Trading platform for cooperation and sharing based on blockchain within multi-agent energy internet

0 Introduction

In 2015, the State Council issued the “No.9 Circular of Several Opinions of the Central Committee of the Communist Party of China on Further Deepening the Reform of the Electricity System,“ stating that the gradual liberalization of the electricity sales market includes both the electricity sales side and the user side.As the electricity sales side links the ups and downs to balance the supply and demand on both sides more effectively, the core values of electricity reform include changing the current electricity sales model of power grid companies, gradually liberalizing the electricity sales market, developing diversified electricity sales entities, encouraging power sellers and users to actively participate in market competition, and restoring the attributes of power commodities [1,2].

However, electricity is unlike general commodities.As it cannot be stored on a large scale, electricity trading has strong randomness and real-time requirements.The transaction volume and transaction price are related to the economic benefits of the electricity market, and also directly affect the safety and reliability of the power grid system.Transactions in the power market are determined by market competition, and are also coordinated and restricted by the market [3].At present, bilateral transactions, centralized transactions, or a combination of the two in medium- and long-term markets are physical contracts that must be rigidly executed.As the executive agency of the transaction plan,dispatching implements the transaction plan by formulating power generation plans.Thus, rapid response and realtime transactions are difficult to achieve in the electricity market.The market must be given greater flexibility and openness through partial compliance management, such that transactions can be autonomous, intelligent, and partially decentralized.In addition, the unified supervision and security protection mode of the electricity spot market must be strengthened to form a bilateral intraday time-of-use electricity price mechanism.Transaction records, financial settlements, and operating data must be protected from tampering by both parties and third-party intruders through transaction data or privacy data, ensuring the integrity,availability, reliability, and traceability of data and other information security functions.

Small-capacity distributed power generation units connected to the distribution network, such as microgrids,integrated energy, and virtual power plants, offer the advantages of low carbon emissions and environmental protection, short transmission distances, and low energy consumption and can be connected on a large scale.A microgrid is a consumer-based power grid.Its main purpose is to autonomously control its own resources to meet the energy demand [4].If there is a power surplus, it is sent to the power grid.Otherwise, the power grid directly provides power to the consumers.The microgrid operates as a consumer but is a new type of electricity market entity with both consumers and producers.

Information technology is constantly improving and developing; the concept of the energy internet has emerged.Compared with the traditional energy system, the energy internet offers “openness, reciprocity, interconnection, and sharing”.With the large-scale integration of distributed energy, the energy market of the energy internet is different from the traditional transmission grid market.It is developing in the direction of diversification of the main body, a flat structure, and diversification of commodities,forming a multi-party competitive market mechanism and a flexible market environment [5].

With changes in the openness, mechanism, and environment of the electricity sales market, traditional centralized transaction methods of the transmission grid are not suitable.An effective solution based on blockchain technology is provided for these market changes in the energy internet.Blockchain technology offers distribution,intelligence, and marketability, consistent with the “open,peer-to-peer, interconnected, and sharing” characteristics of the energy internet.

Blockchain and the energy internet have similar network topologies [6].By mapping the structure and nodes of the blockchain to the energy internet, a trusted interaction mechanism is produced between the energy internet nodes that can support multi-agent and autonomous intelligent transactions between operators, virtual power plants, and microgrid nodes.Blockchain is a new decentralized system based on a peer-to-peer (P2P) network [7].Each node in the network is responsible for the network routing, verifying and disseminating block data, storing data records, and discovering new nodes.Reference [8] designed a distributed bidding transaction algorithm based on blockchain technology with orderly aggregate signature and orderpreserving encryption technology, improving transaction security, reliability, and efficiency.Reference [9] designed a distributed energy transaction model based on blockchain smart contracts and continuous double auction (CDA)and proposed an improved consensus algorithm based on DPOS that offers more transaction benefits for both parties.Reference [10] proposed a blockchain-based distributed electric energy bidding and trading platform design scheme, and designed a method that combined access mechanism and authority control, improving the security and transparency of user transactions.Reference [11]proposed a trusted storage mechanism including identity authentication, sensor data upload, an interplanetary file system (IPFS), blockchain upload, and access verification,effectively solving the problems of storage expansion and data tampering identification.

