Energy scheduling method of data center microgrid and distributed energy scheduling architecture

Abstract

The invention discloses an energy scheduling method of a data center microgrid and a distributed energy scheduling architecture, and relates to the technical field of block chain energy scheduling. The method is applied to a server of a data center micro-grid in a distributed energy scheduling architecture, and the distributed energy scheduling architecture comprises a plurality of data center micro-grids. When the data center microgrid has a transaction demand, the server uploads a power transaction request to a block chain network; receiving transaction matching party information sent by the block chain network; the transaction matching party information is obtained after the block chain network performs matching according to power transaction requests issued by other data center microgrids on the block chain network; carrying out power transaction with a transaction matching party through an intelligent contract of the block chain network; and after the transaction is completed, generating transaction settlement information, and encrypting and storing the transaction settlement information in the block chain network. According to the method, the data security in the power transaction process can be ensured.

Description

Technical Field

[0001] The present invention relates to the field of blockchain energy scheduling technology, and in particular to an energy scheduling method and a distributed energy scheduling architecture for a data center microgrid. Background Art

[0002] With the development of renewable energy and smart grid technology, distributed energy dispatching systems are becoming more and more popular.

[0003] In the prior art, a distributed energy dispatching system can be composed of multiple data center microgrids (DCMG). In the distributed energy dispatching architecture, data centers can buy and sell electricity based on real-time power demand and production conditions. For example, when the load of a data center microgrid is higher than the renewable energy it produces, it can purchase electricity from other data center microgrids. When a data center microgrid generates more electricity than it needs in a certain period of time, the remaining electricity can be sold to other data center microgrids that need electricity, thereby achieving a reasonable allocation of resources and improving energy utilization efficiency.

[0004] However, in the process of electricity trading, a large amount of sensitive data (such as user identity, transaction amount, energy flow, etc.) is involved. These data may be attacked by hackers or illegally obtained during transmission and storage, resulting in data leakage. Summary of the invention

[0005] Based on this, it is necessary to provide an energy scheduling method and distributed energy scheduling architecture for a data center microgrid to address the above technical problems. This method can ensure the security of data during power trading.

[0006] The present invention adopts the following technical solutions:

[0007] The present invention provides an energy scheduling method for a data center microgrid, which is applied to a server of a data center microgrid in a distributed energy scheduling architecture, wherein the distributed energy scheduling architecture includes multiple data center microgrids; the method includes:

[0008] When there is a transaction demand in the data center microgrid, the server generates a power transaction request and uploads the power transaction request to the blockchain network; the power transaction request includes the transaction amount;

[0009] The server receives the transaction matching party information sent by the blockchain network; the transaction matching party information is obtained after the blockchain network matches the power transaction requests issued by other data center microgrids on the blockchain network;

[0010] The server conducts electricity transactions with the transaction matching party through the smart contract of the blockchain network;

[0011] After the transaction is completed, the server generates transaction settlement information and encrypts and stores the transaction settlement information in the blockchain network.

[0012] Preferably, the method further comprises:

[0013] The server obtains the status data of the data center microgrid, which includes the power generation of renewable energy and the power usage of power-consuming devices;

[0014] The server inputs the status data into the action network to obtain the behavior data of the data center microgrid; the behavior data includes the transaction power;

[0015] If the transaction power is within the preset range, the server determines that there is a transaction demand in the data center microgrid.

[0016] Preferably, generating an electricity transaction request includes:

[0017] Obtain the electricity transaction price and electricity transaction time of the data center microgrid;

[0018] Generate an electricity trading request based on the trading volume, electricity trading price and electricity trading time.

[0019] Preferably, the process of constructing the action network includes:

[0020] Obtain the current state data of the simulated data center microgrid, and interact with the simulated data center microgrid through the initial action network to obtain the behavior data, reward value and next moment state data of the simulated data center microgrid;

[0021] The current state data, behavior data, reward value and next moment state data are stored as data quadruple in the experience replay pool, and a preset number of historical experience data are obtained from the experience replay pool; the historical experience data is a data quadruple;

[0022] Input the experience state data in the historical experience data into the target action network to obtain the first behavior data, input the experience state data and the first behavior data into the target evaluation network to obtain the joint state behavior value; determine the target evaluation value according to the reward value, the discount factor and the joint state behavior value;

[0023] Input the experience state data into the initial action network to obtain the second action data, and input the experience state data and the second action data into the evaluation network to obtain the evaluation value;

[0024] According to the target evaluation value and the evaluation value, the initial action network is updated to obtain the action network.

[0025] Preferably, the initial action network interacts with the simulated data center microgrid to obtain behavior data, reward value and next moment state data of the simulated data center microgrid, including:

[0026] The behavior data of the simulated data center microgrid is determined through the initial action network; the behavior data is used to guide the simulated data center microgrid to trade electricity with the outside world;

[0027] Substitute the current state data and behavior data into the reward function to get the reward value;

[0028] The state of the simulated data center microgrid is transferred through behavioral data to obtain the state data of the simulated data center microgrid at the next moment.

[0029] Preferably, the initial action network is updated according to the target evaluation value and the evaluation value to obtain the action network, including:

[0030] Determine the loss function value according to the target evaluation value and the evaluation value;

[0031] Update the parameters of the target action network and the target evaluation network according to the soft update coefficient and the loss function value;

[0032] The parameters of the initial action network and the evaluation network are updated through the loss function value, and the updated initial action network is determined as the action network.

[0033] Preferably, the power transaction is conducted with the transaction matching party through the smart contract of the blockchain network, including:

[0034] If the data center microgrid is the transaction seller and the transaction matcher is the transaction buyer, power transaction data is generated according to the power transaction request of the transaction buyer;

[0035] The power transaction data is signed by a private key, and the signed power transaction data is encrypted by the public key of the transaction buyer to obtain encrypted power transaction data;

[0036] The encrypted power transaction data is sent to the transaction buyer, so that the transaction buyer can conduct power transactions based on the encrypted power transaction data; the transaction buyer decrypts the encrypted power transaction data based on the private key and verifies the integrity of the power transaction data through the public key of the transaction seller;

[0037] Accordingly, the smart contract includes the identity information of the two parties to the transaction, the amount of electricity traded by both parties, the electricity price and the electricity delivery time window; after the electricity transaction is completed, the method further includes:

[0038] The identity information of the two parties to the transaction, the amount of electricity traded by both parties, the electricity price and the electricity delivery time window are determined as the transaction settlement information.

