Blockchain Intelligence Empowered Uncertainty Management in IoT Assisted Smart Grids

Abstract

The recent advances in renewable energy sources impose an urgent need on global community to find an alternative measure for climate-unfriendly fossil fuels. As a result, the energy flow across smart grids has become bidirectional that requires greater attention. Despite the increasing advantages of smart grid application, it deals with greater challenges in managing the supply and demand of power sources. This is because of the reason that power generation, distribution and consumption across smart grids are highly complex. Since all these processes are time-dependent, dynamic management of grid stability has become a significant requirement. Most existing systems adopt a distributed system with a central authority to solve this problem. Such systems are more prone to various security attacks and become a single point of failure in many cases. This paper proposes a blockchain-based decentralized multiparty learning system to ensure smart grid stability with enhanced security and efficiency measures. The experimental observations are made with a power grid simulation dataset taken from Kaggle. From the experiment, it is observed that the proposed approach takes an average of 25ms to read the data across the block, and it takes around 4s to generate a new block. Further with respect to the addition of more intelligent terminals, the proposed approach consumes only 70% of the energy required by conventional methods to perform the task. The prediction and classification accuracy of the proposed system is also analyzed, and it shows 98% accuracy.