Blockchain technology can also be applied to distributed energy transactions in microgrids.P2P transactions and centralized clearing are two distributed energy transaction nodes [12].The former uses blockchain decentralization and traceability; buyers and sellers do not need to negotiate prices.The latter uses decentralization; market participant quotations are matched to form a supplydemand curve and then cleared.Blockchain and microgrid markets have a similar distributed topology [13].Thus,blockchain technology can be used to simplify transaction modes, realize peer-to-peer transactions, promote nearby consumption of distributed energy, and reduce electricity transaction costs, improving the economics of electricity trading.The reward-and-punishment mechanism ensures fairness to individuals participating in trading [14] and compensates for the lack of game models in obtaining global information.The multi-time scale microgrid economic dispatch method based on smart contracts improves the information transparency of the microgrid system [15].Blockchain and a continuous two-way auction mechanism have perfected the direct electricity transaction mode and the strategy between distributed power sources and users in the microgrid [16].Reference [17-18] proposed a power network management system based on blockchain technology using Ethereum as a development platform to establish a distributed energy power interactive management blockchain; they designed a distributed electric energy dispatching strategy using the K-means clustering algorithm and the particle swarm optimization algorithm (PSO).

In some countries, there is a relatively mature blockchain-based energy internet and a novel cooperation model between microgrid distributed energy transaction and blockchain.US energy company LO3 and blockchain development company Consensus Systems have established a distributed photovoltaic power distribution blockchain platform, Transactive Grid, the world’s first energy blockchain market [19].Reference [20] used distributed ledgers to apply blockchain-based distributed demand-side management to balance energy demand and production.Reference [21] built a decentralized power trading model based on the distributed power market framework of P2P network technology.Blockchain can adapt to its operation mode, topology, and security protection, and better support the construction and protection of the power trading market of the energy internet.

Based on the foregoing analysis, a blockchain trading platform that can adapt to multi-agent cooperation and sharing of the energy internet is investigated in this study.The value of the integration of blockchain with electricity market transactions is analyzed, and a blockchain transaction framework is designed for multi-agent cooperation and sharing of the energy internet.The market transaction nodes are modeled according to power system modeling in the physical layer, and the transaction consensus strategy is modeled in the cyber layer; both are verified in a modified IEEE 13 testing feeder of a distribution network.A transaction example based on the Ethereum private blockchain is presented in the multi-agent cooperation and sharing transaction platform.

1 Analysis of Blockchain and Electricity Market

1.1 Application and value of blockchain in distributed energy market

As the electricity trading market continues to open, more trading entities are attracted, the trading volume increases,and the transaction types broaden; this necessitates a more stable trading system.This section discusses the application and value of blockchain in the distributed energy trading market.

As a novel form of data structure, blockchain technology offers decentralization, openness, autonomy, immutability of information, anonymity, and traceability.Thus, with the support of blockchain technology, the energy internet trading platform built with the distributed energy trading market has the following characteristics.The analysis is shown in Fig.1.

Fig.1 Analysis of characteristics and advantages of blockchain and electricity market transaction

· Wisdom and mutual trust: Each layer of power network communication can be interconnected; with support from the autonomy of blockchain technology, each transaction participant can realize intelligent mutual trust.

· Data sharing: With openness, supplier and consumer information can be released on the trading platform during a transaction.After a transaction is completed, transaction data is stored in the blockchain.

· Transparent transaction: The immutability and openness of the blockchain provide greater transparency in distributed electric energy transactions, and a platform allowing information exchange with other industries.