[0039] The present invention provides a distributed energy scheduling architecture, which includes a plurality of data center microgrids, each of which is electrically connected via a common connection point; each of which includes a server;

[0040] The server is used to upload a power transaction request to the blockchain network when there is a transaction demand in the data center microgrid, receive the transaction matching party information sent by the blockchain network, and conduct power transactions with the transaction matching party through the smart contract of the blockchain network. After the transaction is completed, the transaction settlement information is generated and encrypted and stored in the blockchain network;

[0041] The power transaction request includes the transaction quantity; the transaction matching party information is obtained after the blockchain network matches the power transaction requests issued by other data center microgrids on the blockchain network; the power transaction data during the power transaction process is protected by encryption in the smart contract.

[0042] Preferably, each data center microgrid includes servers, energy supply devices and energy consumption devices; the energy supply devices include photovoltaics, wind turbines, gas turbines and energy storage systems, and the energy consumption devices include carbon capture devices, carbon storage devices, power-to-gas devices, cogeneration devices, heat storage devices, absorption chillers and central air conditioners;

[0043] The data center microgrid captures carbon through a carbon capture device and converts the captured carbon into methane through a power-to-gas device to provide fuel for the cogeneration device. The heat energy generated by the cogeneration device is converted into cold energy through a thermal storage device and an absorption chiller.

[0044] Central air conditioning is used to provide cooling energy to the data center microgrid when the absorption chiller cannot meet the cooling needs of the data center.

[0045] Preferably, the distributed energy dispatching architecture also includes a public power grid and a natural gas grid;

[0046] Utility grid and natural gas grid for trading electricity with multiple data center microgrids.

[0047] The present invention provides an energy dispatching device for a data center microgrid, which is applied to a data center microgrid in a distributed energy dispatching architecture, wherein the distributed energy dispatching architecture includes a plurality of data center microgrids; the device includes:

[0048] The upload module is used to generate a power transaction request when there is a transaction demand in the data center microgrid, and upload the power transaction request to the blockchain network; the power transaction request includes the transaction amount;

[0049] A receiving module is used to receive transaction matching party information sent by the blockchain network; the transaction matching party information is obtained after the blockchain network matches the power transaction requests issued by other data center microgrids on the blockchain network;

[0050] The trading module is used to trade electricity with the trading partner through the smart contract of the blockchain network;

[0051] The storage module is used to encrypt the electricity transaction data and store it in the blockchain network after the transaction is completed.

[0052] The present invention provides a computer-readable storage medium, wherein the storage medium stores a computer program, and when the computer program is executed by a processor, the energy scheduling method of the data center microgrid is implemented.

[0053] The present invention provides a computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the energy scheduling method for the data center microgrid when executing the program.

[0054] At least one of the above technical solutions adopted by the present invention can achieve the following beneficial effects:

[0055] In the present invention, the data center microgrid uploads the power transaction request to the blockchain network, and concludes transactions with other data center microgrids through the smart contract of the blockchain. The transaction process of the smart contract is automatically executed based on pre-defined rules. In this way, the corresponding operations are automatically executed when the conditions are met, which greatly reduces the risk of malicious behavior and improves the security of the data center microgrid during the transaction process; and after the transaction is completed, the power transaction data is encrypted and stored, ensuring the protection of sensitive information, preventing data leakage, and ensuring the security of the power transaction data. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0057] Figure 1 A schematic diagram of the structure of a distributed energy scheduling architecture provided by the present invention;

[0058] Figure 2 A schematic diagram of the structure of another distributed energy scheduling architecture provided by the present invention;

[0059] Figure 3 A schematic flow chart of an energy dispatching method for a data center microgrid provided by the present invention;

[0060] Figure 4 A schematic flow chart of another energy dispatching method for a data center microgrid provided by the present invention;

[0061] Figure 5 A distributed DCMG energy management framework based on blockchain is provided by the present invention;

[0062] Figure 6 A schematic diagram of a training process of an action network provided by the present invention;

[0063] Figure 7 A schematic diagram of an energy dispatching device for a data center microgrid provided by the present invention;

[0064] Figure 8 A schematic diagram of a computer device for implementing an energy scheduling method for a data center microgrid provided by the present invention. DETAILED DESCRIPTION

[0065] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the specific embodiments of the present invention and the corresponding drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0066] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by technicians in the technical field to which this application belongs; the terms used herein are only for the purpose of describing specific embodiments and are not intended to limit this application; the terms "including" and "having" in the specification and claims of this application and the above-mentioned figure descriptions and any variations thereof are intended to cover non-exclusive inclusions.

[0067] In the description of the embodiments of the present application, the technical terms "first", "second", etc. are only used to distinguish different objects, and cannot be understood as indicating or implying relative importance or implicitly indicating the number, specific order or primary and secondary relationship of the indicated technical features. In the description of the embodiments of the present application, the meaning of "multiple" is more than two, unless otherwise clearly and specifically defined.

[0068] In the description of the embodiments of the present application, unless otherwise clearly specified and limited, technical terms such as "connection" should be understood in a broad sense, for example, it can be directly connected or indirectly connected through an intermediate medium, it can be the internal connection of two elements or the interaction relationship between two elements. For ordinary technicians in this field, the specific meanings of the above terms in the embodiments of the present application can be understood according to specific circumstances.

[0069] At present, blockchain, as an emerging distributed ledger technology, provides an innovative solution to the data security and privacy issues of energy scheduling. In the process of energy scheduling, data center microgrid power trading is a core link. Choosing a suitable energy scheduling strategy can effectively reduce the power cost and carbon cost of data center microgrid, and using blockchain and smart contract technology can effectively improve the security and tamper-proof of the energy scheduling process.

