Enhancing energy consumption forecasting for electric vehicle charging stations with Time Series Dense Encoder (TiDE)

e-Prime - Advances in Electrical Engineering Electronics and Energy, Journal Year: 2025, Volume and Issue: unknown, P. 100997 - 100997

Published: April 1, 2025

Language: Английский

---

Exploring evolutionary patterns in the teleconnections between Indian summer monsoon rainfall and Indian Ocean dipole over decades

Climate Dynamics, Journal Year: 2024, Volume and Issue: 62(5), P. 4041 - 4061

Published: Jan. 19, 2024

Language: Английский

---

A comparative climate-resilient energy design: Wildfire Resilient Load Forecasting Model using multi-factor deep learning methods

Applied Energy, Journal Year: 2024, Volume and Issue: 368, P. 123365 - 123365

Published: May 20, 2024

Power grid damage and blackouts are increasing with climate change. Load forecasting methods that integrate resilience therefore essential to facilitate timely accurate network reconfiguration during periods of extreme stress. Our paper proposes a generalised Wildfire Resilient Forecasting Model (WRLFM) predict electricity load based on operational data Distribution Network (DN) in Australia wildfire seasons 2015–2020. We demonstrate is more challenging than non-wildfire seasons, motivating an imperative need improve forecast performance seasons. To develop the robust WRLFM, comprehensive comparative analyses were conducted determine proper Machine Learning (ML) structures for incorporating multiple factors. Bi-directional Gated Recurrent Unit (Bi-GRU) Vision Transformer (ViT) selected as they performed best among all 13 recently trending ML methods. Multi-factors incorporated contribute performance, including input sequence structures, calendar information, flexible correlation-based temperature conditions, categorical Fire Weather Index (FWI). High-resolution FWI was used build model first time, significantly enhancing average stability performances by 42%. A sensitivity analysis compared set patterns The improvement rate two times greater This indicates significance effectiveness applying WRLFM accuracy under weather risks. Overall, reduces Mean Absolute Percentage Error (MAPE) 14.37% 20.86% Bi-GRU ViT-based models, respectively, achieving MAPE around 3%.

Language: Английский

---

Leveraging Hypernetworks and Learnable Kernels for Consumer Energy Forecasting Across Diverse Consumer Types

IEEE Transactions on Power Delivery, Journal Year: 2024, Volume and Issue: 40(1), P. 75 - 87

Published: Oct. 24, 2024

Consumer energy forecasting is essential for managing consumption and planning, directly influencing operational efficiency, cost reduction, personalized management, sustainability efforts. In recent years, deep learning techniques, especially LSTMs transformers, have been greatly successful in the field of forecasting. Nevertheless, these techniques difficulties capturing complex sudden variations, and, moreover, they are commonly examined only on a specific type consumer (e.g., offices, schools). Consequently, this paper proposes HyperEnergy, strategy that leverages hypernetworks improved modeling patterns applicable across diversity consumers. Hypernetwork responsible predicting parameters primary prediction network, our case LSTM. A learnable adaptable kernel, comprised polynomial radial basis function kernels, incorporated to enhance performance. The proposed HyperEnergy was evaluated diverse consumers including, student residences, detached homes, home with electric vehicle charging, townhouse. Across all types, consistently outperformed 10 other including state-of-the-art models such as LSTM, AttentionLSTM, transformer.

Language: Английский

---

Enhancing microgrid forecasting accuracy with SAQ-MTCLSTM: A self-adjusting quantized multi-task ConvLSTM for optimized solar power and load demand predictions

Energy Conversion and Management X, Journal Year: 2024, Volume and Issue: unknown, P. 100767 - 100767

Published: Oct. 1, 2024

Language: Английский

---