1.2 Framework of electricity transaction based on blockchain

The energy flow of the traditional power system is oneway transmission through a vertical structure; the recipient of the price is the end customer.With a large number of distributed power generation customers and demand-side response capabilities in the distribution network, many power producers and consumers emerge.At this time, a single distributed generation transaction mechanism in the traditional power system cannot easily satisfy different types of power producers and consumers in different regions, and it is difficult to optimize the dispatch of large-scale distributed resources with fully centralized transaction management.

Blockchain has been widely used in finance and medical treatment and has also been introduced into the energy field.Its characteristics enable it to be used as the underlying technical architecture for information transmission for distributed power facilities and equipment.In the distributed microgrid, each participant acts as a node in blockchain technology to form a distributed power information communication network.Therefore, this study uses blockchain technology to construct a framework diagram of the transaction mode of each node alliance(Fig.2) to promote power transactions between local supply and demand users and thereby realize energy sharing and distributed power generation for nearby consumption.

Fig.2 Analysis of characteristics and value of blockchain and electricity market transaction

During operation, each node in the P2P network competes for the right to establish a new zone by executing the authorization certification mechanism, stores the transaction data that is produced in the block, and connects the block to the main blockchain to form a blockchain structure.Each data block is composed of a block header and a block body.The block header contains most of the functions of the block, including the version number of the current block, the address of the previous block, the Merkle root that provides the transaction data through the Merkle tree hash calculation process, and the timestamp for the generation of the block.On the basis of this management structure, through the Ethereum consensus mechanism and credit scoring system, efficient management of the entire life cycle of grid big data creation, collection, organization,storage, utilization, and removal can be achieved, while ensuring data security, privacy, and reliability.

Electricity trading is different from trading general commodities.The execution of a transaction is affected by the price of transmission and distribution, the scale of electricity, and power dispatch.In terms of transaction methods, bilateral negotiated transactions and matching transactions are currently the mainstay, and bidding transactions are supplemented.In this study, the rules of listing are used for trading.Transaction entities can continue to be listed on the blockchain-based platform.A quotation must include the quotation and transaction volume.Transaction entities can delist if they are willing to trade on a first-come-first-served basis.The price is controlled by the calculation trend of the physical layer, and the priority of delisting depends on the price and the size of the transaction,to provide information for both parties and obtain more transaction opportunities.

Based on Fig.2, the process of completing a transaction is explained as follows.At the physical level, the price of electricity from the grid is calculated per hour based on the energy flow.The nodes (subjects and users) in the cyber layer only use the price calculated by the corresponding physical layer to pre-quote and complete the transaction with other users or the main network on the platform through the information flow when a transaction is necessary.The transaction is initiated by a certain user through a smart contract, and the executed fee in the contract is settled with the corresponding user through the blockchain.The transaction data is confirmed and stored in the block; the block is connected to the chain; the transaction is completed;and the transaction data is stored on the chain.

2 Modeling of Nodes in Market Transaction

2.1 Modeling of power system in physical layer

In the physical layer, the price is calculated based on the distribution location marginal price (DLMP) of the distribution network.It is a mathematical model of the distribution network used to set the wholesale price and reflect the clearing value at different positions in the distribution system.The concept corresponds to the marginal price (LMP) of the transmission system.In the transmission system, if there is no congestion and loss, all LMPs are market clearing price (MCP) because only the congestion cost is reflected; otherwise, there are different bus prices in the system.Generally, the main reasons for difficulty in LMP pricing at different locations are 1) transmission loss and 2) congestion constraints.When the congestion constraint reaches the upper limit, the cost of physical transmission loss changes the LMP for the entire system.Energy cost, congestion charge, and transmission system loss are the most important components of LMP calculation.LMP and DLMP are used to calculate the price in a power system and maximize the cost of energy,congestion, and loss.

The objective function for optimal power flow satisfies all constraints for the unit output to meet the real-time load at the minimum cost.The DC approximation method is used to model the node formula of the distribution network[22-23].The objective function is the system minimum cost function and its constraints.

where x includes the clearance of all units Pn and bus voltage power angle θk; an bn, and cn are the quadratic cost function parameters of the unit; h(x) is part of the constraint equation; g(x) is part of the constraint inequality; PLoadk represents the load on bus k; Xkj is the reactance between buses;is the maximum output value of the unit; is the capacity between buses; Ωk is the set of all buses connecting bus K; ΩN is the set of all nodes; ΩG is the set of all units;ΩL is the set of all buses.