[0070] At present, the existing data center microgrid low-carbon distributed energy scheduling mainly adopts traditional optimization algorithms such as ant colony algorithm, Lyapunov optimization technology and alternating direction multiplier (ADMM) algorithm. Among them, the most mainstream method is to model the low-carbon distributed energy scheduling problem of the data center, and use the algorithm to optimize the output of photovoltaic, wind turbine, gas turbine and other components in the data center microgrid, power demand and other information to reduce its power cost and carbon cost. However, due to the uncertainty of components such as photovoltaic and wind turbines, it is impossible to accurately model them. In addition, the existing methods mainly use traditional optimization algorithms or centralized reinforcement learning methods to optimize the energy scheduling of data center microgrids, but they cannot effectively solve the distributed low-carbon energy problem of data center microgrids, resulting in the inability to effectively reduce the power cost and carbon cost of data center microgrids. In addition, the data security, privacy and security and integrity of power transactions in the low-carbon distributed energy scheduling process of data center microgrids cannot be fully guaranteed.

[0071] In order to overcome the above-mentioned technical shortcomings of the prior art, the purpose of the present invention is to provide an energy scheduling method and distributed energy scheduling architecture for a data center microgrid with strong scalability, high security and low cost, and to model the low-carbon distributed energy scheduling problem of a data center microgrid in detail; and to use a distributed reinforcement learning algorithm to optimize and control the energy scheduling process of a data center microgrid to reduce its electricity cost and carbon cost; finally, to establish a low-carbon distributed energy scheduling architecture for a data center microgrid based on blockchain and smart contracts, to improve the security and privacy of data in the low-carbon energy process of a data center microgrid, and to ensure the security and integrity of electricity transactions.

[0072] The energy scheduling method for a data center microgrid provided in the present application is applied in a data center microgrid in a distributed energy scheduling architecture. Based on this, before describing the energy scheduling method for a data center microgrid provided by the present invention, the distributed energy scheduling architecture is first described.

[0073] The technical solutions provided by various embodiments of the present invention are described in detail below in conjunction with the accompanying drawings.

[0074] The present invention provides a distributed energy scheduling architecture, such as Figure 1 As shown, the distributed energy dispatching architecture includes multiple data center microgrids, each of which is connected to electricity through a public connection point; each data center microgrid includes a server.

[0075] The server is used to upload power transaction requests to the blockchain network when there is a transaction demand in the data center microgrid, receive transaction matching party information sent by the blockchain network, and conduct power transactions with the transaction matching party through the smart contract of the blockchain network. After the transaction is completed, transaction settlement information is generated and encrypted and stored in the blockchain network; the power transaction request includes the transaction quantity; the transaction matching party information is obtained after the blockchain network matches the power transaction requests issued by other data center microgrids on the blockchain network.

[0076] Specifically, whether there is a transaction demand for the data center microgrid can be determined based on the difference between the power consumption and power generation of the data center microgrid; for example, when the absolute value of the difference between the power consumption and power generation of the data center microgrid is greater than a preset difference threshold, it is determined that there is a transaction demand for the data center microgrid.

[0077] The transaction demand may include the demand for buying electricity or the demand for selling electricity. When the power consumption is greater than the power generation, the transaction demand of the data center microgrid is determined to be the demand for buying electricity. When the power consumption is less than the power generation, the transaction demand of the data center microgrid is determined to be the demand for selling electricity. The power transaction request may include the amount of electricity to be traded, the price of electricity, and the time of power transaction.

[0078] Each data center microgrid can upload its own power transaction request to the blockchain network, and the blockchain network can match the power transaction requests of both parties; for example, if the amount of electricity, electricity price and transaction time published by seller A are consistent with the needs of buyer B, the blockchain network will automatically match the two parties for power trading. After the power transaction is successfully matched, the blockchain network triggers the smart contract, automatically executes the smart contract and conducts power trading according to the transaction agreement.

[0079] Optionally, transaction settlement can be performed through power transaction settlement information, which includes: Trsc i ={ts,dcg i ,dcg j ,t DC,ij,t ,c DC,i,t}. Among them, ts is the transaction time, dcg i and dcg j are the information of both parties in the power transaction; t DC,ij,t is the electricity trading volume, c DC,i,tRepresents the electricity transaction cost of data center.

[0080] In addition, the blockchain network uses the Proof of Work (PoW) consensus algorithm to ensure that the supply and demand information of all data center microgrids is updated in real time.

[0081] After the transaction is completed, transaction settlement information is generated and encrypted and stored in the blockchain network.

[0082] In an exemplary embodiment, Figure 2 As shown, each data center microgrid includes servers, energy supply devices and energy consumption devices; the energy supply devices include photovoltaics, wind turbines, gas turbines and energy storage systems, and the energy consumption devices include carbon capture devices, carbon storage devices, power-to-gas devices, cogeneration devices, heat storage devices, absorption chillers and central air conditioners.

[0083] The data center microgrid captures carbon through a carbon capture device, and converts the captured carbon into methane through a power-to-gas device to provide fuel for the cogeneration device, and converts the heat energy generated by the cogeneration device into cold energy through a heat storage device and an absorption chiller; the central air conditioner is used to provide cold energy for the data center microgrid when the absorption chiller cannot meet the cooling needs of the data center.

[0084] DCMG uses carbon capture technology to capture carbon, which is then converted into CH4 through power-to-gas to provide fuel for the cogeneration unit, and the heat energy generated by the cogeneration unit is converted into cold energy through thermal storage and absorption chillers to meet the cooling needs of the data center. In addition, when the absorption chiller cannot meet the cooling needs, central air conditioning can be introduced as an additional cooling measure to reduce the operating temperature of DCMG and ensure its stable operation.

[0085] Optionally, the distributed energy dispatching architecture also includes a public power grid and a natural gas grid; the public power grid and the natural gas grid are used to trade electricity with multiple data center microgrids.

[0086] Multiple DCMGs are interconnected through a public connection point, connected to the utility grid through power lines, and connected to the natural gas grid through natural gas streams. In addition, the distributed energy dispatching architecture meets its power needs through energy storage system charging and discharging, gas turbines, power transactions between DCMGs, and power transactions with the utility grid.

