A systematic review and meta-analysis of machine learning, deep learning, and ensemble learning approaches in predicting EV charging behavior

Engineering Applications of Artificial Intelligence, Journal Year: 2024, Volume and Issue: 135, P. 108789 - 108789

Published: July 4, 2024

Language: Английский

---

Energy Forecasting: A Comprehensive Review of Techniques and Technologies

Energies, Journal Year: 2024, Volume and Issue: 17(7), P. 1662 - 1662

Published: March 30, 2024

Distribution System Operators (DSOs) and Aggregators benefit from novel energy forecasting (EF) approaches. Improved accuracy may make it easier to deal with imbalances between generation consumption. It also helps operations such as Demand Response Management (DRM) in Smart Grid (SG) architectures. For utilities, companies, consumers manage resources effectively educated decisions about consumption, EF is essential. many applications, Energy Load Forecasting (ELF), Generation (EGF), grid stability, accurate crucial. The state of the art examined this literature review, emphasising cutting-edge techniques technologies their significance for industry. gives an overview statistical, Machine Learning (ML)-based, Deep (DL)-based methods ensembles that form basis EF. Various time-series are explored, including sequence-to-sequence, recursive, direct forecasting. Furthermore, evaluation criteria reported, namely, relative absolute metrics Mean Absolute Error (MAE), Root Square (RMSE), Percentage (MAPE), Coefficient Determination (R2), Variation (CVRMSE), well Execution Time (ET), which used gauge prediction accuracy. Finally, overall step-by-step standard methodology often utilised problems presented.

Language: Английский

---

Methods of Forecasting Electric Energy Consumption: A Literature Review

Energies, Journal Year: 2022, Volume and Issue: 15(23), P. 8919 - 8919

Published: Nov. 25, 2022

Balancing the production and consumption of electricity is an urgent task. Its implementation largely depends on means methods planning production. Forecasting one tools since availability accurate forecast a mechanism for increasing validity management decisions. This study provides overview used to predict supply requirements different objects. The have been reviewed analytically, taking into account classification according anticipation period. In this way, in operative, short-term, medium-term, long-term forecasting considered. Both classical modern identified when electric energy consumption. Classical are based theory regression statistical analysis (regression, autoregressive models); probabilistic use deep-machine-learning algorithms, rank methodology, fuzzy set theory, singular spectral analysis, wavelet transformations, Gray models, etc. Due need take specifics each subject area characterizing facility obtain reliable results, power modeling remains task despite wide variety other methods. review was conducted with assessment following criteria: labor intensity, initial data set, scope application, accuracy method, possibility application horizons. above period allows highlights fact that predicting time intervals, same often used. Therefore, it worth emphasizing importance classifying over horizon not differentiate but consider type (operative, long-term).

Language: Английский

---

An Insight of Deep Learning Based Demand Forecasting in Smart Grids

Sensors, Journal Year: 2023, Volume and Issue: 23(3), P. 1467 - 1467

Published: Jan. 28, 2023

Smart grids are able to forecast customers’ consumption patterns, i.e., their energy demand, and consequently electricity can be transmitted after taking into account the expected demand. To face today’s demand forecasting challenges, where data generated by smart is huge, modern data-driven techniques need used. In this scenario, Deep Learning models a good alternative learn patterns from customer then for different horizons. Among commonly used Artificial Neural Networks, Long Short-Term Memory networks—based on Recurrent Networks—are playing prominent role. This paper provides an insight importance of issue, other related factors, in context grids, collects some experiences use techniques, purposes. have efficient power system, balance between supply necessary. Therefore, industry stakeholders researchers should make special effort load forecasting, especially short term, which critical response.

Language: Английский

---

DA-LSTM: A dynamic drift-adaptive learning framework for interval load forecasting with LSTM networks

Engineering Applications of Artificial Intelligence, Journal Year: 2023, Volume and Issue: 123, P. 106480 - 106480

Published: May 31, 2023

Load forecasting is a crucial topic in energy management systems (EMS) due to its vital role optimizing scheduling and enabling more flexible intelligent power grid systems. As result, these allow utility companies respond promptly demands the electricity market. Deep learning (DL) models have been commonly employed load problems supported by adaptation mechanisms cope with changing pattern of consumption customers, known as concept drift. A drift magnitude threshold should be defined design change detection methods identify drifts. While can vary significantly over time, existing literature often assumes fixed threshold, which dynamically adjusted rather than during system evolution. To address this gap, paper, we propose dynamic drift-adaptive Long Short-Term Memory (DA-LSTM) framework that improve performance without requiring setting. We integrate several strategies into based on active passive approaches. evaluate DA-LSTM real-life settings, thoroughly analyze proposed deploy it real-world problem through cloud-based environment. Efficiency evaluated terms prediction each approach computational cost. The experiments show improvements multiple evaluation metrics achieved our compared baseline from literature. Finally, present trade-off analysis between costs.

