Hydrogen Energy Management System and Method Based on Blockchain and Neural Networks

Abstract

A hydrogen energy management system and method based on blockchain and neural networks is proposed. The system includes a wind turbine, a first edge server, a photovoltaic system, a second edge server, an inverter, a rectifier, a fuel cell, an electrolyzer, a hydrogen storage tank, a third edge server, a blockchain composed of multiple interconnected nodes, and a controller. A convolutional neural network is deployed on one node of the blockchain to diagnose faults in the system's equipment. By judging the power generation status of the wind turbine and photovoltaic system, hydrogen energy is used to supplement insufficient power generation, and excess electrical energy is converted into hydrogen energy for storage when power generation is sufficient. During system operation, the convolutional neural network analyzes the equipment's operating data to diagnose the location of faults. When a fault occurs, the system traces it back through data recorded on the blockchain. This invention achieves clear operator permissions and responsibilities, queryable system operation logs, and decentralized data, which can greatly improve system operating efficiency and fault diagnosis accuracy.

Description

Technical Field

[0001] This invention relates to the field of hydrogen energy storage, and in particular to a hydrogen energy management system and method based on blockchain and neural networks. Background Technology

[0002] Building a clean, low-carbon, and efficient energy system is an effective means for my country to achieve carbon peaking and carbon neutrality. The main energy sources in common integrated energy distribution parks are wind, solar, thermal, energy storage, and the power grid. Wind power and solar power are significantly affected by weather, resulting in highly unstable power output. To maintain stable system operation and reduce wind and solar curtailment rates, introducing hydrogen energy storage into integrated energy systems is a common solution.

[0003] Current integrated energy management systems employ a multi-center server approach to manage data. However, this centralized data management poses security risks and hinders timely assessment of equipment malfunctions. For instance, Chinese invention patent application CN114338684A, which uses a multi-center server method to distribute traffic load across multiple operating units, addresses energy management issues. However, improvements are needed in data security and fault diagnosis. Furthermore, the access control for system participants is unclear. Blockchain, with its fault-tolerant distributed database and multi-node network, offers decentralization, traceability, data and transaction security and transparency, and the ability to share accident or fault information among multiple nodes, making it suitable for integrated energy management systems. Achieving accurate and transparent data exchange and timely fault diagnosis are key challenges in hydrogen energy management. Summary of the Invention

[0004] The purpose of this invention is to overcome the above-mentioned shortcomings of the prior art and provide a hydrogen energy management system and method based on blockchain and neural networks.

[0005] The technical solution of the present invention is: a hydrogen energy management system based on blockchain and neural networks, including a wind turbine, a first edge server, a photovoltaic system, a second edge server, an inverter, a rectifier, a fuel cell, an electrolyzer, a hydrogen storage tank, a third edge server, a blockchain composed of multiple interconnected nodes, and a controller.

[0006] The first edge server, the second edge server, and the third edge server are all composed of sensors and intelligent modules.

[0007] The blockchain consists of a data acquisition layer, a transmission layer, a data layer, a service layer, and an application layer. The data acquisition layer receives parameter data collected from the first, second, and third edge servers. The transmission layer improves the network consensus speed and network security. The data layer records operational data, data summaries, and timestamps, and organizes the data received by the data acquisition layer. The service layer is the core of the entire blockchain, maintaining the operation of the entire blockchain system. The application layer, based on the blockchain network, displays the complete hydrogen energy management system platform through the API interfaces provided by the underlying blockchain.

[0008] The controller is electrically connected to the fuel cell, the electrolyzer, and the blockchain, respectively. The controller controls the operating status of the electrolyzer and records its operating parameters; the controller also controls the operating status of the fuel cell and records its operating parameters.

[0009] The sensors of the first edge server are electrically connected to the wind turbine, collecting operating parameters such as wind speed, rotational speed, and output voltage and current values. The sensors of the second edge server are electrically connected to the photovoltaic system, collecting operating parameters such as photovoltaic panel temperature and photovoltaic output voltage and current values. Both the wind turbine and the photovoltaic system are connected to one end of an AC bus, and the other end of the AC bus is connected to an inverter and a rectifier, respectively. The other end of the rectifier is connected to an electrolyzer, and the other end of the electrolyzer is connected to a hydrogen storage tank. The other end of the inverter is connected to a fuel cell, and the other end of the fuel cell is connected to the hydrogen storage tank. The sensors of the third edge server are electrically connected to the hydrogen storage tank, collecting operating parameters such as pressure, temperature, and hydrogen storage capacity. The sensors of the first, second, and third edge servers respectively send the collected data to the data acquisition layer of the blockchain node through the intelligent module. The controller uploads the recorded operating parameters of the fuel cell and electrolyzer to the blockchain node and controls the operating status of the electrolyzer and fuel cell according to the power provided by the system and the load power.

