Electric heating gas state information transmission cooperative power grid energy disturbance suppression method

Abstract

The invention provides a power grid energy disturbance suppression method for electric heating gas state information transmission collaboration, and relates to the technical field of information, and the method comprises the steps: obtaining multi-source heterogeneous state information of an electric heating gas energy network; aiming at the multi-source heterogeneous state information, dynamically converting a heterogeneous communication protocol into a unified protocol format by adopting a self-adaptive communication protocol conversion technology; and for the decoupled state information, parallel processing is carried out by adopting a distributed stream computing framework, so that the real-time performance and expandability of state information processing are improved. A distributed stream computing framework is adopted to construct an efficient processing assembly line, the block chain technology is utilized to guarantee the integrity and reliability of data transmission, a deep learning algorithm is applied to construct an intelligent analysis model, a multi-energy-flow coupling relationship and an evolution law are mined, and a multi-energy-flow analysis result is obtained. According to the method, data processing results of all links are fused to form a global state information view, decision support is provided for power grid energy disturbance suppression, and efficient processing, analysis and visualization of the state information of the electric heating gas multi-energy flow system are achieved.

Description

Technical Field

[0001] The present invention relates to the field of information technology, and in particular to a method for suppressing power grid energy disturbance in coordination with transmission of electric, thermal and gas state information. Background Art

[0002] There are many technical challenges and contradictions in the process of coordinated transmission of electricity, heat and gas state information. First, the suppression of power grid energy disturbances places extremely high demands on the real-time and reliability of electricity, heat and gas state information transmission. Transmission delays and data loss will directly affect the timeliness and accuracy of power grid regulation. Secondly, the physical characteristics of electricity, heat and gas energy systems are significantly different, and the acquisition frequency, data format and time scale of their state information are different. How to achieve the interoperability and integration of state information of heterogeneous energy networks is a major problem. Furthermore, the energy network has a large geographical span and a wide variety of equipment. The communication modes and protocol standards of different regions and equipment are not the same, and it is very difficult to coordinate various communication resources. At the same time, the fluctuations of electricity, heat and gas energy are strongly coupled. The state changes of a single energy system will be quickly transmitted to other energy networks. Multiple energy flows affect each other, making the decoupling and tracing of state information very difficult. Finally, the amount of electricity, heat and gas state information is huge and explosively growing. The real-time processing and intelligent analysis of large-scale multi-source heterogeneous state information also face huge challenges. In short, to achieve efficient and coordinated transmission of electricity, heat and gas status information and lay a solid information foundation for suppressing grid energy disturbances, it is also necessary to work together from multiple technical dimensions such as communication, computing, and control to break through key bottlenecks. Summary of the invention

[0003] The present invention provides a method for suppressing power grid energy disturbances in coordination with electric, thermal and gas state information transmission, which mainly includes:

[0004] Acquire multi-source heterogeneous state information of the electric, thermal and gas energy network; for the multi-source heterogeneous state information, use adaptive communication protocol conversion technology to dynamically convert heterogeneous communication protocols into a unified protocol format; for the converted state information, build a multi-energy flow coupling relationship diagram of the energy network based on the graph model, and realize the decoupling and traceability of the state information through the causal reasoning analysis method; for the decoupled state information, use a distributed stream computing framework for parallel processing to improve the real-time and scalability of state information processing; use a blockchain-based security consensus mechanism to confirm and verify the results of the stream computing to ensure the credibility of the state information; build a state information intelligent analysis model based on a deep learning algorithm to mine the correlation characteristics and evolution laws contained in the multi-energy flow data; generate a global state information view of the electric, thermal and gas energy network through visualization technology to provide auxiliary decision-making for energy network scheduling optimization.

[0005] Furthermore, the acquisition of multi-source heterogeneous status information of the electric, thermal and gas energy network includes: determining the collection frequency, data format and time scale of the status information of each energy system according to the physical characteristics of the electric, thermal and gas energy network; converting the status information of the heterogeneous energy network into a unified format through data standardization processing; and realizing the interoperability and fusion of status information.

[0006] Furthermore, the adaptive communication protocol conversion technology is used to dynamically convert heterogeneous communication protocols into a unified protocol format, including: dynamically matching and converting various communication protocols according to differences in communication modes and protocol standards in different regions and devices; using data segmentation, parallel transmission and other technologies to improve communication efficiency; and ensuring communication security through lightweight encryption and digital signatures.

[0007] Furthermore, the energy network multi-energy flow coupling relationship diagram is constructed based on the graph model, and the decoupling and traceability of state information is achieved through causal reasoning analysis methods, including: targeting the strong coupling of electric, thermal and gas energy fluctuations, using a graph model to characterize the correlation structure of multi-energy flow data; learning the implicit dependency relationship of multi-energy flow data through graph embedding and graph neural networks; combining the law of conservation of energy, building a causal reasoning rule base, and tracing the influence chain of state changes of a single energy system.

[0008] Furthermore, the use of a distributed stream computing framework for parallel processing includes: building an efficient state information processing pipeline for massive multi-source heterogeneous state information; improving the real-time and scalability of state information processing through technologies such as data sharding and parallel computing; and using optimization strategies such as in-memory computing and incremental processing to reduce computing overhead.

[0009] Furthermore, the use of a blockchain-based security consensus mechanism for rights confirmation and verification includes: dividing the stream calculation results into blocks, generating hash fingerprints and timestamps; realizing node collaborative verification in the blockchain network through consensus algorithms and smart contracts; and storing the verified status information on the chain to ensure that the data cannot be tampered with and is traceable.

[0010] Furthermore, the state information intelligent analysis model based on the deep learning algorithm is constructed, including: for multi-source heterogeneous state information, using the transfer learning method to achieve cross-domain knowledge fusion; through the spatiotemporal sequence prediction model, mining the dynamic change rules of state information; combining the reinforcement learning framework to adaptively optimize the scheduling strategy of the energy network.

