the input device are connected to the system bus through the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and the external device. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be realized through WIFI, a mobile cellular network, near field communication (NFC) or other technologies. When the computer program is executed by the processor, a method for identifying industrial control instructions of an energy storage power station is realized. The display unit of the computer device is used to form a visually visible picture, which can be a display screen, a projection device or a virtual reality imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or a button, trackball or touchpad set on the computer device shell, or an external keyboard, touchpad or mouse.

[0162] Those skilled in the art will understand that Figure 5 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0163] In one embodiment, a computer device is further provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above method embodiments when executing the computer program.

[0164] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0165] In one embodiment, a computer program product is provided, including a computer program, which implements the steps in the above method embodiments when executed by a processor.

[0166] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0167] A person of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment method can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, an artificial intelligence (AI) processor, etc., but are not limited to this.

[0168] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0169] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A method for identifying industrial control instructions of an energy storage power station, characterized in that: The method comprises: Obtain network communication data of energy storage power stations in real time; When the network communication data meets the preset conditions, the current industrial control instruction recognition model is updated according to the network communication data to obtain a new industrial control instruction recognition model; Acquire current network communication data of the energy storage power station, parse and process the current network communication data according to a preset industrial control protocol, and obtain key fields of industrial control instructions in the current network communication data; The key field of the industrial control instruction is used as feature information and input into the new industrial control instruction recognition model to obtain the industrial control instruction recognition result corresponding to the current network communication data.

2. The method according to claim 1, characterized in that When the network communication data meets the preset condition, the current industrial control instruction recognition model is updated according to the network communication data to obtain a new industrial control instruction recognition model, including: Determine the frequency of sending industrial control instructions and the number of new devices connected within a preset time window according to the network communication data; When the frequency of sending the industrial control instructions and / or the number of new devices connected exceeds the corresponding preset threshold, the model weight and input feature weight of the current industrial control instruction recognition model are updated according to the network communication data within the preset time window to obtain a new industrial control instruction recognition model.

3. The method according to claim 1, characterized in that When the network communication data meets the preset condition, the current industrial control instruction recognition model is updated according to the network communication data to obtain a new industrial control instruction recognition model, including: Determining the amount of network communication data and the number of types of industrial control instructions sent within a preset time window according to the network communication data; When the network communication data volume is less than the corresponding preset threshold, the model structure of the current industrial control instruction recognition model is trimmed according to the network communication data volume to obtain a new industrial control instruction recognition model; and / or When the number of the industrial control instruction sending types is less than the corresponding preset threshold, the input features of the current industrial control instruction recognition model are trimmed according to the number of the industrial control instruction sending types to obtain a new industrial control instruction recognition model.

4. The method according to claim 1, characterized in that: The step of inputting the key field of the industrial control instruction as feature information into the new industrial control instruction recognition model to obtain the industrial control instruction recognition result corresponding to the current network communication data includes: Preprocessing the industrial control instruction key field to obtain the preprocessed industrial control instruction key field; Performing standardization processing on the preprocessed industrial control instruction key fields to obtain standardized industrial control instruction key fields; Converting the standardized key fields of the industrial control instructions into feature vectors; The feature vector is input into the new industrial control instruction recognition model to obtain the industrial control instruction recognition result corresponding to the current network communication data.

5. The method according to claim 4, characterized in that The new industrial control instruction recognition model includes a spatial feature extraction module, a short-time sequence feature extraction module, a long-time sequence feature extraction module and a recognition module; The step of inputting the feature vector into the new industrial control instruction recognition model to obtain the industrial control instruction recognition result corresponding to the current network communication data includes: Using the spatial feature extraction module to perform spatial feature extraction processing on the feature vector to obtain the spatial feature corresponding to the feature vector; The short-time time series feature extraction module is used to perform short-time time series feature extraction processing on the spatial feature to obtain the short-time time series feature corresponding to the feature vector; Using the long-term time series feature extraction module to perform long-term time series feature extraction processing on the short-term time series feature to obtain the long-term time series feature corresponding to the feature vector; The long-term timing characteristics are input into the recognition module to obtain the industrial control instruction recognition result corresponding to the current network communication data.

6. The method according to claim 1, characterized in that After obtaining the industrial control instruction recognition result corresponding to the current network communication data, the method further includes: Monitoring the status information and execution result of a target industrial control instruction; the target industrial control instruction is an industrial control instruction corresponding to the industrial control instruction identification result; Analyze the status information and the execution result to obtain an execution report of the target industrial control instruction, and upload the execution report to the blockchain network.

7. An energy storage power station industrial control instruction recognition device, characterized in that: The device comprises: A data acquisition module is used to acquire network communication data of the energy storage power station in real time; A model updating module, used to update the current industrial control instruction recognition model according to the network communication data to obtain a new industrial control instruction recognition model when the network communication data meets the preset conditions; A field parsing module, used to obtain the current network communication data of the energy storage power station, parse and process the current network communication data according to a preset industrial control protocol, and obtain the key fields of the industrial control instructions in the current network communication data; The instruction recognition module is used to input the key field of the industrial control instruction as feature information into the new industrial control instruction recognition model to obtain the industrial control instruction recognition result corresponding to the current network communication data.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images