[0010] A further technical solution of the present invention is as follows: the transmission layer includes the Gossip transmission protocol based on Grpc, the Ethernet transmission protocol, TCP / IC and WiFi protocols; the data layer stores data in the distributed ledger of the blockchain network; the service layer includes smart contracts, encryption algorithms, consensus algorithms, key management and permission management; and the application layer includes an operating platform for park administrators, maintenance personnel and operators.

[0011] A further technical solution of the present invention is: the smart contract includes a personnel permission module, which is used to determine the identity of system operators and assign different permissions; the system operators include administrators, maintainers and operators.

[0012] A further technical solution of the present invention is that the operating parameters of the electrolytic cell include the electrolytic cell voltage, the electrolytic cell current, the rate of hydrogen production, and the rectifier power; the operating parameters of the fuel cell include the fuel cell voltage, the fuel cell current, the rate of hydrogen consumption, the remaining power in the fuel cell, and the inverter power.

[0013] Another technical solution of the present invention is: a method for a hydrogen energy management system based on blockchain and neural networks, as described above, comprising the following steps,

[0014] A. The administrator registers their administrator identity on the blockchain node through a smart contract, then adds maintainers and operators to the system and grants them the corresponding permissions.

[0015] B. Start the wind turbines and photovoltaics to transmit the generated electricity to the AC bus, which supplies the load in the park. At the same time, the first and second edge servers monitor the real-time operating parameters of the wind turbines and photovoltaics, respectively, and upload the real-time operating parameters to the blockchain node. The third edge server monitors the working information of the hydrogen storage tank and uploads the working information to the blockchain node.

[0016] C. After the data is uploaded to the blockchain network, the computer nodes on the blockchain node network are used to organize the data so that it can be stored and retrieved on the blockchain.

[0017] D. The controller obtains the current power generation of the wind turbine and photovoltaic through the blockchain node and compares it with the current load power of the system. When the power generation is greater than the load power, the controller controls the electrolyzer to work and convert the excess electrical energy into hydrogen energy to be stored in the hydrogen storage tank. When the power generation is less than the load power, the controller controls the fuel cell to work and the hydrogen storage tank to release hydrogen, converting hydrogen energy into electrical energy, which is finally input into the AC bus to supply the load.

[0018] A further technical solution of the present invention is as follows: In step C, the blockchain node that deploys the improved convolutional neural network performs calculations based on the collected data to diagnose whether the system has malfunctioned. If a malfunction occurs, the malfunction is promptly notified to the administrator through the blockchain node.

[0019] A further technical solution of the present invention is as follows: the process of deploying blockchain nodes with improved convolutional neural networks to diagnose faults in the hydrogen energy management system is as follows:

[0020] (1) Establish training data samples: Use data augmentation to extract data stored in the blockchain, increase training samples, and obtain multiple training data samples, thereby improving the generalization ability of the network model.

[0021] (2) Constructing the feature transfer module: Load the AlexNet pre-trained network, freeze the first half of the network, fine-tune the second half of the network, and complete the construction of the feature transfer module.

[0022] (3) Split the datasets in multiple training data samples into an 85:15 ratio, with 85% of the data used for training and 15% used for testing the training results; at the same time, add corresponding labels to the training data and test data randomly to shuffle the data training set.

[0023] (4) The training data is put into the AlexNet network for transfer learning, and the parameters are optimized using the Bayesian function to obtain different fault models.

[0024] (5) Use the remaining 15% of the test data to test the obtained fault model, identify the actual health status of the system, and transmit the identification results to the application layer of the blockchain for post-processing.

[0025] A further technical solution of the present invention is as follows: In step C, during hydrogen production and release, the controller compares the amount of hydrogen in the hydrogen storage tank on the blockchain with the amount of hydrogen stored through the third edge server to determine whether external hydrogen addition is required.

[0026] A further technical solution of the present invention is as follows: the first edge server, the second edge server, and the third edge server collect and upload data once every ten minutes; the data regularization is achieved by removing some redundant data through smart contracts in the blockchain nodes, and organizing the data format into the unified format requirements of the blockchain.