[0011] Furthermore, the global status information view of the electric, heat and gas energy network is generated by visualization technology, including: designing a multi-level visualization view for multi-dimensional and multi-granular status information; supporting dynamic query and correlation analysis of status information through a human-computer interaction interface; combining virtual reality and augmented reality technologies to construct an immersive visualization scene to improve the intuitiveness and understandability of system operation.

[0012] The technical solution provided by the embodiment of the present invention may have the following beneficial effects:

[0013] The present invention discloses a method for suppressing power grid energy disturbances in coordination with the transmission of electric, thermal and gas state information. In view of the strong coupling and heterogeneity of the electric, thermal and gas energy network, the present invention adopts adaptive communication protocol conversion technology to achieve unified coordination of various communication resources, and constructs a multi-energy flow coupling relationship diagram through a causal reasoning method based on a graph model to achieve decoupling and traceability of state information. In order to process massive multi-source heterogeneous state information, the present invention adopts a distributed stream computing framework to construct an efficient processing pipeline. At the same time, blockchain technology is used to ensure the integrity and reliability of data transmission, and a deep learning algorithm is applied to construct an intelligent analysis model to explore the multi-energy flow coupling relationship and evolution law. Finally, the present invention integrates the data processing results of each link to form a global state information view, provides decision support for power grid energy disturbance suppression, and realizes efficient processing, analysis and visualization of the state information of the electric, thermal and gas multi-energy flow system. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 The present invention is a flow chart of the method for suppressing power grid energy disturbances in coordination with transmission of electric, heat and gas state information.

[0015] Figure 2 It is a schematic diagram of the grid energy disturbance suppression method coordinated with the transmission of electric, heat and gas state information of the present invention.

[0016] Figure 3 It is another schematic diagram of the grid energy disturbance suppression method coordinated with the transmission of electricity, heat and gas state information of the present invention. DETAILED DESCRIPTION

[0017] In order to enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below in conjunction with the drawings in the embodiments of this specification. Obviously, the described embodiments are only part of the embodiments of this specification, not all of the embodiments. Based on the embodiments in this specification, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of this specification.

[0018] like Figure 1-Figure 3 The method for suppressing power grid energy disturbances in coordination with transmission of electric, heating and gas state information in this embodiment may specifically include:

[0019] Step S101, according to the physical characteristics of the electric, thermal and gas energy network, determine the collection frequency, data format and time scale of the status information of each energy system, and convert the status information of the heterogeneous energy network into a unified format through data standardization processing to achieve interoperability and integration of status information.

[0020] According to the physical characteristics of the electric, thermal and gas energy network, the state information of each energy system is obtained, including parameters such as energy flow, temperature and pressure. In view of the problem of inconsistent data acquisition frequency and time scale of different energy systems, the acquisition frequency and time scale are unified through the linear interpolation algorithm, so that the state information of each energy system is aligned in the time dimension. In view of the problem of inconsistent data formats of different energy systems, a unified data format is designed, including the data type and unit of parameters such as energy flow, temperature and pressure. Then, through data mapping and conversion rules, the state information of heterogeneous energy networks is converted into a unified data format to achieve data standardization. According to the unified data format, the data model of energy network state information is designed using the relational model, the data table and association relationship of parameters such as energy flow, temperature and pressure are defined, and a standardized data storage structure is established. Through the Kalman filter algorithm, the state information of different energy systems is fused, and the measurement error and process noise of each energy system are comprehensively considered to obtain the global state estimation value of the entire energy network, thereby improving the accuracy and reliability of the state information. Based on the global state information, the linear programming algorithm is used to consider factors such as energy balance constraints, equipment capacity constraints, and energy prices, and the optimal configuration of electricity, heat, gas, and other energies is calculated to achieve economic optimization scheduling of the energy network. The optimized scheduling strategy is sent to each energy system, and the operating parameters of each energy system are dynamically adjusted through the PID control algorithm according to the set value of the scheduling strategy and the feedback value of the actual state information, so as to achieve coordinated optimization control of the electric, heat, and gas energy network and improve energy utilization efficiency and system stability.

[0021] Specifically, the electric, thermal and gas energy network is a complex system composed of an electric power system, a thermal system and a gas system. In order to achieve collaborative optimization control, it is necessary to first obtain the state information of each energy system. For example, the state information of the electric power system includes voltage, current, frequency, power, etc.; the state information of the thermal system includes temperature, pressure, flow, etc.; the state information of the gas system includes pressure, flow, composition, etc. This information is collected by sensors and measuring instruments, such as smart meters, temperature sensors, pressure sensors, etc. Since the data collection frequency and time scale of different energy systems are inconsistent, they need to be unified. For example, the collection frequency of the electric power system may be at the second level, while the collection frequency of the thermal system may be at the minute level. At this time, a linear interpolation algorithm can be used to interpolate low-frequency data to a high-frequency time scale. Assuming that the temperature of the thermal system at 0 minutes and 1 minute is 60℃ and 62℃ respectively, the temperature at 5 minutes can be estimated to be 61℃ by linear interpolation. This can align the data of different energy systems in the time dimension, facilitating subsequent data analysis and processing. The data formats of different energy systems may also be inconsistent. For example, the unit of temperature may be Celsius or Fahrenheit, and the unit of power may be kilowatt or megawatt. Therefore, it is necessary to design a unified data format, for example, Celsius is used for temperature and megawatt is used for power. Then, through data mapping and conversion rules, the state information of heterogeneous energy networks is converted into a unified data format to achieve data standardization. For example, Fahrenheit is converted into Celsius, and kilowatt is converted into megawatt. After data standardization, it is convenient to exchange and share data between different systems. The unified data can be stored in a relational database. Data tables can be designed to store the state information of electricity, heat, and gas respectively. For example, the "electricity" table contains fields such as time, voltage, current, frequency, and power; the "heat" table contains fields such as time, temperature, pressure, and flow; and the "gas" table contains fields such as time, pressure, flow, and composition. The "time" field is used as the primary key to associate data from different energy systems. This standardized data storage structure facilitates data query and management. In order to improve the accuracy and reliability of state information, the Kalman filter algorithm can be used to fuse the state information of different energy systems. The Kalman filter algorithm can comprehensively consider the measurement errors and process noise of each energy system. For example, there may be errors in the voltage measurement value of the power system, and there may also be errors in the temperature measurement value of the thermal system. The Kalman filter algorithm can estimate more accurate state values ​​based on these measurement values ​​and the state equation of the system. This can reduce the impact of measurement errors on state estimation. Based on the global state information, a linear programming algorithm can be used for economic optimization scheduling of energy networks. For example, considering the energy balance constraints of electricity, heat, and gas, equipment capacity constraints, energy prices and other factors, the optimal energy configuration plan is calculated. Assuming that the electricity price is low, electricity can be used for energy supply first; if the thermal demand is high, the output of the thermal system can be increased.Linear programming algorithms can help find the optimal solution that meets all constraints, thereby achieving economic optimization scheduling of energy networks. After the optimized scheduling strategy is issued to each energy system, the operating parameters of each energy system are dynamically adjusted through the PID control algorithm. The PID control algorithm calculates the control quantity, such as adjusting the opening of the valve and the output of the generator, based on the set value of the scheduling strategy and the feedback value of the actual state information. For example, if the set temperature of the thermal system is 70°C and the actual temperature is 65°C, the PID control algorithm will calculate a control quantity based on the deviation to increase the output of the thermal system so that the temperature reaches the set value. The PID control algorithm can realize the coordinated optimization control of the electric, thermal and gas energy network and improve energy utilization efficiency and system stability.

