SynergyGrids: Blokzincir destekli gorevdes enerji ticareti ve tuketici yonetim sistemi

With depletion in fossil energy and increasing environmental problems, energy markets have shifted towards renewable energy resources, such as wind and solar. The ultimate aim is to bridge the gap between generation and demand. As these sources can be harvested on a small scale, increasing number of producers are becoming capable of selling surplus energy. This fostered the concept of peer-to- peer (P2P) energy trading and sharing, by encouraging consumers to become prosumers (i.e. capable of both producing and selling surplus energy). Nowadays, the concept of energy trading and sharing is no longer a dream due to the emergence of the smart city concept. However, due to the heterogeneity of these energy resources (i.e. difference in energy generation and dependency) and a low degree of reliability, the task of efficiently managing these energy resources becomes difficult. Various energy trading systems and management tools have been developed, however, they mostly use centralized cloud systems and face inherent problems of high latency, low scalability, and security problems. Hence, a strong tool that can protect the prosumer data using emerging technology, log the changes for audit purposes, and eventually improve the performance of the system is needed. In this context, blockchain technology as a distributed ledger has emerged as an alternative to centralized systems. In this thesis, state-of-the-art P2P energy trading and sharing solutions are investigated, highlighting their benefits, drawbacks and open issues by classifying them into sub-categories of Infrastructure-based Energy Trading, Large Scale Energy Storage Based Trading, and Ad hoc P2P Energy Trading. Considering the investigated open issues, a blockchain assisted P2P energy trading and prosumer management framework, namely SynergyGrids, that consists of two main modules is developed. As the first module of the framework, a decentralized grouping mechanism, named SynergyChain, is proposed for improving scalability and decentralization of the prosumer grouping mechanism in the context of P2P energy trading. Smart contracts are used for storing transaction information and for the creation of the prosumer groups. SynergyChain integrates a reinforcement learning module to create a self-adaptive grouping technique for improving the overall system performance and profitability. The proposed SynergyChain model improves the time complexity for each energy request as it reduces the search space of available energy from n (number of prosumers) to k (number of groups formed). As the second module, a hybrid trading system, named FederatedGrids, that manages both the prosumers from local microgrids and across the microgrids to allow energy trading and sharing is proposed. This allows consumers to buy energy from local resources and reduce reliance on utility grids. FederatedGrids proposes a decentralized blockchain-based federated learning model that is used to predict the energy demand of the consumers. The proposed SynergyGrids modules are developed using Python and Solidity and have been analyzed using Ethereum test nets. The comprehensive experimental analysis using the Hourly Energy Consumption data-set has been conducted. SynergyChain shows a 39.7% improvement in the performance and scalability of the system as compared to the centralized system, and 18.3% improvement in the profitability of the prosumers. The blockchain-based federated learning to predict the demand load proposed in FederatedGrids shows comparable accuracy with an average 97.8% improved network load. The energy trading and sharing results show a 17.8% decrease in energy cost for consumers and a 76.4% decrease in load over utility grids when compared to other models from the literature.