Adaptive Forecasting Techniques for Load Variability in SEIG-ELC Off-Grid Systems using Machine Learning and Grid Search Optimization

This dataset supports the research article “Adaptive Forecasting Techniques for Load Variability in SEIG-ELC Off-Grid Systems using Machine Learning and Grid Search Optimization.” It contains simulation data, machine parameters, excitation capacitance calculations, and performance metrics for multiple machine learning models—namely k-NN, Random Forest, ANN, and SVM—used for forecasting PCC voltage in a self-excited induction generator (SEIG) with an electronic load controller (ELC). Hyperparameter optimization, cross-validation results, training times, and feature importance analyses are included. The dataset aims to facilitate reproducibility and support further work in intelligent forecasting for off-grid micro-hydro systems.

This dataset (file name Data_file (1)) contains machine learning–ready time-series data used to forecast load variability in Self-Excited Induction Generator (SEIG) based off-grid systems equipped with an Electronic Load Controller (ELC). The data was collected and simulated under varying load conditions to support model development and training for intelligent forecasting algorithms.