[0022] Step S102, in view of the differences in communication modes and protocol standards of different regions and devices, adaptive communication protocol conversion technology is used to dynamically match and convert various communication protocols, achieve unified coordination of various communication resources, and ensure efficient transmission of electric, heating and gas status information.

[0023] According to the differences in communication modes and protocol standards of different regions and devices, a pre-established communication protocol knowledge base is obtained, which contains information such as the format, parameters, and conversion rules of various communication protocols. For the received electric heating gas status information, its communication protocol type and parameters are extracted, and it is matched and compared with the protocol in the communication protocol knowledge base to determine whether protocol conversion is required. If protocol conversion is required, the Scapy library in Python is used to dynamically generate a protocol conversion script according to the conversion rules in the communication protocol knowledge base to convert the format and parameters of the original communication protocol. After protocol conversion, the electric heating gas status information data in a unified format is obtained. According to the preset data format standard, the Python Pandas library is used to format the converted data, and the NumPy library is used to perform numerical range and integrity verification to ensure the accuracy and integrity of the data. The verified electric heating gas status information data is efficiently transmitted through the Mosquitto message queue service in accordance with the unified JSON format and MQTT protocol standard, and Grafana is used to monitor and control the transmission process in real time. In the process of communication resource coordination, a dynamic resource allocation algorithm based on device priority is adopted. According to the communication mode, data volume and preset device priority of the equipment in different regions, the Python Schedule library is used to dynamically schedule and allocate MQTT topics, QoS and other communication resources to ensure the efficiency and reliability of data transmission. In view of the real-time requirements of the status information of electric heating and gas, the retention flag and QoS level of the MQTT message are set, and the CoAP protocol is used to asynchronously report the status change data to minimize the data transmission delay and ensure that the status information can be transmitted to the data center in a timely and accurate manner. On the data center side, the Kafka message queue is used to cache the received status information, and Spark Streaming is used for real-time data processing and analysis to extract key feature parameters to achieve real-time monitoring and fault warning of electric heating and gas equipment.

[0024] Specifically, in the electric, thermal and gas energy network, the communication modes and protocol standards of different regions and devices vary greatly, so a communication protocol knowledge base is needed to store the formats, parameters and conversion rules of various communication protocols. For example, the power system may use the Modbus protocol, while the thermal system may use the BACnet protocol. Through the pre-established communication protocol knowledge base, data from different systems can be quickly identified and processed. When receiving the status information of the power system, the system first identifies that it uses the Modbus protocol, and then uses the Scapy library to generate a protocol conversion script according to the conversion rules in the knowledge base to convert the Modbus protocol into a unified communication format of the system. The converted data requires further format verification and value range check to ensure the accuracy and integrity of the data. The Pandas library of Python can be used to verify the data format to ensure that all data fields meet the preset standards, such as timestamp format, value type, etc. At the same time, the NumPy library is used to check the value range of the data, such as whether the current and voltage values ​​are within a reasonable range. This step ensures the quality of the data before entering the analysis and processing process. In order to efficiently transmit these verified data, the system uses the JSON format and MQTT protocol. Through the Mosquitto message queue service, data can be efficiently transmitted between different systems and devices. For example, the data transmitted from the power station to the dispatch center can be guaranteed to be reliable by setting the MQTT QoS level. At the same time, using Grafana to monitor the transmission process in real time, you can view the data flow status and possible transmission bottlenecks in real time. In the process of communication resource coordination, the dynamic resource allocation algorithm dynamically schedules communication resources according to the priority and data volume of the equipment. For example, for data transmission of key equipment such as substations, the system may allocate a higher QoS level and a more frequent message sending interval. This priority-based resource allocation strategy ensures the timely transmission of key data without affecting the operating efficiency of the entire energy network due to network congestion. In response to the real-time requirements of the electric, hot and gas status information, the system greatly reduces the data transmission delay by setting the retention flag and QoS level of the MQTT message and using the CoAP protocol to asynchronously report the status change. This is critical for quickly responding to system changes, such as sudden failures or changes in demand. For example, if a sudden drop in voltage in a certain area is detected, the system can immediately report this change through the CoAP protocol, and the dispatch center can respond quickly and adjust the operating status of the relevant equipment. On the data center side, Kafka message queues are used to cache received status information, and Spark Streaming is used for real-time data processing and analysis. This step can extract key feature parameters from a large amount of real-time data to achieve real-time monitoring and fault warning of electric, heating and gas equipment.For example, by analyzing continuously received temperature data, the system can predict the risk of overheating of equipment and promptly notify the maintenance team to conduct inspections or adjustments, thereby avoiding possible equipment failure.

