Abstract
In recent years, rapid advancements in smart grid technology and smart metering systems have raised serious privacy concerns about the collection of customers’ real-time energy usage behaviors. Due to cybersecurity attacks and threats, data aggregation operations in a smart grid are challenging. The majority of existing techniques have high computation and communication costs and are still vulnerable to various security and privacy concerns. This paper proposes a deep learning and homomorphic encryption-based privacy-preserving data aggregation model to mitigate the negative impact of a flash workload on the accuracy of prediction models. The model also ensures a secure data aggregation process with low computational overhead. The proposed model is 80% more effective than the traditional approach in detecting smart meter manipulation, and the computation cost is 20% to 80% less than existing techniques. Thus, the proposed blockchain and homomorphic encryption-based data aggregation (BHDA) scheme shows a significant improvement in performance and privacy preservation with minimal computation overhead for data aggregation in smart grids.
•A homomorphic encryption-based data aggregation model for the smart grid system.•A blockchain-based data aggregation framework to enhance security in a smart grid system.•The framework reduces the computation to work with smart meters with low computation power.•The computation cost of the model is 20% to 80% less as compared to existing techniques.