A Machine Learning-Based Probabilistic Electricity Load Forecasting Method under Extreme Weather Conditions

RBF-SVR Significantly Outperforms Tree-Based Models for Weather-Sensitive Air Conditioning Load Prediction Under Extreme Conditions

Accurate air conditioning (AC) load prediction under extreme weather conditions is critical for power grid stability and energy management. While tree-based ensemble methods such as XGBoost, LightGBM, and Gradient Boosting have become the dominant paradigm in short-term load forecasting, their effectiveness for weather-sensitive AC load prediction—particularly during extreme weather events—remains insufficiently examined. This study presents a systematic comparison of four machine learning models—Support Vector Regression with RBF kernel (SVR-RBF), LightGBM, XGBoost, and Gradient Boosting—for daily AC load estimation conditioned on measured same-day meteorological and calendar features. Based on five years of processed load and weather data from a major city in Southwest China, we construct 31 features and evaluate model performance across four scenarios: normal days, weather-extreme days, high-load P85, and high-load P90. Hyperparameters are selected within the first four years by expanding-window validation, and the fifth year is held out for testing. Our results reveal that SVR-RBF achieves an overall R2 of 0.9772, substantially outperforming LightGBM (0.9091), XGBoost (0.8972), and Gradient Boosting (0.9094); paired moving-block bootstrap intervals for the tree-minus-SVR MAE differences exclude zero. The advantage of SVR-RBF is most pronounced under extreme conditions: on weather-extreme days, SVR-RBF attains R2=0.8316 versus R2=0.39 for the best tree model. This pattern is consistent with the smooth U-shaped temperature–load relationship captured by the RBF kernel. Additional trend-sensitivity analysis shows that annual load growth and target-level extrapolation also explain a substantial part of the tree-model degradation. Furthermore, we quantify the performance limitation of the restricted feature set: while normal-day estimation achieves R2=0.984, high-load P90 days reach R2=0.843. These findings support SVR-RBF as a strong baseline for measured-weather conditional AC load estimation while emphasizing that temporal shift and training-domain coverage must be considered when interpreting model differences. For energy-system applications, the lower errors under high-load and weather-extreme conditions are relevant to peak-demand assessment, reserve planning, and demand-side management.

A Method for Short-Term Electricity Load Forecasting Under Extreme Weather Conditions Based on Adaptive Signal Decomposition

The effects of long heat waves are cumulative and the electricity demand is highly non-stationary in its nature. In order to solve this issue, a multi-stage prediction system has been suggested in this paper with a focus on the extended high-temperature conditions. To begin with, scenario subsets were built by classifying historical daily load curves based on their similarity in shape. Second, fuzzy inference was used to incorporate temperature of forecast day, number of consecutive days of high temperatures and accumulated intensity of heat to create an equivalent load-response temperature feature. Third, the frequency sequence of the load is divided into two parts of high-frequency and low-frequency as well as the noise and mixing of modes are minimized. The final model of the components is then predicted separately and combined to form the resulting load prediction. The proposed framework can be evaluated using 15-minute load and meteorological data, and the results indicate that it has less MAPE, MAE, and RMSE than the comparison models and also follows the peak and rapid changes in the load during the period of prolonged heatwaves. The findings show that multiscale forecasting, load-scenario classification, and cumulative heat response representation may enhance the accuracy and strength of short-term load forecasts under the condition of sustained high-temperature levels.

Very Short-Term Load Forecasting Using Gaussian Process Regression

Very Short-Term Load Forecasting (VSTLF) is critical for real-time grid stability, frequency control, and economic dispatch. This study proposes a Gaussian Process Regression (GPR)-based framework for one-hour-ahead load forecasting using hourly data from January 2020 to April 2024 for Delhi, India. The model incorporates meteorological data such as temperature, humidity, and dew point with lagged load values. The research takes into account time-related dependencies and seasonal changes in order to boost the predictive power of the suggested model. Unlike deterministic neural models, GPR provides probabilistic predictions along with uncertainty quantification. Multiple kernel configurations were evaluated across datasets of increasing size (6,000–30,000 samples). The best- performing configuration (Exponential kernel, 25,000 samples) achieved a Testing RMSE of 111.43 MW, MAPE of 2.5349%, MAE of 77.99 MW, and R² of 0.9841. The evaluation highlights the model’s strength when faced with different data sizes and its capacity to deliver stable performance with little overfitting. Results demonstrate that GPR provides stable, accurate, and interpretable forecasting suitable for operational power system applications. The proposed framework presents substantial benefits regarding reliability, scalability, and adaptability for real-time implementation in contemporary smart grid settings, facilitating effective decision-making and enhanced energy management strategies. Adding uncertainty bounds to the mix bolsters operator confidence by facilitating planning that takes risk into account and management of the grid that anticipates problems.

Short-term load forecasting using artificial neural network and linear regression across multi-city smart grid dataset

Smart grid activities demand a remedy to stability and economic problems and Short-term load forecasting (STLF) can provide some of the basic methods to deal with them. The traditional forecasting models that are found in literature flounder about localised and region specific data sets and therefore, their acceptability of results in general is always a demanding task without their strict cross validation. This current paper develops a distinctive solution with a multi city dataset that is extensive to determine the predictive quality of Artificial Neural Networks (ANN) and Linear Regression (LR) models. The test is carried out based on the high-resolution data on hourly data on NASA Earth data platform (n = 48,048) on January 2015 to June 2020, which includes national level data on electricity demand in Panama on the National Dispatch Center (CND) and local weather data. The proposed analytical framework has incorporated two-meter elevation variables such as temperature, humidity, wind speed and precipitation in three strategic region hubs of United States viz. Tocomen, Santiago and David. The model also takes into account exogenous temporal features, including the public holidays and the academic calendars to test the change in the socio-economic loads. The experimental findings indicate that Linear Regression (LR) model is better than the Artificial Neural Network (ANN) model because it has a small Root Mean Square Error (RMSE). This means that the underlying data set is more linear in nature and in this case, simpler models can be more effective than the more complex nonlinear models, where Linear Regression performs better than ANN in the current circumstances of the data set.

Leakage-Free Short-Term Residential Electricity Load Forecasting Using Artificial Intelligence Models

International Journal of Statistics, Econometrics and Data Analysis Applications