[0025] Step S103, in view of the strong coupling of electricity, heat and gas energy fluctuations, a causal reasoning method based on a graphical model is used to construct a multi-energy flow coupling relationship diagram. Through causal reasoning, the impact of the state change of a single energy system on other energy networks is traced to achieve decoupling and tracing of state information.

[0026] The real-time status data of electric energy, thermal energy and gas energy are obtained, and the data is preprocessed to extract the key features reflecting the energy fluctuation. According to the physical coupling relationship between each energy, a multi-energy flow coupling relationship diagram is constructed using a graph model to characterize the mutual influence between energies. Among them, the nodes of the graph represent different energies, and the edges represent the coupling relationship between energies. For the nodes and edges in the graph model, the quantitative causal relationship strength is defined through physics knowledge and expert experience to form the basis of causal reasoning. When it is detected that the state characteristic value of a certain energy system exceeds the preset threshold, it is judged that the energy system has a significant state change, and the potential impact of the change on other energies is traced on the causal graph. The propagation path and influence range of the state change in the multi-energy flow network are determined by the graph-based shortest path search algorithm. According to the causal reasoning results, the affected energy system and the corresponding state change amplitude are obtained. For the state change information of each energy system, the principal component analysis method is used to perform feature dimension reduction and information decoupling, and the independent state change information of each energy system is obtained to achieve the traceability and decoupling of state information.

[0027] Specifically, obtain the real-time status data of electric energy, thermal energy and gas energy, for example, obtain the electricity consumption of each household per minute through smart meters, obtain the heating temperature and flow through sensors at thermal power stations, and obtain the natural gas usage through flow meters on gas pipelines. Preprocess the data, such as interpolating missing data, removing abnormal data, and smoothing the data to reduce noise interference. Extract key features that reflect energy fluctuations, such as peak, valley and average values ​​of electricity consumption, the rate of change of heating temperature, and the fluctuation amplitude of natural gas usage. These features can reflect the operating status and change trend of the energy system. In order to study the mutual influence between different energies, it is necessary to construct a multi-energy flow coupling relationship graph. The nodes of the graph represent different types of energy, such as electric energy, thermal energy and gas energy. The edges represent the coupling relationship between energies. For example, natural gas can be used to generate electricity, thereby converting gas energy into electric energy; electric energy can be used to drive heat pumps to convert electric energy into thermal energy; natural gas can be used directly for heating to convert gas energy into thermal energy. The weights on nodes and edges can represent the efficiency or proportional relationship of energy conversion. For example, the efficiency of gas-fired power generation can be set to 4, indicating that 4 units of electricity can be generated for every 1 unit of natural gas input. For the nodes and edges in the graph model, the quantitative causal relationship strength is defined through physics knowledge and expert experience. For example, rising gas prices will cause more users to use electricity for heating, so the causal relationship strength of gas energy to electricity can be set to 8, indicating that rising gas prices have a strong impact on electricity demand. If a region uses a large number of heat pumps for heating, then the impact of electricity price fluctuations on thermal energy demand will also be relatively large, so the causal relationship strength of electricity to thermal energy can be set to 7. These causal relationship strength values ​​form the basis of causal reasoning. Assuming that a sudden increase in power load is detected, exceeding the preset threshold, it is judged that a significant state change has occurred in the power system. At this time, it is necessary to trace the potential impact of this change on other energy on the causal graph. The impact path of the increase in power load can be found through the graph-based shortest path search algorithm. For example, if the path is "electricity-thermal energy", it means that the increase in power load mainly affects the thermal energy system. Assuming that the area uses a large number of electric heat pumps for heating, due to rising electricity prices or other reasons, the power load increases, and users may reduce the use of heat pumps, resulting in a decrease in heat demand. According to the results of causal reasoning, the affected energy system and the corresponding state change amplitude can be obtained. For example, through causal reasoning, it can be concluded that a 10% increase in power load may lead to a 5% decrease in heat demand. This information can be used to guide the coordinated operation and optimal control of multi-energy systems. Finally, in order to distinguish between the state changes of different energy systems themselves and the state changes affected by other energy systems, the state information needs to be decoupled. For example, suppose the power load increases by 10%, and at the same time, the heat demand increases by 8% due to the decrease in temperature.By using the principal component analysis method to perform feature dimensionality reduction and information decoupling, the increase in power load and the increase in thermal energy demand can be separated, and independent state change information of the power system and thermal energy system can be obtained respectively. This helps to more accurately analyze the operating status of each energy system and conduct more targeted regulation. For example, based on the decoupled information, it can be determined whether the main reason for the increase in power load is external factors (such as lower temperature) or internal factors of the system (such as equipment failure), so as to take corresponding measures.

[0028] Step S104, for real-time processing of massive multi-source heterogeneous state information, a distributed stream computing framework is used to build an efficient state information processing pipeline, and the real-time and scalability of state information processing are improved through data sharding, parallel computing and other technologies.

