Research On Distributed Power Trading System Based On Blockchain

With the rapid growth of energy demand and the rapid promotion of market-oriented reform of energy trading,the transaction and utilization of distributed renewable energy has gradually become the mainstream development trend of the current energy market.However,this has also caused a huge impact on the traditional centralized energy trading mode,and the decentralized and market-oriented reform of energy trading is imminent.As an emerging distributed storage technology,blockchain has the characteristics of decentralization,secure and reliable information storage.It is an important technical support to solve the problem of distributed energy transaction.At the same time,intelligent algorithms such as deep learning and reinforcement learning can also be used to help users solve transaction decision-making problems.Based on the above,this paper makes an intensive study on the energy trading system based on blockchain and the optimization of energy trading strategy.The main research work includes:Aiming at the problems of low security and poor scalability of the traditional centralized energy trading network.This paper designs an energy trading platform based on blockchain,and formulates the processes of registration and authentication,transaction execution,supervision and traceability.At the same time,an encrypted election consensus algorithm based on transaction participation is designed for the energy trading scenario to ensure the stability of decentralized network and the enthusiasm of node participation.Design a pre-sale matching energy trading mechanism based on blockchain and deep learning,use deep learning to learn the historical data recorded on the blockchain,predict the future energy production and load data,then record on the blockchain.Users can conduct power pretransaction in advance according to the prediction results on the chain.Meanwhile,CNN-LSTM prediction model is designed to improve the prediction accuracy.Through simulation analysis,the feasibility and superiority of the pre-sale matching energy trading method is proved,and the CNN-LSTM hybrid model has good predictive performance.In order to help users formulate high quality trading strategies,maximize the benefits of all parties in the power grid on the basis of realizing power balance.The energy trading system model is designed,considering many influencing factors such as trading,storage,transmission and balance,and comprehensively setting the optimization objectives of trading strategy.At the same time,an energy trading strategy optimization scheme based on MADDPG is proposed.Simulation results show that compared with the other two algorithms,the proposed algorithm can maximize the overall utility function,reduce user costs,and is more consistent with the actual application scenario.