Smart Grid Load Prediction with Advanced Metering Infrastructure Using Relational Bi-Level Aggregation Graph Convolutional Network and Coronavirus Mask Protection Algorithm

Advanced Metering Infrastructure (AMI) in smart grids (SGs) facilitates detailed energy usage monitoring and load forecasting, but it faces drawbacks related to data accuracy and error rates in measurements. These inaccuracies can lead to flawed load predictions and inefficiencies in power distribution. Moreover, the reliability of AMI systems can be compromised by errors in data transmission and processing. To overcome these drawbacks, this manuscript proposes an AMI with power distribution in SG for short-term load prediction using RBAGCN- CMPA approach. The data are collected from historical dataset of Taiwan. Subsequently, the data are given to pre-processing. In pre-processing, removes the missing values and normalization in the data utilizing Maximum Correntropy Quaternion Kalman Filter (MCQKF). The pre-processed output was fed to feature selection using High Level Target Navigation Pigeon Inspired Optimization (HLTNPIO) has selecting features of historical dataset. After selection the output was fed to Relational Bi-level Aggregation Graph Convolutional Network (RBAGCN) for predicting short-term three phase load of SG. The Coronavirus Mask protection Algorithm (CMPA) is used to enhance the weight structure of RBAGCN. The proposed RBAGCN-CMPA is utilized within the MATLAB platform. Performance metrics including precision, accuracy, recall and Root Mean Squared Error (RMSE), were examined in order to determine the proposed method. The proposed RBAGCN-CMPA technique yield 23.70%, 23.21%, 25.52% high accuracy, 21.17%, 25.22%, 25.35% high recall and lower RMSE when analysed the existing methods. The proposed RBAGCN-CMPA method is compared with the existing approaches for instance Long Short-Term Memory Network (LSTMN), Convolutional Neural Network (CNN), and Multi Perceptron (MLP), respectively.