[0029] Obtain massive heterogeneous state information from multiple data sources, preprocess and clean the raw data, filter out invalid, duplicate and abnormal data, and improve data quality. Parse and convert the data formats and protocols of different data sources and unify them into standardized data formats, such as JSON or Avro. Design a reasonable data sharding strategy based on the characteristics and processing requirements of the state information, and divide the state information into multiple data shards, each of which contains a part of the state information. The data sharding strategy can be divided according to dimensions such as the timestamp and device ID of the state information to ensure that the data volume of each shard is balanced. For each data shard, use distributed stream computing frameworks such as Apache Flink for parallel processing. In Flink, build a computing topology for state information processing, define operators such as data source (Source), data transformation (Transformation) and data output (Sink), and specify the data flow and dependency between operators. In the data transformation operator, the state information of each data shard is processed in real time. Through operations such as data conversion and data aggregation, extract key information, calculate relevant indicators, and generate processing results. Common conversion operations include mapping, filtering, and aggregation. The processing results of the distributed stream computing framework are aggregated and merged using message queues such as Kafka to obtain complete state information processing results. The processing results are formatted and stored, and written to the distributed file system HDFS or the database system HBase for subsequent analysis and application. Monitor the running status and performance indicators of the stream computing framework, and use Flink's built-in monitoring and measurement system to collect indicators such as processing delay and throughput of each operator. For operators with high processing delay or heavy load, optimize them by increasing parallelism and adjusting the resource allocation of operators to ensure the real-time and stability of state information processing. Flexible expansion of the computing resources of the stream computing framework according to changes in state information processing requirements and data volume growth. Improve the parallel processing capability of the system by increasing the number of working nodes in the Flink cluster. At the same time, optimize the data sharding strategy, adjust the granularity and distribution of data shards, balance the data volume and processing load of each shard, and improve the scalability and throughput of the system.

[0030] Specifically, in modern data processing scenarios, faced with massive heterogeneous state information from multiple data sources, the first thing to do is preprocessing and cleaning the data. For example, a large-scale power monitoring system may collect data from thousands of smart meters. These data may contain abnormal values ​​caused by equipment failure or transmission errors. By setting thresholds, these abnormal data can be identified and eliminated, such as the meter suddenly reporting a negative or abnormally high power consumption. In addition, data deduplication is another key step to ensure the accuracy of the analysis, such as duplicate data records at the same time point need to be merged or deleted. Unification of data formats is another important link in processing heterogeneous data. Suppose that the data obtained from different meters has different formats, some in XML format and some in CSV format. By writing a conversion script, all data is converted to JSON format. The flexibility and easy parsing characteristics of this format make subsequent processing more efficient. In the design of data sharding strategy, data can be divided according to timestamps and device IDs. For example, data within a day can be sharded by hour, and each shard contains data from all devices within that hour. This approach not only helps to balance the amount of data in each shard, but also optimizes the subsequent parallel processing performance. When using Apache Flink for real-time data stream processing, the built computing topology includes multiple components such as data source, transformation, and output. During the data transformation process, for example, a timestamp field can be added to each data through the Map operation, and the total power consumption per hour can be calculated through the Aggregate operation. The results of these real-time calculations can be summarized through Kafka and then stored in HDFS or HBase to provide support for subsequent data analysis and decision-making. Monitoring the running status of the stream computing framework is the key to ensuring system stability. Through Flink's monitoring system, the processing delay and throughput of each operator can be collected and analyzed in real time. If the delay of an operator is found to be abnormally high, its resource allocation may need to be adjusted, such as increasing the degree of parallelism, to ensure the efficiency and stability of the entire data processing process. With the increase in data volume and changes in processing requirements, the scalability of the system is also particularly important. By increasing the number of nodes in the Flink cluster, the processing capacity of the system can be improved. At the same time, optimizing the data sharding strategy and adjusting the granularity and distribution of the shards can further improve the processing efficiency and load balancing of the system. The implementation of these steps and strategies not only improves the efficiency and accuracy of data processing, but also ensures that the system can flexibly respond to changing business needs and data scale, thereby supporting complex data analysis and intelligent decision-making.

[0031] Step S105, in response to the high reliability requirement for state information transmission in response to power grid energy disturbance suppression, a data security transmission mechanism based on blockchain is adopted, and the decentralized and tamper-proof characteristics of blockchain are utilized to ensure data integrity and non-repudiation during state information transmission and improve transmission reliability.

[0032] The state information change data caused by the power grid energy disturbance is obtained, and the data is preprocessed, including denoising, normalization and other operations, to improve the data quality. The state information is judged whether it exceeds the allowable range according to the preset threshold. If it exceeds, the blockchain data transmission mechanism is triggered. The decentralized state information transmission network is constructed by blockchain technology. Each node verifies the authenticity and integrity of the state information through the Raft consensus algorithm to ensure that the data is not tampered with. The state information data is encrypted and stored on the chain using the SHA-256 encryption algorithm. The chain structure and hash pointer of the blockchain are used to ensure the timeliness and immutability of the data. In the process of state information transmission, the smart contract is written in Solidity to automatically perform data verification and integrity checks. Data verification rules and integrity requirements are set in the contract. Once the data is found to be tampered or lost, the data recovery mechanism is immediately started to obtain complete data from other nodes. Based on the differential privacy algorithm, the state information fusion calculation involving multiple parties is realized without leaking the original data. Each node locally uses the differential privacy algorithm to add random noise to the original data, and then uploads the noisy data to the blockchain network for fusion calculation to improve the accuracy of data analysis and decision-making. Utilize the non-repudiation characteristics of blockchain to record the data transmission and reception behavior of each node and store the operation log on the blockchain. Once a dispute occurs, the responsibility can be traced based on the log information to improve the reliability of cooperation among all parties. Optimize the throughput and latency of the blockchain network. Introduce the Merkle tree structure to store status information and reduce storage overhead. Use the PBFT consensus algorithm instead of the Raft algorithm to improve consensus efficiency. Improve the real-time transmission efficiency of status information while ensuring security.