[0087] It should be noted that Figure 1 and Figure 2 All of them are illustrated with the distributed energy scheduling architecture including three data center microgrids. In practical applications, the number of data center microgrids in the distributed energy scheduling architecture can be determined according to actual needs, and the present invention will not be repeated here.

[0088] It can be understood that the design of each structure in the distributed energy scheduling architecture in the above embodiment is only one example of achieving the technical effect of the present application. In practical applications, it can also be adaptively deformed to achieve easily conceivable technical effects. The embodiment of the present application does not limit its structure.

[0089] Next, the energy dispatching method of the data center microgrid provided in the embodiment of the present application is described. The energy dispatching method of the data center microgrid in the embodiment of the present application is applied to the above Figure 1 The distributed energy dispatching architecture in the paper is described with the server of the data center microgrid as the execution entity.

[0090] In one embodiment, Figure 3 As shown, this embodiment includes the following steps:

[0091] S301, when there is a transaction demand in the data center microgrid, the server generates a power transaction request and uploads the power transaction request to the blockchain network; the power transaction request includes the transaction amount.

[0092] The server can determine whether there is a transaction demand in the data center microgrid by: obtaining status data of the data center microgrid, the status data including the power generation of renewable energy and the power usage of electrical equipment; inputting the status data into the action network to obtain the behavior data of the data center microgrid; the behavior data includes the transaction amount; if the transaction amount is within a preset range, it is determined that there is a transaction demand in the data center microgrid.

[0093] Among them, the status data of the data center microgrid may include photovoltaic power generation, wind turbine power generation, gas turbine power generation, charge state of the energy storage system, carbon quota, power demand of the data center microgrid, carbon usage of power-to-gas, electricity price, natural gas price and carbon price.

[0094] Optionally, at time t, the i-th data center microgrid DCMG in the distributed energy dispatch architecture i The state space of can be defined as follows:

[0095] s i,t =(s pv,i,t ,s wd,i,t ,s pgu,i,t ,s soc,i,t ,s quota,i,t ,P i,t ,Q p2g,i,t ,s e,i,t ,s g,i,t ,s c,i,t ) (1)

[0096] Among them, s pv,i,t is the photovoltaic power generation of the data center microgrid; s wd,i,t is the power generation of the wind turbine; s pgu,i,t is the power generation of the gas turbine; s soc,i,t is the state of charge of the energy storage system; s quota,i,t is the carbon quota; P i,t is the power demand of the data center microgrid; Q p2g,i,t is the carbon usage of power-to-gas; s e,i,t is the price of electricity; g,i,t is the price of natural gas; s c,i,t It’s the carbon price.

[0097] The action network can be used to predict the behavior data of the data center microgrid. Specifically, the status data of the data center microgrid is input into the action network, and the action network analyzes the status data and outputs the behavior data. The behavior data may include the charge and discharge amount of the data center microgrid's energy storage system, the natural gas usage, the transaction amount of the data center microgrid, the carbon usage of carbon capture, and the carbon usage of power-to-gas.

[0098] Optionally, at time t, DCMG i The behavior space of is defined as follows:

[0099] a i,t =(p gas,i,t ,p bat,i,t ,t DC,ij,t ,Q cc,i,t ,Q p2g,i,t ) (2)

[0100] Among them, p bat,i,t is the charge and discharge capacity of the energy storage system of the data center microgrid; p gas,i,t is the natural gas usage; t DC,ij,t is the transaction power of the data center microgrid; Q cc,i,t is the carbon usage of carbon capture; Q p2g,i,t is the carbon usage of power-to-gas. Each of the above behaviors is a continuous variable.

[0101] The transaction power can be positive, negative or 0. If the transaction power is positive, it means that the data center microgrid has surplus power. If the transaction power is greater than the preset positive threshold, it can be determined that the data center microgrid has a demand to sell electricity. If the transaction power is negative, it means that the data center microgrid has insufficient power. If the transaction power is less than the preset negative threshold, it can be determined that the data center microgrid has a demand to buy electricity.

[0102] Optionally, if the transaction power is a positive number, it means that the data center microgrid has surplus power, and it can be determined that the data center microgrid has a demand to sell electricity; if the transaction power is a negative number, it means that the data center microgrid has insufficient power, and it can be determined that the data center microgrid has a demand to buy electricity.

[0103] In an exemplary embodiment, generating a power trading request includes: obtaining a power trading price and a power trading time of a data center microgrid; and generating a power trading request according to the trading quantity, the power trading price and the power trading time.

[0104] Specifically, the transaction quantity can represent the total amount of electricity that the seller can provide (unit: kWh), or the amount of electricity required by the buyer; the electricity transaction price can be the electricity price in the status data, or a preset price range; the electricity transaction time can be the time when the data center wants to trade, or a preset trading time period, for example, the time period when electricity is available for sale, such as 10:00 to 14:00, etc. The transaction quantity, electricity transaction price and electricity transaction time are determined as a power transaction request, that is, the power transaction request includes the transaction quantity, electricity transaction price and electricity transaction time.

[0105] After generating a power transaction request, the data center microgrid can upload the power transaction request to the blockchain network.

[0106] S302, the server receives the transaction matching party information sent by the blockchain network; the transaction matching party information is obtained after the blockchain network matches the power transaction requests issued by other data center microgrids on the blockchain network.

[0107] The transaction matcher can be other data center microgrids that meet the supply and demand requirements of the data center microgrid; for example, the data center microgrid is the seller, and the transaction matcher is the buyer, and the transaction volume, power transaction time and power transaction price of the two match.

[0108] The blockchain network can match the power transaction requests issued by both parties based on market supply and demand. For example, if the quantity, price and time issued by seller A match the needs of buyer B, the blockchain network will automatically match the two parties for the transaction. After the blockchain is successfully matched, the transaction matching party information can be sent to the data center microgrid, and the data center microgrid receives the transaction matching party information sent by the blockchain network.

[0109] The transaction matching party information may include the transaction matching party's identity information, public key, digital certificate, etc.