[0027] A further technical solution of the present invention includes step E: the administrator assigns tasks to the members and operators in a visual interface, and queries the log information recorded in the blockchain through the visual interface to realize the allocation of system responsibilities and the traceability of working conditions.

[0028] Compared with the prior art, the present invention has the following characteristics:

[0029] (1) This invention uses blockchain technology to build a hydrogen energy management system, which realizes the purpose of automatic data collection and ensures data storage security, avoids the cost of maintaining the central node, and effectively avoids irreparable economic losses caused by the attack on the central server; the data encryption and digital signature of the blockchain ensure the immutability of data transmission, and explores a new path for the low-carbon development of society.

[0030] (2) By deploying a convolutional neural network on a blockchain node, the present invention can accurately diagnose system faults when judging system operation faults, ensuring normal system operation and reducing economic losses caused by faults.

[0031] (3) The present invention sets the permissions of administrators, maintenance personnel and operators through smart contracts. The new smart contract design makes the rule judgment simpler and clearer, and improves the efficiency of blockchain operation.

[0032] The detailed structure of the present invention will be further described below with reference to the accompanying drawings and specific embodiments. Attached Figure Description

[0033] Appendix Figure 1 This is a schematic diagram of the system structure of the present invention;

[0034] Appendix Figure 2 This is a schematic diagram of the system data flow of the present invention;

[0035] Appendix Figure 3 This is a schematic diagram of the blockchain system architecture of the present invention;

[0036] Appendix Figure 4 This is a flowchart of the system management method of the present invention;

[0037] Appendix Figure 5 The graphs show the accuracy of fault diagnosis using SVM, KNN, and the present invention, respectively. Detailed Implementation

[0038] Example 1, as shown in the appendix Figure 1-4 As shown, the hydrogen energy management system based on blockchain and neural networks includes a wind turbine, a first edge server, a photovoltaic system, a second edge server, an inverter, a rectifier, a fuel cell, an electrolyzer, a hydrogen storage tank, a third edge server, a blockchain consisting of multiple interconnected nodes, and a controller.

[0039] The first, second, and third edge servers are all composed of sensors and intelligent modules. The first edge server is electrically connected to the wind turbine to collect its operating parameters; the second edge server is electrically connected to the photovoltaic system to collect its operating parameters; and the third edge server is electrically connected to the hydrogen storage tank to collect its operating parameters.

[0040] The blockchain comprises a data acquisition layer, a transmission layer, a data layer, a service layer, and an application layer. The data acquisition layer receives parameter data collected from the first, second, and third edge servers. The transmission layer includes the GRPC-based Gossip transmission protocol, Ethernet transmission protocol, TCP / IC, and WiFi protocol to improve network consensus speed and network security. The data layer stores data in the distributed ledger of the blockchain network, recording operational data, data summaries, and timestamps. It also organizes the data received by the data acquisition layer; duplicate data collection prevents it from being recorded on the blockchain, avoiding wasted blockchain space. The service layer is the core of the entire blockchain, including smart contracts, encryption algorithms, consensus algorithms, key management, and access control, maintaining the operation of the entire blockchain system. The application layer includes an operating platform for park administrators, maintenance personnel, and operators. Built upon the existing blockchain network, it utilizes the API interfaces provided by the underlying blockchain to present a complete hydrogen energy management system platform, fulfilling the system's business requirements.

[0041] The smart contract includes a personnel permission module for identifying system operators and assigning them different permissions. System operators include administrators, maintainers, and operators. Administrators have the highest permissions within the system, enabling them to monitor and query the operational status information of all devices or components within the system. Maintainers and operators can only query information relevant to themselves; viewing data beyond their authorized scope requires a special request to the administrator. Once an operator's identity is confirmed as an administrator, that operator is granted the right to query all log information. The administrator can also assign tasks to maintainers and operators and approve special requests. This permission module clearly defines the responsibilities of different personnel, ensuring data security.

[0042] The controller is electrically connected to the fuel cell, the electrolyzer, and the blockchain, respectively. The controller controls the operating status of the electrolyzer and records its operating parameters, including the electrolyzer voltage, electrolyzer current, hydrogen production rate, rectifier power, etc.; the controller controls the operating status of the fuel cell and records its operating parameters, including the fuel cell voltage, fuel cell current, hydrogen consumption rate, remaining energy in the fuel cell, inverter power, etc.

