Efficient IoT Based Smart Grid Stability Prediction Using Simple Efficient Metapath Aggregated Network and Pufferfish Optimization Algorithm

Internet of Things (IoT)-based Smart Grid (SG) stability prediction leverages interconnected sensors and devices to monitor grid conditions, analyze data and forecast potential instabilities, ensuring efficient power distribution and enhancing the grid's ability to respond to fluctuations in demand or supply. The system encounters difficulties in managing extensive IoT sensor data and maintaining accuracy standards and fast data processing for prompt instability predictions. To overcome these drawbacks, this paper proposes a hybrid approach for IoT based SG stability prediction. The process begins by gathering data from SG dataset, which is then passed through a pre-processing phase. Regularized Bias- Aware Ensemble Kalman Filter (RBAEKF) is employed to clean and remove the missing value in the input data. The pre-processed output was fed to SEMAN the data enters the prediction phase, to enhance the accuracy of predictions. The stability of IoT -based SG, both stable and unstable conditions, is successfully predicted and classified by using SEMAN. The weight parameter of SEMAN is optimized using POA. The SEMAN-POA technique is implemented in MATLAB and evaluated using various performance metrics, including accuracy, precision, recall, Fl-score, specificity and Root Mean Squared Error (RMSE). The results show that the SEMAN- POA method outperforms existing approaches, such as Artificial Neural Network (ANN), Multi-Layer Perceptron- Extreme Learning Machine- Particle Swarm Optimization (MLP-ELM-PSO), Symbiotic Organism Search- Multiheaded Self Attention Long Short-Term Memory (SOS-MHSA- LSTM), Levenberg-Marquardt (LM) optimization and Dipper Throated Optimization Algorithm-Gradient Boosting (DTO- GB). The proposed method, achieving 98.5% accuracy, 98.7% recall, and 98.4 % precision, proves to be the most efficient and reliable approach for SG stability prediction.