The Lagrange solution of the minimum cost function of the operating unit is expressed as

The necessary conditions of Karush-Kuhn-Tucker Conditions (KKT) are expressed as

2.2 Transaction consensus strategy

Currently, there are three main modes for distributed consumers to participate in market transactions.With the small capacity and large fluctuation of distributed energy,the profit model is based mainly on trading through policy subsidies or a third-party aggregation agent.However,the price provided by the third party is lower than the actual market price, and long-term policy subsidies are not conducive to technological progress and sustainable development of distributed renewable energy.In the distributed market trading mechanism, although the third party is removed and can directly trade with the power market, a trading mode between distributed energy entities is still not realized.A P2P decentralized trading mechanism can provide negotiated shared energy, fair competition, a user-centered trading environment for each market subject,and encourage users to participate in the market more actively.

In the node modeling of a distribution network,considering the P2P transaction mechanism, the traditional direct current optimal power flow (DCOPF) modeling and decentralized consensus C+I algorithm are combined to realize a co-governance transaction strategy based on a decentralized DCOPF.

Currently, there are four decentralized DC optimal power flow algorithms with a mature and strong theoretical basis: optimal message passing (PMP), auxiliary problem principle (APP), optimization condition decomposition(OCD), and consensus + innovations (C+I).The classification, performance comparison, and other algorithm information are summarized in Table 1 [22-24].

Table 1 Comparison of optimal power flow solutions

Algorithm Classification Decomposition method Central regulation PMP Decentralization Enhanced Lagrange relaxation No APP Decentralization Enhanced Lagrange relaxation No OCD Decentralization KKT conditions No C+I Decentralization KKT conditions No

The decentralized multi-agent C+I method is used in this study.The C+I method automatically optimizes the global environment according to the local node and neighbor node information.Generally, the main parameters λ and P must be updated at different times using Eq.(14),where i is the number of iterations; j is the number of nodes; t is the number of hours; αi and βi are the super parameters in the iterative formula, with values of 0 when the number of iterations increases infinitely; ωj represents the communication information of the bus connection in the physical layer; an and bn are the cost parameters of the node unit.

3 Case Study

In this section, the data assumption and data of the example are analyzed.The decentralized C+I algorithm is used to complete the consensus strategy.The application display interface of the blockchain is provided and analyzed.

3.1 Hypothesis and data

The example system is a modified IEEE 13 bus test distribution network feeder system, as shown in Fig.3.In this power system topology, we consider two microgrids and three solar cells as the distributed energy supply of the system.There are power supply devices on nodes 650, 646,633, 684, 692, and 680.Node 650 (source bus) is directly connected to the main power grid; nodes 646, 633, and 684 are distributed solar energy nodes; nodes 692 and 680 are microgrids.The generation cost parameters and maximum power output data are shown in Table 2.The loads are proportionally located on nodes 646, 645, 632, 634, 611, 671,675, and 652.Load data are presented in Table 3, including line segment and impedance configuration parameters such as feeder impedance, connection, and length.

Fig.3 Modified IEEE 13 bus test feeder system for distribution network

Table 2 Power generation data sheet

?

Table 3 Load data sheet

?

3.2 Analysis of Consensus Results

The example data is applied to the decentralized method of C+I to calculate the DCOPF of the modified IEEE 13 bus test distribution network feeder system and obtain the iterative co-governance trading strategy (clearing and price).The combined iterative formula is expressed as

The algorithm is presented in Table 4.

