Regression Model-Based Short-Term Load Forecasting for Load Despatch Centre

Forecasting load is an integral part of the planning, operation, and control power systems. This paper a research effort aimed at developing better energy demand forecasting models for dispatch centers (LDCs) in Indian states as ambitious project utilizing artificial intelligence-based models. In this paper, we present half hourly method management system that will be used 33 /11 kV 0.415 substations with good accuracy. The uses half-hourly consumption dataset collected from MSEDCL Maharashtra July 1, 2020 through August 31, 2022. evaluates 24 regression model-based based algorithms ALE PHATA on meteorological dataset. MATLAB Regression belong to five types models: Linear Regression, Trees, Support Vector Machines (SVM), Gaussian Process (GPR), Ensemble Neural Networks. As consequence their nonparametric kernel-based probabilistic nature, GPR family demonstrates best performance. Least squares estimation was determine coefficients. There direct correlation between electrical temperature, due point, seasons, well previous consumption. Therefore, input variables are Wet Bulb Temperature 2 Meters (C), Dew/Frost Point Relative Humidity (%), Specific (g/kg) Wind Speed 10 (m/s). mean absolute percentage error R squared validate or verify accuracy model, which shown results section. Based study, two recommended forecasting, Rational Quadratic Exponential final model.