[0110] S303, the server conducts electricity transactions with the transaction matching party through the smart contract of the blockchain network.

[0111] Smart contracts are pre-defined on the blockchain. After the matching conditions are met, buyers and sellers can conduct electricity transactions according to the preset process.

[0112] Preferably, electricity transactions are conducted with transaction matching parties through smart contracts of the blockchain network, including: if the data center microgrid is the transaction seller and the transaction matching party is the transaction buyer, electricity transaction data is generated according to the electricity transaction request of the transaction buyer; the electricity transaction data is signed by a private key, and the signed electricity transaction data is encrypted by the public key of the transaction buyer to obtain encrypted electricity transaction data; the encrypted electricity transaction data is sent to the transaction buyer, so that the transaction buyer conducts electricity transactions according to the encrypted electricity transaction data; the transaction buyer decrypts the encrypted electricity transaction data according to the private key and verifies the integrity of the electricity transaction data through the public key of the transaction seller.

[0113] Optionally, during the initialization phase of the data center microgrid, the elliptic curve digital signature algorithm (ECDSA) and asymmetric encryption technology can be used to assign a unique digital certificate, public key, and private key to each DCMG. Specifically, each DCMG generates a pair of public keys through ECDSA. and private key Using private keys in electricity trading Signing transactions, using public keys Verify the signature. At the same time, the distributed energy dispatching architecture uses its public key Assign a unique digital certificate to each DCMG

[0114]

[0115] Among them, dcg i Represents the identity information of the i-th data center microgrid, and ID is the identity identifier of the data center microgrid; is the public key of the data center microgrid; It is the private key of the data center microgrid; It is the authorization certificate for the data center microgrid.

[0116] The power transaction data may include the power transaction price, digital certificate, transaction quantity, status data and power transaction time of the corresponding data center microgrid.

[0117] Specifically, the transaction seller and the transaction buyer each have their own public key and private key. The transaction seller generates power transaction data, and in the process of power transaction, first encrypts the power transaction data with the public key of the transaction buyer, and then signs it with the private key of the transaction seller to obtain the encrypted power transaction data; after receiving the encrypted power transaction data, the transaction buyer can first use the public key of the transaction seller to verify the other party's digital signature. If it is confirmed that the power transaction data is signed by the transaction seller, it means that the transaction process has not been tampered with. Then, after verifying the signature, the transaction buyer uses his own private key to decrypt the power transaction data encrypted by the public key to obtain the power transaction data. The transaction seller uses his own private key to sign the power transaction data, which can prove that the transaction was initiated by the transaction seller.

[0118] The method of encrypting the power transaction data and signing the encrypted power transaction data can be shown in formula (4) and formula (5).

[0119]

[0120]

[0121] in, Represents the encrypted power transaction data, Indicates the public key of the buyer of the transaction Encrypt electricity transaction data, e,i is the electricity trading price, The digital certificate of the seller of the transaction, t DC,ij,t is the transaction power, s i,t is the status data of the transaction seller, ts is the electricity transaction time, The encrypted power transaction data is the encrypted power transaction data signed by the user. Indicates the private key of the transaction seller Sign the encrypted power transaction data. Each DCMG uses its own public key Generate a unique digital wallet address, as shown in formula (5), and upload it to the global account pool for fund flow and payment in the electricity trading process.

[0122] The seller uses his private key Sign the transaction data to generate a unique transaction hash, and generate a wallet address based on the hash value to ensure the uniqueness and non-repudiation of the transaction.

[0123] The smart contract includes the identity information of the two parties to the transaction, the amount of electricity traded by both parties, the electricity price and the electricity delivery time window; after confirming the electricity transaction data, the transaction buyer can determine the transaction resource volume based on the amount of electricity traded by both parties and the electricity transaction price; within the electricity delivery time window, the transaction buyer transfers the transaction resource volume to the transaction seller based on the transaction seller’s identity information; after receiving the transaction resource volume, the transaction seller transfers the corresponding amount of electricity to the transaction buyer.

[0124] S304: After the transaction is completed, the server generates transaction settlement information and encrypts and stores the transaction settlement information in the blockchain network.

[0125] After the power transaction is completed, the identity information of the two parties to the transaction, the amount of power traded by both parties, the power price and the power delivery time window can be determined as the transaction settlement information. The power transaction settlement information can be expressed by formula (6):

[0126] Trsc i ={ts,dcg i ,dcg j ,t DC,ij,t ,c DC,i,t} (6)

[0127] Among them, dcg i It can be the identity information of the transaction seller, dcg j The identity information of the buyer of the transaction; DC,i,t is the electricity transaction cost.

[0128] The transaction settlement information is encrypted through the public key of the data center microgrid and stored in the blockchain network; in this way, when the data center microgrid wants to view the transaction, it can decrypt and view it through its own private key.

[0129] In an exemplary embodiment, Figure 4 As shown in Figure 2, the construction process of the action network includes the following steps:

[0130] S401, obtaining the current state data of the simulated data center microgrid, and interacting with the simulated data center microgrid through the initial action network to obtain the behavior data, reward value and next moment state data of the simulated data center microgrid.

[0131] Firstly, the energy dispatch model of the data center microgrid is established. Each data center microgrid is modeled as an intelligent agent (simulated data center microgrid). Each simulated data center microgrid can include two groups of neural networks, namely the initial action network (μ i ) and the target action network (μ i ′), evaluation network (Q i ) and the target evaluation network (Qi ′).

[0132] First, observe the state of the simulated data center microgrid to obtain the current state data of the simulated data center microgrid, and then interact with the simulated data center microgrid through the initial action network to obtain the behavior data, reward value and next moment state data of the simulated data center microgrid. Specifically, it includes: determining the behavior data of the simulated data center microgrid through the initial action network; the behavior data is used to guide the simulated data center microgrid to trade electricity with the outside world; substituting the current state data and behavior data into the reward function to obtain the reward value; transferring the state of the simulated data center microgrid through the behavior data to obtain the state data of the simulated data center microgrid at the next moment.