[0043] The sensors of the first edge server are electrically connected to the wind turbine, collecting operating parameters such as wind speed, rotational speed, and output voltage and current values. The sensors of the second edge server are electrically connected to the photovoltaic system, collecting operating parameters such as photovoltaic panel temperature and photovoltaic output voltage and current values. Both the wind turbine and the photovoltaic system are connected to one end of an AC bus, and the other end of the AC bus is connected to an inverter and a rectifier, respectively. The other end of the rectifier is connected to an electrolyzer, and the other end of the electrolyzer is connected to a hydrogen storage tank. The other end of the inverter is connected to a fuel cell, and the other end of the fuel cell is connected to the hydrogen storage tank. The sensors of the third edge server are electrically connected to the hydrogen storage tank, collecting operating parameters such as pressure, temperature, and hydrogen storage capacity. The sensors of the first, second, and third edge servers respectively send the collected data to the data acquisition layer of the blockchain node through the intelligent module. The controller uploads the recorded operating parameters of the fuel cell and electrolyzer to the blockchain node and controls the operating status of the electrolyzer and fuel cell according to the power provided by the system and the load power.

[0044] A convolutional neural network (CNN) is deployed on one of the nodes of the blockchain to diagnose device faults. The CNN includes an input layer, convolutional layers, activation functions, pooling layers, and fully connected layers. When using the CNN to analyze and process data, it needs to be trained first; the trained network is then used for fault diagnosis in the hydrogen energy management system.

[0045] Since the working parameter data collected by the first, second, and third edge servers are all one-dimensional time-series signals and do not involve two-dimensional images, only one-dimensional convolution needs to be performed, which saves computing resources and time and is beneficial for real-time monitoring of the equipment. However, in actual system operation, fault data is very scarce compared to normal data, resulting in insufficient sample data and low generalization performance of the fault diagnosis model trained by the neural network. To solve this problem, data augmentation is used to process the collected working parameter data. Specifically, the data of different health states of the equipment stored in the blockchain are extracted by overlapping. The working parameter data collected by the first, second, and third edge servers every 10 minutes are recorded as a group, which generates 24 groups of data in four hours. These 24 groups of data are used as the first training set. After training, the 6 groups of data from the first hour are discarded, and the training data from the fifth hour is added. The data from the second, third, fourth, and fifth hours are used as the second training set for training. That is, the sample length of the overlapping extraction is 3 hours of data.

[0046] Example 2, a method for a hydrogen energy management system based on the aforementioned blockchain and neural network, includes the following steps:

[0047] A. The administrator registers their administrator identity on the blockchain node through a smart contract, then adds maintainers and operators to the system and grants them corresponding permissions. Under their respective permissions, the administrator, maintainers, and operators perform corresponding operations based on the fault identification results received by the blockchain's application layer.

[0048] B. Start the wind turbines and photovoltaics to transmit the generated electricity to the AC bus, which supplies the load in the park. At the same time, the first and second edge servers monitor the real-time operating parameters of the wind turbines and photovoltaics, and upload the real-time operating parameters to the blockchain node via WiFi. The third edge server monitors the working information such as the pressure and hydrogen storage capacity of the hydrogen storage tank, and uploads the working information to the blockchain node via WiFi.

[0049] The first edge server, the second edge server, and the third edge server collect and upload data once every ten minutes.

[0050] C. After the data is uploaded to the blockchain network, the computer nodes on the blockchain node network are used to standardize the data format so that it can be stored and retrieved on the blockchain.

[0051] Data standardization involves the smart contracts in the blockchain nodes standardizing the received data, removing redundant data, and organizing the data into a format that meets the unified requirements of the blockchain.

[0052] D. The controller obtains the current power generation of the wind turbine and photovoltaic system through blockchain nodes and compares it with the current system load power. When the power generation is greater than the load power, the controller controls the electrolyzer to operate, converting the excess electrical energy into hydrogen energy and storing it in the hydrogen storage tank. When the power generation is less than the load power, the controller controls the fuel cell to operate and the hydrogen storage tank to release hydrogen, converting the hydrogen energy into electrical energy, which is ultimately input into the AC bus to supply the load.

[0053] In step C, to understand the working status of the hydrogen storage tank and ensure system safety, during hydrogen production and release, the controller compares the amount of hydrogen in the hydrogen storage tank on the blockchain with the amount of hydrogen stored through the third edge server to determine whether external hydrogen addition or other operations are required.

[0054] In step C, in order to make a clear judgment on the operating status of the system, the blockchain node with the improved convolutional neural network is deployed to perform calculations based on the collected data to diagnose whether the system has failed. If a failure occurs, the failure will be promptly notified to the administrator through the blockchain node.