Table 4 Algorithm flow

Algorithm 1 decentralized co-governance trading strategy 1: Initial parameters λ, θ, Pn, and μ 2: while the convergence value is not satisfied do 3: for i = 1: n (number of units) do 4: update λi jt+,1 5: update Pnt i,+1 6: update θi jt+,1 7: update Pnt i,+1 8: Sharing parameters with connected buses λi jt+,1 and θi jt+,1 9: end if 10: end while

1) and represent the transmission limit between feeders, which is set as 500 kW; 2) Ωi represents the neighbor communication information of the bus connection in the physical layer; 3) ΩN represents the collection of all nodes; ΩG represents the set of all units; ΩL represents the set of all buses; 4) The other Lagrange multipliersand are updated using the new λ iteration results;5) α, β, γ, and δ are the four super parameters in the iterative formula.

The co-governance trading strategy results for day-ahead clearing and price after 1000 iterations are shown in Fig.4 and Fig.5.Fig.4 is a 24-hour day-ahead clearing and price overview of all nodes.Fig.5 shows the iteration results of the consensus strategy for all nodes in the 10th hour (maximum photovoltaic) and the 18th hour (maximum price).

Fig.4 24-hour unit clearing and marginal price strategy

Fig.5 Unit clearing and marginal price after 1000 iterations

3.3 Application of blockchain

An example of the distributed energy trading platform based on node.js is introduced in this section.

The blockchain is based on the underlying design of Ethereum private chain technology, as shown in Fig.6;its main programming language is Go.The distributed consensus trading strategy results are obtained using the decentralized C+I algorithm simulated by MATLAB, and the strategy result is transmitted to the smart contract in the blockchain through API.By designing a model view controller for the web page, direct transaction between users is possible with the integration of blockchain and a decentralized consensus mechanism.

Fig.6 Trading platform framework based on node.js

The platform address is localhost:3000.The home page of the trading platform is shown in Fig.7; it contains the account number and balance display.Except for the miner account, the initial account values are 0.A miner account maintains the ecology of the blockchain and confirms the transaction to form a transaction block for the blockchain.The platform also includes a block information query, a transaction block information query, an intelligent contract query, and a transaction implementation module.Due to the limited number of pages, they are not listed individually.In the platform, transaction transfers are implemented by blockchain node users through the transaction implementation module.Fig.8 shows detailed transaction information.Fig.9 shows the home page after the miner confirms the successful completion of two transactions.

Fig.7 Trading platform home page

Fig.8 Transaction block information queried in background

Fig.9 Platform home page after transaction

With decentralization and data security, blockchain technology can achieve safe and efficient P2P energy transaction.The token incentive mechanism based on blockchain provides token rewards for users with high participation.Using smart contract technology,the transaction process for distributed generation is automatically executed; the transaction data are written into a Merkle tree in real time, allowing transaction data tracking for the entire process.Organic combination of blockchain technology and power trading is realized; the security and convenience of power trading are improved, and clean energy production and nearby consumption are promoted.

4 Conclusion and Future Work

This study examines the value of combining the distributed new energy trading mode and blockchain,constructs the DCOPF model of a power system in the physical layer and a C+I strategy of distributed new energy node pricing in the cyber layer, and verifies the feasibility of a modified IEEE 13 node test distribution network feeder system.After the strategy iteration of the consensus algorithm is completed, the day-ahead clearing and price results are quickly calculated and converged.

A power market trading platform with a high proportion of distributed new energy was built based on blockchain technology.Two blockchain transactions based on the Ethereum private blockchain with multi-agent cooperation and sharing were displayed on the platform, and the transaction transfer between blockchain node users based on the day-ahead price was realized.

With the development of distributed transaction mode and blockchain technology, the distributed transaction platform based on blockchain and characterized by intelligent mutual trust, data sharing, and transparent transaction can be deeply mined.Its autonomy, openness,and security bring more possibilities to the future trading platform.

Acknowledgements

This work was supported by the Smart Grid Joint Fund of the National Natural Science Foundation of China (No.U2066209) and the Science and Technology Project of the China Electric Power Research Institute (No.AI83-20-002).

Declaration of Competing Interest

We declare that we have no conflict of interest.