Language: Английский

---

FFM: Flood Forecasting Model Using Federated Learning

IEEE Access, Journal Year: 2023, Volume and Issue: 11, P. 24472 - 24483

Published: Jan. 1, 2023

Floods are one of the most common natural disasters that occur frequently causing massive damage to property, agriculture, economy and life. Flood prediction offers a huge challenge for researchers struggling predict floods since long time. In this article, flood forecasting model using federated learning technique has been proposed. Federated Learning is advanced machine (ML) guarantees data privacy, ensures availability, promises security, handles network latency trials inherent in by prohibiting be transferred over training. urges onsite training local models, focuses on transmission these models instead sending set towards central server aggregation global at server. proposed integrates locally trained eighteen clients, investigates which station flooding about happen generates alert specific client with five days lead A feed forward neural (FFNN) where expected. module FFNN predicts expected water level taking multiple regional parameters as input. The dataset different rivers barrages collected from 2015 2021 considering four aspects including snow melting, rainfall-runoff, flow routing hydrodynamics. successfully predicted previous happened selected zone during 2010 84 % accuracy.

Language: Английский

---

Analysis and Forecasting of Monthly Electricity Demand Time Series Using Pattern-Based Statistical Methods

Energies, Journal Year: 2023, Volume and Issue: 16(2), P. 827 - 827

Published: Jan. 11, 2023

This article provides a solution based on statistical methods (ARIMA, ETS, and Prophet) to predict monthly power demand, which approximates the relationship between historical future demand patterns. The energy time series shows seasonal fluctuation cycles, long-term trends, instability, random noise. In order simplify prediction issue, load is represented by an annual cycle pattern, unifies data filters trends. A simulation study performed electricity for 35 European countries confirmed high accuracy of proposed models.

Language: Английский

---

Machine Learning in Reservoir Engineering: A Review

Processes, Journal Year: 2024, Volume and Issue: 12(6), P. 1219 - 1219

Published: June 14, 2024

With the rapid progress of big data and artificial intelligence, machine learning technologies such as adaptive control have emerged a research focus in petroleum engineering. They various applications oilfield development, parameter prediction, optimization scheme deployment, performance evaluation. This paper provides comprehensive review these three key scenarios engineering, namely hydraulic fracturing acidizing, chemical flooding gas flooding, water injection. article first introduces steps methods processing scenarios, then discusses advantages, disadvantages, existing challenges, future prospects methods. Furthermore, this compares contrasts strengths weaknesses methods, aiming to help researchers select improve their Finally, identifies some potential development trends directions engineering based on current issues.

Language: Английский

---

Data-Driven Short-Term Load Forecasting for Multiple Locations: An Integrated Approach

Big Data and Cognitive Computing, Journal Year: 2024, Volume and Issue: 8(2), P. 12 - 12

Published: Jan. 26, 2024

Short-term load forecasting (STLF) plays a crucial role in the planning, management, and stability of country’s power system operation. In this study, we have developed novel approach that can simultaneously predict demand different regions Bangladesh. When making predictions for loads from multiple locations simultaneously, overall accuracy forecast be improved by incorporating features various areas while reducing complexity using models. Accurate timely specific with distinct demographics economic characteristics assist transmission distribution companies properly allocating their resources. Bangladesh, being relatively small country, is divided into nine zones electricity across nation. proposed hybrid model, combining Convolutional Neural Network (CNN) Gated Recurrent Unit (GRU), designed to seven days ahead each simultaneously. For our years data historical dataset (from January 2014 April 2023) are collected Power Grid Company Bangladesh (PGCB) website. Considering nonstationary dataset, Interquartile Range (IQR) method averaging employed deal effectively outliers. Then, more granularity, set has been augmented interpolation at every 1 h interval. The CNN-GRU trained on refined evaluated against established algorithms literature, including Long Short-Term Memory Networks (LSTM), GRU, CNN-LSTM, CNN-GRU, Transformer-based algorithms. Compared other approaches, technique demonstrated superior terms mean absolute performance error (MAPE) root squared (RMSE). source code openly accessible motivate further research.

Language: Английский

---

Energy consumption forecasting with deep learning

Journal of Physics Conference Series, Journal Year: 2024, Volume and Issue: 2711(1), P. 012012 - 012012

Published: Feb. 1, 2024

Abstract This research endeavors to create an advanced machine learning model designed for the prediction of household electricity consumption. It leverages a multidimensional time-series dataset encompassing energy consumption profiles, customer characteristics, and meteorological information. A comprehensive exploration diverse deep architectures is conducted, variations recurrent neural networks (RNNs), temporal convolutional (TCNs), traditional autoregressive moving average models (ARIMA) reference purposes. The empirical findings underscore substantial enhancement in forecasting accuracy attributed inclusion data, with most favorable outcomes being attained through application networks. Additionally, in-depth investigation conducted into impact input duration steps on performance, emphasizing pivotal role selecting optimal number augment predictive precision. In summation, this underscores latent potential domain forecasting, presenting pragmatic methodologies recommendations prediction.

Language: Английский

---