[0133] The current state data is input into the initial action network to obtain the behavior data of the simulated data center microgrid, and then the current state data and behavior data are substituted into the reward function to obtain the reward value. According to the behavior data, the state of the simulated data center microgrid is transferred to the corresponding behavior to obtain the state data of the simulated data center microgrid at the next moment.

[0134] Input the current state data into the initial action network to obtain the behavior data of the simulated data center microgrid. Specifically, the current state data is input into the initial action network, the initial action network outputs the initial behavior data, and the sum of the initial behavior data and the noise value is determined as the behavior data. As shown in formula (7). At the beginning of each time slot t, the observed state s of the simulated data center microgrid i is i,t As the input of the initial action network, the initial action network is based on the policy function μ i Generate the behavior data of the current simulated data center microgrid i:

[0135]

[0136] in, is the initial action network μ i Network parameters; N i,t For its noise.

[0137] The optimization objective of the evaluation network is to minimize the electricity cost and carbon cost of DCMG. Therefore, the reward function at time step t can be shown as formula (8).

[0138] r i,t =c DC,i,t +c battery,i,t +c ug,i,t +c co2,i,t (8)

[0139] c DC,i,t =t DC,ij,t ×se,i,t (9)

[0140] c battery,i,t =p bat,i,t ×η wo,i,t (10)

[0141] c co2,i,t =Q ct,i,t ×s c,i,t (11)

[0142] c ug,i,t =t ug,i,t ×s e,i,t +p gas,i,t ×s g,i,t (12)

[0143] Among them, r i,t is the reward value, c DC,i,t is the electricity transaction cost of the data center microgrid; t DC,ij,t is the transaction power of the data center microgrid; c battery,i,t is the depreciation cost of the data center microgrid energy storage system; η wo,i,t is the depreciation factor of the energy storage system; c co2,i,t is the carbon cost of the data center microgrid; Q ct,i,t is the carbon trading volume of the data center microgrid; c ug,i,t is the cost of electricity trading between the data center microgrid and the public grid; t ug,i,t is the transaction power between the data center microgrid and the public grid; p gas,i,t is the natural gas usage of the data center microgrid.

[0144] S402, storing the current state data, behavior data, reward value and next moment state data as a data quadruple in an experience replay pool, and obtaining a preset number of historical experience data from the experience replay pool.

[0145] Among them, the historical experience data is a data quadruple.

[0146] Observe the current state data of the simulated data center microgrid i,t (including data center server load, power demand, remaining carbon storage capacity, carbon quota, solar irradiation rate, remaining battery capacity, carbon price, natural gas price and electricity price) to generate a behavior a i,t , through the interaction between behavior and environment, the simulation data center microgrid is transferred to the next state s i,t+1 , calculate the reward value r i,t , and the obtained data quadruple (s i,t ,a i,t ,r i,t ,s i,t+1) are stored in the experience replay pool D, and then a preset number of data quadruple (S data quadruple) are sampled in small batches from the experience replay pool D, and the preset number of data quadruple are used as historical experience data, which includes experience state data, experience action data, experience reward value and next experience state data.

[0147] S403, input the experience state data in the historical experience data into the target action network to obtain the first behavior data, input the experience state data and the first behavior data into the target evaluation network to obtain the joint state behavior value; determine the target evaluation value according to the reward value, the discount factor and the joint state behavior value.

[0148] The experience state data is input into the target action network to obtain the first behavior data, and the experience state data and the first behavior data are input into the target evaluation network to obtain the joint state behavior value, as shown in formula (13).

[0149] Each simulated data center microgrid has an independent target evaluation network, which integrates the status s of all simulated data center microgrids. n,t and behavior a n,t As the input of the target evaluation network, the target evaluation network outputs the joint state behavior value (Q value) function of each agent:

[0150] Q i (s t ,a t ) = E{γ t r i (s t ,a t )} (13)

[0151] Among them, Q i (s t ,a t ) is the joint state behavior value of Agenti at time t.

[0152] The reward value of the experience state data can be obtained by substituting the experience state data and the first behavior data into the reward function. Therefore, the target evaluation value is determined according to the reward value, the discount factor and the joint state behavior value, as shown in formula (14).

[0153] When training a neural network, a target evaluation value y can be calculated based on the approximate form of the Bellman equation i,t , which is used to measure the difference between the current estimate and the target value:

[0154]

[0155] Among them, y i,t is the target evaluation value; ri,t is the reward value, γ is the discount factor; Evaluate the network parameters of the network for the target; is the joint state behavior value at time t calculated by the target evaluation network, where s i,t a is the state data of the simulated data center microgrid i at time t; i,t is the set of behavioral data of all simulated data center microgrids at time t, obtained by state estimation using the target action network, that is, a i,t ={a 1,t ,a 2,t ,…,a M,t}.

[0156] S404, input the experience state data into the initial action network to obtain second behavior data, and input the experience state data and the second behavior data into the evaluation network to obtain an evaluation value.

[0157] The experience state data is input into the initial action network to obtain the second behavior data output by the initial action network, and the experience state data and the second behavior data are input into the evaluation network to obtain the evaluation value (estimated value).

[0158] S405, updating the initial action network according to the target evaluation value and the evaluation value to obtain the action network.

[0159] Specifically, it includes: determining the loss function value according to the target evaluation value and the evaluation value; updating the parameters of the target action network and the target evaluation network according to the soft update coefficient and the loss function value; updating the parameters of the initial action network and the evaluation network through the loss function value, and determining the updated initial action network as the action network.

[0160] First, by evaluating the target evaluation value y i,t The difference between the evaluation value and the mean square error is used as the loss function, and the gradient descent optimization is used to update the network parameters, as shown in formula (15). Formula (15) is the calculated loss function value.

[0161]

[0162] in, is the loss function value, is the network parameter of the evaluation network; N is the number of empirical state data.

[0163] According to the gradient of the loss function, update the parameters of the comment network:

[0164]

[0165] Among them, α Qi To evaluate the learning rate of the network.

[0166] Action network parameter update: During the training process, the action network updates parameters according to the direction of increasing joint state behavior value to obtain the maximum value. The gradient of action network parameter update is:

[0167]

[0168] in, is the joint state behavior value.

