Network analysis in a peer-to-peer energy trading model using blockchain and machine learning

Existing technology like smart grid (SG) and smart meters play a significant role in meeting the everlasting demand of energy consumption, supply, and generation for peer-to-peer (P2P) energy trading between different distributed prosumers. Whereas blockchain when used with P2P energy trading plays a major role in cost and security by eliminating any involvement of outsiders and third parties. However, existing works related to the blockchain with P2P energy trading are engaged in increasing the cost related to resource allocation, latency, computational processing, and large network setup. The objective of this paper is to design and develop a three-tier architecture, an analytical model, and a hybrid algorithm for network analysis in a blockchain-based P2P energy trading system using reinforcement learning (RL) and feed forward neural network (FFNN) techniques. In this model, we will examine the various parameters and tradeoffs which affect the delay, throughput, and security in P2P energy trading. This will lead to profitable P2P energy trading between different distributed prosumers. By analyzing the simulation results of the proposed model and algorithm by benchmarking with the existing state-of-the-art techniques it's clear that the proposed algorithm shows marked improvement over network latency generated results. The simulation of the model is conducted using the iFogSim simulator, Ganache with Ethereum platform, Truffle, Python editor tool, and ATOM IDE with solidity. © 2023 Elsevier B.V.