Efficient and secure integration of renewable energy sources in smart grids using hybrid fuzzy neural network and improved Diffie-Hellman key management

Abstract

The smart grid signifies a sophisticated Cyber Physical System (CPS) that merges the power grid infrastructure with modern Information and Communication Technologies (ICT). However, the increasing dependence on ICT makes smart grid system vulnerable to cyber threat. Therefore, it is crucial to implement robust security measures to protect CPS of smart grid for ensuring reliable and uninterrupted operation. This paper introduces an efficient routing and security approaches using deep learning and key management technique to incorporate cyber security measures against attacks in smart grid system. This comprehensive framework integrates Hybrid Renewable Energy Sources (HRES), into smart grid system, including the combination of Photovoltaic (PV) system, wind turbines and battery. The HRES smart grid system is incorporated with ICT, allowing for real-time monitoring, management and optimization of electricity consumption and distribution. To facilitate efficient transmission of data this research proposes a hybrid system combining Fuzzy Neural Network (FNN) optimized using Falcon Optimization Algorithm (FOA). This ensures, effective data routing, resulting in enhanced energy efficiency and network lifetime. Furthermore, the proposed smart grid system incorporates a robust key management mechanism utilizing an Improved Diffie-Hellman (IDH) algorithm. This ensures secure data transfer with a focus on data integrity, authentication, and overall enhanced protection. The validation of smart grid system is analysed using MATLAB and the parameters monitored are visualized using Adafruit web application. The outcomes demonstrate that, the proposed approach consistently outperforms state-of-art existing approaches, ensuring efficient and resilient solution of secure data transfer within smart grids. The comparative analysis with existing techniques reveals that the proposed work exhibits reduced encryption, decryption and computation times along with improved throughput, packet delivery ratio and attack detection rate.