[0169] According to the gradient of the loss function, update the parameters of the action network:

[0170]

[0171] in, is the learning rate of the action network.

[0172] Therefore, the loss function value can be calculated according to formula (15), the parameters of the initial action network and the evaluation network can be updated according to formula (16) and formula (18), and the parameters of the latest initial action network can be used as the parameters of the action network to obtain the action network.

[0173] Both the target action network and the target evaluation network use soft update to update network parameters to improve the stability and convergence of training, as shown in formula (19).

[0174]

[0175] Where, τ is the soft update coefficient, τ∈[0,1]; Evaluate network parameters for the target; are the target action network parameters.

[0176] It should be noted that the optimization objectives of the evaluation network and the target evaluation network are formula (20), and the constraints formula (21) and formula (22) are satisfied.

[0177]

[0178] Constraints:

[0179]

[0180]

[0181] Wherein, the discount factor is represented by γ. Formulas (21) and (22) are the power constraints and thermal constraints of the data center microgrid, which aim to ensure the safe and efficient operation of the data center microgrid.

[0182] This invention focuses on three aspects: first, building a distributed energy dispatching architecture; second, data center microgrid power trading based on blockchain; and third, the algorithm flow. Figure 2 shown.

[0183] In an exemplary embodiment, this embodiment provides a distributed DCMG energy management framework based on blockchain, which regards each DCMG as an energy node, conducts power transactions with other DCMGs, and transmits their power transaction data to the blockchain. The specific process can be as follows: Figure 5 shown.

[0184] First, the data center microgrid low-carbon distributed energy management is used to observe the status of the data center microgrid and determine the behavioral data, and then electricity transactions are carried out with other data center microgrids based on the behavioral data.

[0185] The power trading part specifically includes: identity establishment and system initialization, triggering smart contracts, power trading, and storing power trading information. After the power transaction is completed, the settlement process of the smart contract is triggered, each energy node is managed according to the energy and carbon optimization strategy, and the power transaction is settled using smart contracts.

[0186] The encrypted transaction settlement information is generated into block data, and each transaction is stored in the blockchain network after verification and consensus by the nodes in the blockchain network.

[0187] The algorithm part can be Figure 6 As shown, the data center interacts with the environment to gain and store experience for training the neural network.

[0188] In the data center microgrid low-carbon distributed energy scheduling model established by the present invention, each data center is modeled as an agent, and each agent includes two groups of neural networks, namely, action network (μ i ) and the target action network (μ i ′), evaluation network (Q i ) and the target evaluation network (Q i ′). The action network outputs behavior according to the state observed by the agent, inputs the state and behavior of the environment into the evaluation network, and the neural network outputs the Q value of the corresponding behavior, corrects the strategy of the action network, and uses the action target network and the evaluation target network to stabilize the training process, such as Figure 3 Specifically, first, the data center initializes the experience playback buffer D, μ i , μ i ′, Q i and Q i ′; Secondly, the data center observes the current state s i,t(including data center server load, power demand, remaining carbon storage capacity, carbon quota, solar irradiation rate, remaining battery capacity, carbon price, natural gas price and electricity price) to generate a behavior a i,t , through the interaction between behavior and environment, the system moves to the next state s i,t+1 Get rewards i,t , and the quadruple (s i,t ,a i,t ,r i,t ,s i,t+1 ) is stored in the experience replay pool D. Then, the experience in the experience replay pool is sampled in small batches to train the neural network and reduce the correlation between samples. Finally, the state in the experience is input into the neural network to obtain the behavior and the joint state behavior value, and the TDtarget and loss function are calculated based on the final behavior and Q value, and the neural network parameters are updated using gradient descent.

[0189] The specific training process is shown in Table 1.

[0190]

[0191]

[0192] The present invention is based on distributed reinforcement learning. Through the interaction between each agent and the environment and the communication between multiple agents, the uncertainty of renewable energy output is alleviated, and the electricity cost and carbon cost of the data center are reduced. In addition, a low-carbon distributed energy dispatching architecture for data center microgrids based on blockchain is constructed to improve the security and privacy of data in the energy dispatching process of the data center and the security and integrity of power transactions.

[0193] When applying the energy dispatching method for the data center microgrid provided by the present invention, it is not necessary to Figure 3 The steps are executed in the order shown. The specific execution order of the steps can be determined according to needs, and the present invention does not limit this.

[0194] The above is an energy dispatching method for a data center microgrid provided by one or more embodiments of the present invention. Based on the same idea, the present invention also provides a corresponding energy dispatching device for a data center microgrid, which is applied to a data center microgrid in a distributed energy dispatching architecture. The distributed energy dispatching architecture includes multiple data center microgrids, such as Figure 7 shown.

[0195] Figure 7 A schematic diagram of an energy dispatching device for a data center microgrid provided by the present invention, the device 700 includes:

[0196] The upload module 701 is used to generate a power transaction request when there is a transaction demand in the data center microgrid, and upload the power transaction request to the blockchain network; the power transaction request includes the transaction amount;

[0197] The receiving module 702 is used to receive the transaction matching party information sent by the blockchain network; the transaction matching party information is obtained after the blockchain network matches the power transaction requests issued by other data center microgrids on the blockchain network;

[0198] The transaction module 703 is used to conduct electricity transactions with transaction matching parties through smart contracts of the blockchain network;

[0199] The storage module 704 is used to encrypt the power transaction data and store it in the blockchain network after the transaction is completed.

[0200] For the specific definition of the energy dispatching device of the data center microgrid, please refer to the definition of the energy dispatching method of the data center microgrid above, which will not be repeated here. Each module in the energy dispatching device of the above-mentioned data center microgrid can be implemented in whole or in part by software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0201] The present invention also provides a computer-readable storage medium, which stores a computer program, which can be used to execute the above Figure 3 An energy scheduling method for a data center microgrid is provided.