[0033] Specifically, in modern power grid management systems, real-time monitoring and processing of power grid energy disturbances are crucial. First, denoising and normalization are necessary steps for data collected from sensors and smart meters. For example, if a certain area in the power grid generates noise due to equipment failure, such as abnormal voltage fluctuations, the denoising algorithm can effectively identify and filter out these abnormal fluctuations, while normalization ensures that data from different devices are compared under the same standard, thereby improving the accuracy of data processing. After data preprocessing, the system will determine whether the status information is normal based on the preset threshold. For example, the normal operating range of voltage is set to 220V±5V. If the voltage at a certain point in time is detected to be beyond this range, the system will regard this as an abnormal event and trigger the blockchain data transmission mechanism. The introduction of this mechanism is to utilize the decentralized and tamper-proof characteristics of blockchain to ensure that the authenticity and integrity of abnormal data are guaranteed. In the blockchain network, each node uses the Raft consensus algorithm to verify the authenticity of the data. The Raft algorithm ensures that the leader node in the network has the latest and correct data status through an election mechanism, and other nodes will be updated based on the leader's data, so that the status information can be synchronized quickly and accurately throughout the network. At the same time, the SHA-256 encryption algorithm is used to encrypt the data to ensure the security of the data during transmission. Smart contracts play the role of automatically performing data verification and integrity checks in this process. By writing specific data verification rules in the contract, such as checking whether the data is within the normal range, and the integrity requirements of the data, such as whether the data packet is complete and has not been truncated or tampered with. Once the smart contract finds a data problem, it will automatically trigger the data recovery mechanism, such as requesting missing or tampered data segments from other nodes in the blockchain network. In addition, the application of differential privacy algorithms allows cross-node data fusion calculations without leaking specific user data. Each node adds random noise to the data locally and then uploads it to the blockchain for centralized calculation. This method not only protects user privacy, but also improves the accuracy and reliability of data processing. Finally, in order to improve the processing efficiency of the blockchain network, the Merkle tree structure can be introduced to optimize data storage, and the more efficient PBFT consensus algorithm may be used to replace the Raft algorithm, which not only ensures the security of the network, but also improves the real-time transmission efficiency of state information. The combined application of these technologies not only improves the efficiency and accuracy of grid status information processing, but also ensures that the system can flexibly respond to changing business needs and data scale, thereby supporting complex data analysis and intelligent decision-making.

[0034] Step S106, in response to the demand for intelligent analysis of state information, a deep learning algorithm is used to build an intelligent analysis model for state information. By training and learning historical state information, the multi-energy flow coupling relationship and evolution law are mined to achieve intelligent analysis and prediction of state information.

[0035] According to the historical state information, data cleaning techniques such as removing missing values ​​and outliers are first used to preprocess the data and standardize it, scaling the data to a uniform range to obtain a standardized state information data set. For the standardized state information data set, feature engineering methods such as statistical features, time domain features, and frequency domain features are used to extract key features that can reflect the state characteristics. Then feature selection techniques such as filtering and wrapping are used to select the most discriminative and relevant feature subsets. The processed data set is divided into training sets and test sets according to a certain ratio, such as 8:2. The training set is used for model training, and the test set is used for model evaluation. According to the dimension of the optimal feature subset, the input layer of the convolutional neural network is designed. Then, high-level features of the data are extracted by stacking multiple convolutional layers and pooling layers. The convolutional layer uses a 3x3 or 5x5 convolution kernel, and the activation function uses ReLU. The pooling layer uses maximum pooling to reduce the size of the feature map. Finally, the fully connected layer and the Softmax layer are used to achieve classification prediction of the state. When training the model, the cross entropy loss function is used, and the Adam optimization algorithm is used for optimization. Through grid search and other methods, hyperparameters such as learning rate, batch size, and number of iterations are tuned to improve the performance of the model. For new input data, preprocessing and feature extraction are first performed in the same way as the training data, and then input into the trained model to obtain the state prediction results. Using dimensionality reduction algorithms such as t-SNE, high-dimensional state features are mapped to a two-dimensional plane, and the distribution and relationship between different states are intuitively displayed through scatter plots. Using methods such as Markov chain, the transition probability between states is analyzed, and a state transition map is constructed to reveal the law of state evolution. Using visualization libraries such as Matplotlib and Seaborn, line charts, bar charts, heat maps, etc. are generated to show the trend, distribution, and correlation of state changes. An interactive interface is designed to allow users to explore and analyze state data through operations such as filtering, zooming, and hovering, providing intuitive decision support.

