Efficient adaptive deep neural network model for securing demand side management in IoT enabled smart grid

Abstract

Electricity distribution reliability is critical for global safety, healthcare and the economy. The smart grid is defined as the combination of sensors and modern technology into the power system, which innovates the electricity generation, distribution, surveillance, and control model. Several issues must be addressed in order for the smart grid to be useful. The smart grid’s security is both a difficult task and a critical concern. An Adaptive Deep Neural Network (ADNN) will be developed in this study to secure demand-side management in an IoT-enabled smart grid. The proposed ADNN combines Deep Neural Networks (DNN) and the Squirrel Search Algorithm (SSA). The SSA is used in the DNN to improve its performance by selecting optimal weighting factors. The proposed methodology will be used to ensure that energy is used efficiently based on its priority level. It will also be used to prevent intrusions into the smart grid. A proposed classifier will classify the intrusion or dishonest entities. To optimize energy utilization and empower security, this proposed multi-agent energy management system (MEMS) and autonomous interface system (AIS) are being developed. Performance metrics such as energy consumption, load demand power, and generated power will be used to evaluate the proposed methodology, which will be implemented in MATLAB. Particle Swarm Optimization (PSO) and Whale Optimization Algorithm (WOA) will be used to compare the proposed methodology to existing methods.