[0202] The present invention also provides Figure 8 The structural diagram of the computer device shown in FIG. Figure 8 As shown in the figure, at the hardware level, the computer device includes a processor, an internal bus, a network interface, a memory, and a non-volatile memory, and may also include other hardware required for the business. The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs it to achieve the above Figure 3 An energy scheduling method for a data center microgrid is provided.

[0203] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided by the present invention can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory or optical memory, etc. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0204] The technical features of the above embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of the present invention.

Claims

1. A method for energy dispatching of a data center microgrid, characterized in that: A server for a data center microgrid applied to a distributed energy scheduling architecture, wherein the distributed energy scheduling architecture includes a plurality of data center microgrids; the method includes: When there is a transaction demand in the data center microgrid, the server generates a power transaction request and uploads the power transaction request to the blockchain network; the power transaction request includes the transaction amount; The server receives transaction matching party information sent by the blockchain network; the transaction matching party information is obtained after the blockchain network matches the power transaction requests issued by other data center microgrids on the blockchain network; The server conducts electricity trading with a transaction matching party through a smart contract of the blockchain network; The server generates transaction settlement information after the transaction is completed, and encrypts and stores the transaction settlement information in the blockchain network.

2. The method according to claim 1, characterized in that The method further comprises: The server obtains status data of the data center microgrid, wherein the status data includes power generation of renewable energy and power usage of power-consuming devices; The server inputs the state data into the action network to obtain the behavior data of the data center microgrid; the behavior data includes the transaction power; If the transaction power is within a preset range, the server determines that there is a transaction demand in the data center microgrid.

3. The method according to claim 2, characterized in that The generating of the power transaction request comprises: Obtaining the power transaction price and power transaction time of the data center microgrid; The power trading request is generated according to the trading power, the power trading price and the power trading time.

4. The method according to claim 3, characterized in that The construction process of the action network includes: Acquire the current state data of the simulated data center microgrid, and interact with the simulated data center microgrid through the initial action network to obtain the behavior data, reward value and next moment state data of the simulated data center microgrid; The current state data, the behavior data, the reward value and the next moment state data are stored as a data quadruple in an experience replay pool, and a preset number of historical experience data are obtained from the experience replay pool; the historical experience data is a data quadruple; Inputting the experience state data in the historical experience data into the target action network to obtain first behavior data, inputting the experience state data and the first behavior data into the target evaluation network to obtain a joint state behavior value; determining the target evaluation value according to the reward value, the discount factor and the joint state behavior value; Inputting the experience state data into the initial action network to obtain second behavior data, and inputting the experience state data and the second behavior data into an evaluation network to obtain an evaluation value; The initial action network is updated according to the target evaluation value and the evaluation value to obtain the action network.

5. The method according to claim 4, characterized in that The interaction with the simulated data center microgrid through the initial action network to obtain the behavior data, reward value and next moment state data of the simulated data center microgrid includes: Determine the behavior data of the simulated data center microgrid through the initial action network; the behavior data is used to guide the simulated data center microgrid to trade electricity with the outside world; Substituting the current state data and the behavior data into a reward function to obtain the reward value; The state of the simulated data center microgrid is transferred through the behavior data to obtain the state data of the simulated data center microgrid at the next moment.

6. The method according to claim 4, characterized in that The updating of the initial action network according to the target evaluation value and the evaluation value to obtain the action network comprises: Determine the loss function value according to the target evaluation value and the evaluation value; Updating parameters of the target action network and the target evaluation network according to the soft update coefficient and the loss function value; The parameters of the initial action network and the parameters of the evaluation network are updated according to the loss function value, and the updated initial action network is determined as the action network.

7. The method according to claim 1, characterized in that The power transaction with the transaction matching party through the smart contract of the blockchain network includes: If the data center microgrid is a transaction seller and the transaction matcher is a transaction buyer, generating power transaction data according to the power transaction request of the transaction buyer; Signing the power transaction data with a private key, and encrypting the signed power transaction data with the public key of the transaction buyer to obtain encrypted power transaction data; The encrypted power transaction data is sent to the transaction buyer, so that the transaction buyer conducts power transaction according to the encrypted power transaction data; the transaction buyer decrypts the encrypted power transaction data according to the private key and verifies the integrity of the power transaction data through the public key of the transaction seller; Accordingly, the smart contract includes the identity information of the two parties to the transaction, the amount of electricity traded by both parties, the electricity price and the electricity delivery time window; after the electricity transaction is completed, the method further includes: The identity information of the two parties to the transaction, the amount of electricity traded by the two parties, the electricity price and the electricity delivery time window are determined as the transaction settlement information.

8. A distributed energy scheduling architecture, characterized in that: The distributed energy dispatching architecture includes a plurality of data center microgrids, each of which is electrically connected via a common connection point; each of which includes a server; The server is used to upload a power transaction request to the blockchain network when there is a transaction demand in the data center microgrid, receive transaction matching party information sent by the blockchain network, and conduct power transactions with the transaction matching party through the smart contract of the blockchain network, generate transaction settlement information after the transaction is completed, and encrypt and store the transaction settlement information in the blockchain network; The power transaction request includes the transaction quantity; the transaction matching party information is obtained after the blockchain network matches the power transaction requests issued by other data center microgrids on the blockchain network.

9. The distributed energy dispatching architecture according to claim 8, characterized in that: Each data center microgrid includes servers, energy supply devices and energy consumption devices; the energy supply devices include photovoltaics, wind turbines, gas turbines and energy storage systems, and the energy consumption devices include carbon capture devices, carbon storage devices, power-to-gas devices, cogeneration devices, heat storage devices, absorption chillers and central air conditioners; The data center microgrid captures carbon through the carbon capture device, and converts the captured carbon into methane through the power-to-gas device to provide fuel for the cogeneration device, and converts the heat energy generated by the cogeneration device into cold energy through the heat storage device and the absorption chiller; The central air conditioner is used to provide cold energy for the data center microgrid when the absorption chiller cannot meet the cooling demand of the data center.

10. The distributed energy dispatching architecture according to claim 8 or 9, characterized in that: The system also includes a public power grid and a natural gas grid; The public power grid and the natural gas grid are used to trade electricity with the multiple data center microgrids.

Images