[0036] Specifically, the high reliability demand for state information transmission in power grid energy disturbance suppression has prompted the adoption of more advanced data processing and machine learning technologies. In order to ensure the accuracy and reliability of state information, the original data needs to be cleaned first. For example, there may be missing values ​​or outliers in the voltage data collected by the sensor, such as a reading of -1000V due to sensor failure, which is obviously unreasonable. Missing values ​​can be supplemented by interpolation, and outliers can be eliminated using the 3σ criterion, replacing the abnormal -1000V with the voltage value at the adjacent moment. The cleaned data is more authentic and reliable, laying the foundation for subsequent analysis. Next, the cleaned data needs to be standardized. For example, the unit of voltage is volt and the unit of current is ampere, and the numerical range of the two is very different. In order to eliminate the influence of dimension, the voltage and current data can be scaled to between 0 and 1, so that data with different features can be compared at the same scale, avoiding some features from having too much impact on model training due to large values. After obtaining the normalized state information data set, key features need to be extracted. For example, statistical features such as the mean and variance of voltage can be calculated to reflect the overall level and fluctuation of voltage. It is also possible to extract time domain features such as voltage zero-crossing rate and frequency to reflect the speed and periodicity of voltage change. In addition, it is also possible to extract voltage spectrum features to reflect the energy distribution of different frequency components in the voltage. These features can more comprehensively describe the state of the power grid and provide richer information for subsequent model training. Not all extracted features are helpful for state prediction. Therefore, feature selection is required. For example, the filtering method can be used to calculate the correlation coefficient between each feature and the target variable and select features with higher correlation. Or the wrapping method can be used to repeatedly train the model to select the feature subset that has the greatest improvement in model performance. Assume that after feature selection, the three features of voltage mean, variance and zero-crossing rate are finally selected as the optimal feature subset. Dividing the data set into training set and test set is also a very important step. For example, 80% of the data can be used as a training set to train the model; the remaining 20% ​​of the data can be used as a test set to evaluate the performance of the model. Such a division can effectively avoid overfitting of the model and improve the generalization ability of the model. According to the selected feature subset, the structure of the convolutional neural network can be designed. For example, if three features are selected, the input layer dimension of the convolutional neural network is 3. The convolution layer can use a 3x3 convolution kernel, and the ReLU activation function can effectively avoid the gradient vanishing problem. The pooling layer can use maximum pooling to reduce the size of the feature map and reduce the amount of calculation. The fully connected layer integrates the extracted high-level features and finally outputs the state prediction results through the Softmax layer. For example, the power grid state is divided into three states: normal, warning, and fault. During the model training process, the cross entropy loss function can be used to measure the gap between the model prediction results and the true label. The Adam optimization algorithm can adaptively adjust the learning rate to speed up the model convergence.Through methods such as grid search, the optimal hyperparameters such as learning rate, batch size and number of iterations can be found to further improve the performance of the model. For new input data, it is necessary to process it according to the same preprocessing and feature extraction steps as the training data, and then input it into the trained model to obtain the state prediction results. For example, if the new input data is the voltage and current values ​​at a certain moment, it is necessary to first clean and standardize these data, and then extract features such as voltage mean, variance and zero-crossing rate, and finally input these features into the trained model to obtain the prediction results of the power grid state. In order to more intuitively display the relationship between states, dimensionality reduction algorithms such as t-SNE can be used to map high-dimensional features to a two-dimensional plane. For example, sample points of different states can be represented by different colors to observe their distribution on the two-dimensional plane. If the sample points of different states can be clearly separated, it means that the classification effect of the model is good. Markov chain can be used to analyze the transition probability between states. For example, the probability of the power grid transitioning from the normal state to the warning state and the probability of the power grid transitioning from the warning state to the fault state can be calculated. These probabilities can help better understand the evolution law of the power grid state and take preventive measures in advance. Visualization libraries such as Matplotlib and Seaborn can be used to generate various charts. For example, a line chart can be used to show the trend of voltage over time, a bar chart can be used to show the number of samples in different states, and a heat map can be used to show the correlation between features. The interactive interface allows users to explore and analyze data more conveniently. For example, users can click on a sample point to view its detailed information, or zoom in and out the chart by dragging the mouse. These visualization methods can provide more intuitive support for decision-making.

[0037] Step S107, finally, the data processing results of each link are integrated and associated to form a global electric, heat and gas status information view. Through visualization technology, the real-time operation status of the multi-energy flow system is intuitively presented to provide information support and decision-making basis for power grid energy disturbance suppression.

[0038] The real-time operation data of the power grid, heat network and gas network are obtained, and the operation status of each energy network is fully perceived and uniformly modeled to form the global state information of the multi-energy flow system. The multi-source heterogeneous data obtained are preprocessed, and the data quality is improved through data cleaning, normalization and other operations, and key characteristic parameters are extracted. For the real-time state data of the multi-energy flow system after preprocessing, data fusion technology such as Kalman filtering is used to perform correlation analysis, explore the internal connection and mutual influence between different energy networks, and realize the integrated representation of multi-source heterogeneous data. The historical operation data of the multi-energy flow system is trained through the long short-term memory (LSTM) neural network, and an energy disturbance prediction model is established to judge the energy disturbance risk of the power grid in real time. When a large energy disturbance is predicted, the warning information is output in time. According to the global state information of the multi-energy flow system and the energy disturbance prediction results, the particle swarm optimization algorithm is used to automatically generate the power grid energy disturbance suppression strategy, and the strategy decision information is fed back to the power grid dispatching control system. The JavaScriptDjs library is used to intuitively present the real-time operating status, energy disturbance prediction and suppression strategy of the electric, thermal and gas multi-energy flow system, and a global multi-energy flow monitoring view is formed through dynamic interactive charts to provide auxiliary decision support for power grid dispatchers. The historical operating data of the multi-energy flow system is continuously analyzed through the association rule mining algorithm, the characteristic patterns and response experience of the power grid energy disturbance are summarized, and the LSTM model hyperparameters and particle swarm algorithm parameters are dynamically adjusted to continuously improve the system's adaptive ability. A distributed data management platform based on Hadoop and HBase is built to realize the distributed storage, MapReduce parallel processing and API sharing application of multi-source data such as electric, thermal and gas, and provide efficient and scalable data infrastructure support for power grid energy disturbance suppression.

[0039] Specifically, obtaining real-time operating data of the power grid, heat network, and gas network is the first step in managing a multi-energy flow system. For example, in a city's energy management center, through the intelligent sensors installed at different nodes, the voltage and current data of the power grid, the temperature and flow data of the heat network, and the pressure and flow data of the gas network can be monitored in real time. These data are transmitted back to the data center through a high-speed communication network, forming a comprehensive data set of the energy network operation status. For these multi-source heterogeneous data, it is necessary to first perform data cleaning and normalization. Data cleaning includes removing erroneous readings and duplicate data. For example, if a sensor failure causes an abnormal voltage reading at a certain moment, these data need to be identified and processed. Normalization is to convert data of different dimensions to the same scale, such as converting both voltage and temperature data to the range of 0 to 1, which is helpful for subsequent data analysis and model training. After data preprocessing, data fusion technologies such as Kalman filtering are used to perform correlation analysis on the data, which can effectively explore the mutual influence between different energy networks. For example, a sudden increase in the load of the power grid may affect the operating efficiency of the heat network. Through Kalman filtering, the extent of this impact can be estimated in real time and the future state can be predicted. By training historical data with a long short-term memory (LSTM) neural network, an energy disturbance prediction model can be established. This model can learn the time series characteristics of grid load changes, so as to judge and warn possible energy disturbance risks in real time during actual operation. For example, if the model predicts that there will be a large-scale load increase in the grid within the next hour, the system can adjust the power generation plan in advance or start the backup energy system. Based on the prediction results, the particle swarm optimization algorithm is used to automatically generate the grid energy disturbance suppression strategy. The particle swarm optimization algorithm simulates the foraging behavior of bird flocks to find the optimal solution, and can quickly find the best energy allocation strategy among multiple possible scheduling schemes. These strategies can then be implemented through the grid dispatch system to ensure the stable operation of the grid. Through the Djs library of JavaScript, the real-time operating status, energy disturbance prediction and suppression strategy of the electric, hot and gas multi-energy flow system can be intuitively presented. Dynamic interactive charts can not only display the current energy flow situation, but also simulate the future energy distribution situation, providing dispatchers with intuitive decision support tools. In addition, through continuous association rule mining, patterns in historical data can be analyzed to summarize the characteristics and response strategies of grid energy disturbances. This information can be used to dynamically adjust the parameters of the LSTM model and particle swarm algorithm to improve the adaptability and prediction accuracy of the system. Finally, a distributed data management platform based on Hadoop and HBase can be built to achieve efficient processing and storage of multi-source data such as electricity, heat and gas. Using MapReduce parallel processing technology, large-scale data sets can be processed quickly, ensuring the response speed of real-time data analysis and providing strong data support for power grid energy disturbance suppression.