[0055] The process of using convolutional neural networks to diagnose faults in the hydrogen energy management system is as follows:

[0056] (1) Establish training data samples: Use data augmentation to extract data stored in the blockchain, increase training samples, and obtain multiple training data samples, thereby improving the generalization ability of the network model.

[0057] (2) Constructing the feature transfer module: Load the AlexNet pre-trained network, freeze the first half of the network, fine-tune the second half of the network, and complete the construction of the feature transfer module.

[0058] (3) Split the datasets in multiple training data samples into an 85:15 ratio, with 85% of the data used for training and 15% used for testing the training results; at the same time, add corresponding labels to the training data and test data randomly to shuffle the data training set.

[0059] (4) The training data is put into the AlexNet network for transfer learning, and the parameters are optimized using the Bayesian function to obtain different fault models.

[0060] (5) Use the remaining 15% of the test data to test the obtained fault model, identify the actual health status of the system, and transmit the identification results to the application layer of the blockchain for post-processing.

[0061] In practical applications, to further enable the query of operating status information of all devices or components within the system, as well as the supervision of system operators, step E is also included: the administrator assigns tasks to maintenance personnel and operators through the visual interface, and queries the log information recorded in the blockchain through the visual interface to realize the allocation of system responsibilities and the traceability of operating status.

[0062] To verify the superiority of this system in judging system operating status and fault status, the data collected during system operation were used to judge system faults using the aforementioned hydrogen energy management method based on blockchain and neural networks, as well as the traditional SVM and KNN methods. The accuracy rates of system fault judgment obtained by the three methods are shown in the appendix. Figure 5 As shown, from the appendix Figure 5 As can be seen, the hydrogen energy management method based on blockchain and neural networks adopted in this system can achieve normal operation of both single and multiple devices. Furthermore, when detecting a single device, the accuracy of all three methods can reach over 90%; when detecting all devices within the system, the accuracy of traditional SVM and KNN drops to 60%, while the diagnostic method based on blockchain and neural networks still maintains an accuracy of 85%, thus improving the system's accuracy and the efficiency of blockchain operation.

Claims

1. A hydrogen energy management system based on blockchain and neural networks, characterized by: It includes wind turbines, first edge servers, photovoltaics, second edge servers, inverters, rectifiers, fuel cells, electrolyzers, hydrogen storage tanks, third edge servers, a blockchain consisting of multiple interconnected nodes, and a controller; The first edge server, the second edge server, and the third edge server are all composed of sensors and intelligent modules; The blockchain consists of a data acquisition layer, a transmission layer, a data layer, a service layer, and an application layer; the data acquisition layer receives parameter data collected by the first edge server, the second edge server, and the third edge server; The transport layer is used to improve network consensus speed and network security; The data layer is used to record runtime data, data summaries, timestamps, and to organize the data received by the data acquisition layer; The service layer is the core of the entire blockchain, used to maintain the operation of the entire blockchain system; the application layer is used to display the complete hydrogen energy management system platform on the basis of the blockchain network through the API interface provided by the underlying blockchain. The controller is electrically connected to the fuel cell, the electrolyzer, and the blockchain respectively; the controller controls the working status of the electrolyzer and records its working parameters; the controller controls the working status of the fuel cell and records its working parameters. The sensors of the first edge server are electrically connected to the wind turbine, collecting the wind speed, rotational speed, and output voltage and current values ​​of the wind turbine. The sensors of the second edge server are electrically connected to the photovoltaic system, collecting the photovoltaic panel temperature and output voltage and current values ​​of the photovoltaic system. Both the wind turbine and the photovoltaic system are connected to one end of the AC bus. The other end of the AC bus is connected to the inverter and the rectifier, respectively. The other end of the rectifier is connected to the electrolyzer, and the other end of the electrolyzer is connected to the hydrogen storage tank. The other end of the inverter is connected to the fuel cell, and the other end of the fuel cell is connected to the hydrogen storage tank. The sensors of the third edge server are electrically connected to the hydrogen storage tank, collecting the pressure, temperature, and hydrogen storage capacity of the hydrogen storage tank. The sensors of the first, second, and third edge servers respectively send the collected data to the data acquisition layer of the blockchain node through the intelligent module. The controller uploads the recorded operating parameters of the fuel cell and the electrolyzer to the blockchain node and controls the operating status of the electrolyzer and the fuel cell according to the power provided by the system and the load power.

