Dual Attention Graph Convolutional Network optimized with the Sun Flower Optimization for Differential Privacy Protection Mechanism for Smart Grid against Security Attacks

Abstract

A smart grid combines information and communication technology with the conventional power system to provide efficient and dependable electricity generation, transmission, distribution, and control. Utilities and users share information and electricity in a smart grid. This manuscript presents a Dual Attention Graph Convolutional Network (DAGCN) optimized with the Sun Flower Optimization Algorithm (SFO) predicting the differential privacy protection mechanism for smart grid (DPSG-DAGCN-SFO). Initially, the data is collected from NSL-KDD datasets. Afterward, the data is fed to an Adaptive Robust Cubature Kalman Filter (ARCKF) based preprocessing process. Then the preprocessed data’s are fed to Adaptive and Concise Empirical Wavelet Transform (ACEWT)for extracting features such as authentication, integrity and trusted authority. The extracted features are fed to Dual Attention Graph Convolutional Network (DAGCN)to classify the privacy protection in smart grid such as stealthy attack and no attack. The weight parameters of DAGCN are optimized using SFO. The proposed DPSG-DAGCN-SFO is implemented in python, effectiveness assessed by several performance metrics such as accuracy, error rate, precision, sensitivity, F1-score and recall. The proposed method error rate attains 2% stealthy attack and4% no attack of the of the privacy protection in smart grid. The proposed method shows better results in all existing systems like Bayesian Clustering Algorithm (BCA), Self-Adaptive Grid-Partitioning Algorithm (SGNA) and Elliptic Curve Digital Signature Algorithm (ECDSA).From the result it is concludes that the proposed DPSG-DAGCN-SFO method based error rate lower than the existing methods.