[0040] The preferred embodiments of the present invention disclosed above are only used to help explain the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to only specific implementation methods. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present invention, so that those skilled in the art can understand and use the present invention well. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A method for suppressing power grid energy disturbances by coordinating the transmission of electric, thermal and gas state information, characterized in that: include: Obtain multi-source heterogeneous state information of electric, thermal and gas energy networks; For the multi-source heterogeneous state information, an adaptive communication protocol conversion technology is used to dynamically convert the heterogeneous communication protocols into a unified protocol format; For the converted state information, a multi-energy flow coupling relationship diagram of the energy network is constructed based on the graph model, and the decoupling and traceability of the state information are achieved through the causal reasoning analysis method; For the decoupled state information, a distributed stream computing framework is used for parallel processing to improve the real-time and scalability of state information processing; The results of stream computing are authenticated and verified using a blockchain-based security consensus mechanism to ensure the credibility of status information. Build an intelligent analysis model of state information based on deep learning algorithms to mine the correlation characteristics and evolution laws contained in multi-energy flow data; The global status information view of the electric, heat and gas energy network is generated through visualization technology to provide auxiliary decision-making for energy network scheduling optimization.

2. The method according to claim 1, characterized in that The method of obtaining multi-source heterogeneous state information of the electric, thermal and gas energy network includes: According to the physical characteristics of the electric, thermal and gas energy network, determine the collection frequency, data format and time scale of each energy system status information; Through data standardization, the status information of heterogeneous energy networks is converted into a unified format; Realize interoperability and integration of status information.

3. The method according to claim 1, characterized in that The adaptive communication protocol conversion technology is used to dynamically convert heterogeneous communication protocols into a unified protocol format, including: Dynamically match and convert various communication protocols according to the differences in communication modes and protocol standards of different regions and devices; Adopt data slicing and parallel transmission technology to improve communication efficiency; Communication security is ensured through lightweight encryption and digital signatures.

4. The method according to claim 1, characterized in that The energy network multi-energy flow coupling relationship diagram is constructed based on the graph model, and the decoupling and traceability of state information is achieved through the causal reasoning analysis method, including: In view of the strong coupling of electric, thermal and gas energy fluctuations, a graphical model is used to characterize the correlation structure of multi-energy flow data; Through graph embedding and graph neural network, we learn the implicit dependencies of multi-energy flow data; Combined with the law of conservation of energy, a causal reasoning rule base is constructed to trace the influence chain of state changes of a single energy system.

5. The method according to claim 1, characterized in that The parallel processing using a distributed stream computing framework includes: Build an efficient state information processing pipeline for massive multi-source heterogeneous state information; Improve the real-time and scalability of state information processing through data sharding and parallel computing technology; Adopt the optimization strategy of memory computing and incremental processing to reduce computing overhead.

6. The method according to claim 1, characterized in that The blockchain-based security consensus mechanism is used for rights confirmation and verification, including: Divide the stream calculation results into blocks and generate hash fingerprints and timestamps; Through consensus algorithms and smart contracts, nodes in the blockchain network can be collaboratively verified; The verified status information is stored on the chain to ensure that the data cannot be tampered with and is traceable.

7. The method according to claim 1, characterized in that The state information intelligent analysis model based on the deep learning algorithm is constructed, including: For multi-source heterogeneous state information, transfer learning method is used to achieve cross-domain knowledge fusion; Through the spatiotemporal series prediction model, the dynamic change rules of state information are explored; Combined with the reinforcement learning framework, the scheduling strategy of the energy network is adaptively optimized.

8. The method according to claim 1, characterized in that The method of generating a global state information view of the electric, heat and gas energy network by using visualization technology includes: Design multi-level visualization views for multi-dimensional and multi-granular status information; Support dynamic query and correlation analysis of status information through human-computer interaction interface; Combine virtual reality and augmented reality technologies to build immersive visualization scenes and improve the intuitiveness and understandability of system operations.

9. The method according to claim 3, characterized in that The dynamic matching and conversion of various communication protocols include: For the received electric and heating gas status information, extract its communication protocol type and parameters, match it with the protocol in the communication protocol knowledge base, and determine whether protocol conversion is required; If protocol conversion is required, the Scapy library in Python is used to dynamically generate a protocol conversion script based on the conversion rules in the communication protocol knowledge base to convert the format and parameters of the original communication protocol.

10. The method according to claim 3, characterized in that: The use of data slicing and parallel transmission technology to improve communication efficiency means that in the process of communication resource coordination, a dynamic resource allocation algorithm based on device priority is used to dynamically schedule and allocate MQTT topics and QoS communication resources using Python's Schedule library according to the communication mode, data volume and preset device priority of devices in different regions to ensure the efficiency and reliability of data transmission; In response to the real-time requirements of electric, heating and gas status information, the retention flag and QoS level of the MQTT message are set, and the CoAP protocol is used to asynchronously report the status change data.

Images