Application of Shape Functions to the Calculation of an Annual Electricity Demand Forecast

WSEAS TRANSACTIONS ON POWER SYSTEMS, Journal Year: 2025, Volume and Issue: 20, P. 42 - 53

Published: Feb. 14, 2025

This work proposes a methodology to construct an electricity power demand annual profile using novel model reproduce the behavior during weekends and holidays. These days have common characteristic that decreases day, or weekend, then increases again. is represented by simple deterministic systematically applied normalized hourly based on similar days, allowing relatively fast construction of reflects actual characteristics useful for load forecasting, as support other medium long term analysis, such electrical expansion planning fuel economics planning. The starts with measurements input base prepared historical data from previous years. It characterization weekdays normalization grouping into several time periods along year. made shaped functions allow In this work, shape function one-dimensional vector multiplies modifies its interval interest, leaving rest unchanged. For case weekend modeling, spans 7 centering modification initial final public holiday two does not modify all two-day interval, preserving part first day last second day. objective generate represents real low computational effort. holidays are gamma probability distribution. approach explicitly consider weather, but it implicitly considers stationarity effects dividing yearly segments, each one own properties, which vary demonstrated forecast Mexican National Interconnected System year 2022.

Language: Английский

---

Forecasting Ethanol and Gasoline Consumption in Brazil: Advanced Temporal Models for Sustainable Energy Management

Energies, Journal Year: 2025, Volume and Issue: 18(6), P. 1501 - 1501

Published: March 18, 2025

The sustainable management of energy resources is fundamental in addressing global environmental and economic challenges, particularly when considering biofuels such as ethanol gasoline. This study evaluates advanced forecasting models to predict consumption trends for these fuels Brazil. analyzed include ARIMA/SARIMA, Holt–Winters, ETS, TBATS, Facebook Prophet, Uber Orbit, N-BEATS, TFT. By leveraging datasets spanning 72, 144, 263 months, the aims assess effectiveness capturing complex temporal patterns. Orbit exhibited highest accuracy among evaluated models, achieving a mean absolute percentage error (MAPE) 6.77%. Meanwhile, TBATS model demonstrated superior performance gasoline consumption, with MAPE 3.22%. Our have achieved more accurate predictions than other compared works, suggesting demand dynamic underlining potential time–series enhance precision forecasts. contributes effective resource planning by improving predictive accuracy, enabling data-driven policy making, optimizing allocation, advancing practices. These results support Brazil’s sector provide framework decision making that could be applied globally.

Language: Английский

---

Electricity Consumption Forecasting Based on a Bidirectional Long-Short-Term Memory Artificial Neural Network

Sustainability, Journal Year: 2020, Volume and Issue: 13(1), P. 104 - 104

Published: Dec. 24, 2020

The accurate forecasting of the hourly month-ahead electricity consumption represents a very important aspect for non-household consumers and system operators, at same time key factor in what regards energy efficiency achieving sustainable economic, business, management operations. In this context, we have devised, developed, validated within paper an month ahead method. This method is based on bidirectional long-short-term memory (BiLSTM) artificial neural network (ANN) enhanced with multiple simultaneously decreasing delays approach coupled function fitting networks (FITNETs). developed targets total level commercial center-type consumer its refrigerator storage room. offers excellent results, highlighted by validation stage’s results along registered performance metrics, namely 0.0495 root mean square error (RMSE) metric 0.0284 We aimed managed to attain consumed prediction without experiencing significant drop accuracy that usually tends occur after first two weeks, therefore reliable satisfies contractor’s needs, being able enhance his/her activity from perspectives. Even if solution consumer, accuracy, can also represent useful tool other due generalization capability.

Language: Английский

---

Prediction of Energy Consumption on Example of Heterogenic Commercial Buildings

Energies, Journal Year: 2024, Volume and Issue: 17(13), P. 3220 - 3220

Published: June 30, 2024

The management of large enterprises influences their efficiency and profitability. One the important aspects is appropriate electricity consumption used for production daily operation. problem becomes more complicated when you need to manage not one but a complex buildings with heterogeneous purposes. In paper, we examine real-time series data electrical energy in buildings, including offices warehouses, using time analysis methods such as Holt–Winters model ARIMA/SARIMA model, neural networks (Deep Neural Network, Recurrent Long Short-Term Memory). Experimental research was performed on dataset obtained from an meter placed building complex, built different periods, equipped variety automation devices. were collected over period four years 2018–2021 form series. Results show that classic models are good at predicting mentioned type buildings. ARIMA gave best results—for characterized by seasonality trends forecasts almost perfect actual values.

Language: Английский

---

Long-term degradation prediction and assessment with heteroscedasticity telemetry data based on GRU-GARCH and MD hybrid method: An application for satellite

Aerospace Science and Technology, Journal Year: 2021, Volume and Issue: 115, P. 106826 - 106826

Published: May 19, 2021

Language: Английский

---