2. The hydrogen energy management system based on blockchain and neural networks as described in claim 1, characterized in that: The transport layer includes the GRPC-based Gossip transport protocol, Ethernet transport protocol, TCP / IC and WiFi protocol; the data layer stores data in the distributed ledger of the blockchain network; the service layer includes smart contracts, encryption algorithms, consensus algorithms, key management and permission management; the application layer includes an operating platform for park administrators, maintenance personnel and operators.

3. The hydrogen energy management system based on blockchain and neural networks as described in claim 2, characterized in that: The smart contract includes a personnel permission module, which is used to identify system operators and assign different permissions; the system operators include administrators, maintainers, and operators.

4. The hydrogen energy management system based on blockchain and neural networks as described in claim 3, characterized in that: The operating parameters of the electrolyzer include the electrolyzer voltage, electrolyzer current, hydrogen production rate, and rectifier power; the operating parameters of the fuel cell include the fuel cell voltage, fuel cell current, hydrogen consumption rate, remaining energy in the fuel cell, and inverter power.

5. The method for a hydrogen energy management system based on blockchain and neural networks as described in any one of claims 1-4, characterized in that: Includes the following steps, A. The administrator registers their administrator identity on the blockchain node through a smart contract, then adds maintainers and operators to the system and grants them the corresponding permissions; B. Start the wind turbines and photovoltaics to transmit the generated electricity to the AC bus, which supplies the load in the park; at the same time, the first edge server and the second edge server monitor the real-time operating parameters of the wind turbines and photovoltaics respectively, and upload the real-time operating parameters to the blockchain node; the third edge server monitors the working information of the hydrogen storage tank and uploads the working information to the blockchain node. C. After the data is uploaded to the blockchain network, the computer nodes on the blockchain node network are used to organize the data so that it can be stored and retrieved on the blockchain. D. The controller obtains the current power generation of the wind turbine and photovoltaic through the blockchain node and compares it with the current load power of the system. When the power generation is greater than the load power, the controller controls the electrolyzer to work and convert the excess electrical energy into hydrogen energy to be stored in the hydrogen storage tank. When the power generation is less than the load power, the controller controls the fuel cell to work and the hydrogen storage tank to release hydrogen, converting hydrogen energy into electrical energy, which is finally input into the AC bus to supply the load.

6. The method for a hydrogen energy management system based on blockchain and neural networks as described in claim 5, characterized in that: in In step C, the blockchain node deploying the improved convolutional neural network performs calculations based on the collected data to diagnose whether the system has malfunctioned. If a malfunction occurs, it will promptly notify the administrator through the blockchain node.

7. The method for a hydrogen energy management system based on blockchain and neural networks as described in claim 6, characterized in that: The process for deploying blockchain nodes with improved convolutional neural networks to diagnose faults in the hydrogen energy management system is as follows: (1) Establish training data samples: Use data augmentation to extract data stored in the blockchain, increase training samples, and obtain multiple training data samples, thereby improving the generalization ability of the network model. (2) Constructing the feature transfer module: Load the AlexNet pre-trained network, freeze the first half of the network, fine-tune the second half of the network, and complete the construction of the feature transfer module; (3) Split the datasets in multiple training data samples into an 85:15 ratio, with 85% of the data used for training and 15% used for testing the training results; at the same time, randomly add corresponding labels to the training data and test data to shuffle the data training set; (4) The training data is put into the AlexNet network for transfer learning, and the parameters are optimized using the Bayesian function to obtain different fault models; (5) Use the remaining 15% of the test data to test the obtained fault model, identify the actual health status of the system, and transmit the identification results to the application layer of the blockchain for post-processing.

8. The method for a hydrogen energy management system based on blockchain and neural networks as described in claim 5, characterized in that: in In step C, during hydrogen production and release, the controller compares the amount of hydrogen in the hydrogen storage tank on the blockchain with the amount of hydrogen stored through the third edge server to determine whether external hydrogen replenishment is required.

9. The method for a hydrogen energy management system based on blockchain and neural networks as described in any one of claims 5, characterized in that: The first edge server, the second edge server, and the third edge server collect and upload data every ten minutes; the data standardization process uses smart contracts in the blockchain nodes to remove redundant data and organize the data into a unified format required by the blockchain.

10. The method for a hydrogen energy management system based on blockchain and neural networks as described in any one of claims 6-9, characterized in that: It also includes step E, where the administrator assigns tasks to maintenance personnel and operators through a visual interface, and queries the log information recorded in the blockchain through the visual interface to realize the allocation of system responsibilities and the traceability of working conditions.

Images