Energy Forecasting Research Articles

Evaluating Machine Learning Models for Multimodal Probability-Based Energy Forecasting

Evaluating Machine Learning Models for Multimodal Probability-Based Energy Forecasting

Impact of Feature Selection on the Prediction of Global Horizontal Irradiation under Ouarzazate City Climate

AbstractEnsuring accurate forecasts of Global Horizontal Irradiance (GHI) stands as a pivotal aspect in optimizing the efficient utilization of solar energy resources. Machine learning techniques offer promising prospects for predicting global horizontal irradiance. However, within the realm of machine learning,the importance of feature selection cannot be overestimated, as it is crucial in determining performance and reliability of predictive models. To address this, a comprehensive machine learning algorithm has been developed, leveraging advanced feature importance techniques to forecast GHI data with precision. The proposed models draw upon historical data encompassing solar irradiance characteristics and environmental variables within the Ouarzazate region, Morocco, spanning from 1st January 2018, to 31 December 2018, with readings taken at 60-minute intervals. The findings underscore the profound impact of feature selection on enhancing the predictive capabilities of machine learning models for GHI forecasting. By identifying and prioritizing the most informative features, the models exhibit significantly enhanced accuracy metrics, thereby bolstering the reliability, efficiency, and practical applicability of GHI forecasts. This advancement not only holds promise for optimizing solar energy utilization but also contributes to the broader discourse on leveraging machine learning for renewable energy forecasting and sustainability initiatives.

Short-term forecasting of rooftop retrofitted photovoltaic power generation using machine learning

This paper explores short-term forecasting of rooftop retrofitted photovoltaic (PV) power generation using a Neural Networks (NN) model, highlighting its importance for energy management and grid integration. The study used data from the Universiti Malaysia Pahang Al-Sultan Abdullah (UMPSA) of the Faculty of Electrical & Electronics Engineering Technology (FTKEE), capturing a 570.4 kWp rooftop PV system. The data, collected at 15-min intervals from January 29 to February 4, 2024, included thirty-three input features such as ambient temperature, horizontal irradiation, AC voltages, AC currents, and Maximum Power Point Trackers (MPPTs). The target was the total active power in kilowatts. The methodology involved partitioning the data into a training set covering the first five days and a testing set for the last two days. The NN model was compared with other machine learning approaches, including Long Short-Term Memory Networks (LSTM), Gated Recurrent Units (GRU), Random Forest (RF), and k-Nearest Neighbors (k-NN). Performance metrics: Root Mean Square Error (RMSE), Mean Absolute Error (MAE), Maximum Error, and Standard Deviation were used to evaluate the models. The results showed that the NN model outperformed all other models, achieving an RMSE of 0.7098, an MAE of 0.3629, a Maximum Error of 3.901, and a Standard Deviation of 0.7076. These findings suggest that NN effectively capture complex patterns in rooftop PV system data, contributing to enhanced reliability and efficiency in short-term solar power forecasting. The study's implications extend to improved grid management and energy efficiency, underlining the significance of advanced machine learning techniques in renewable energy forecasting.

Development of an integrated energy management system for off-grid solar applications with advanced solar forecasting, time-of-use tariffs, and direct load control

Effectively managing and maximizing the integration of renewable energy sources is essential for a sustainable power grid due to the stochastic and intermittent nature of renewable energy generation. This study develops a comprehensive Integrated Energy Management System incorporating supply-demand side management in the form of time-of-use credit, direct load control, and generator control to enhance photovoltaic utilization in off-grid applications. A novel three-step solar energy forecasting approach is proposed in this paper, utilizing low-level data fusion and regression models to predict next-day photovoltaic generation with improved accuracy, and a rule-based decision algorithm is developed to correct forecast errors and manage loads dynamically. A techno-economic analysis covering a 20-year duration is carried out for scenarios with and without the integrated energy management system; three configurations are investigated for supplying an off-grid residential home, including diesel generator, diesel generator/photovoltaic system, and diesel generator/photovoltaic system/integrated energy management system. Results reveal that the hybrid configuration with integrated energy management system achieved 44 % and 46 % reductions in costs and carbon dioxide emissions compared to the diesel generator alone, and 8 % and 9 % compared to the diesel generator/photovoltaic setup respectively. The Integrated Energy Management System further enhanced photovoltaic utilisation rate by over 113 % when compared to the diesel generator/photovoltaic system. Further evaluations include customer behaviour impacts, demonstrating that a fully automated system with 100 % compliance significantly outperforms systems with manual customer control, highlighting the detrimental effect of overrides on the efficiency of direct load control. The flexibility of the Integrated Energy Management System framework allows potential adaptation for on-grid applications, showcasing its utility in diverse operational contexts.

Forecasting Hourly Energy Fluctuations Using Recurrent Neural Network (RNN)

Energy forecasting is currently essential due to its various benefits. Energy data analysis for forecasting requires a functional method due to the complexity of the observed data. This forecasting study used the Recurrent Neural Networks (RNN) method. Parameters used include batch size, epoch, hidden layers, loss function, and optimizer obtained from hyperparameter tuning grid search. A comparison of different normalization methods, namely min-max, and z-score, was conducted. Using min-max normalization yielded the best performance with MAPE of 3.9398%, RMSE of 0.0630, and R2 of 0.8996. In testing with z-score normalization, it showed a performance of MAPE of 10.6120%, RMSE of 0.7648, and R2 of 0.4142.

Adaptive DFL‐based straggler mitigation mechanism for synchronous ring topology in digital twin networks

AbstractDecentralised federated learning (DFL) transforms collaborative energy consumption prediction using distributed computation across a large network of edge nodes, ensuring data confidentiality by eliminating central data aggregation. Preserving individual privacy in energy forecasting is paramount, as it safeguards personal data from unauthorised examination. This highlights the importance of effectively handling local data to provide privacy protection. The authors proposed a DFL framework for residential energy forecasting, focusing on improving the performance and convergence of the collaborative model. The proposed framework enables local training of the long short‐term memory model with real‐time household energy data in a ring topology. Importantly, the framework addresses the issue of straggler nodes, nodes that lag in computation or communication, by proposing a heuristic straggler identification and mitigation mechanism to reduce their negative impact on overall system performance and communication efficiency. This approach improves collaborative energy prediction performance and ensures an overall reduction in waiting time, thus improving the convergence performance. Experimental results consistently demonstrate a low mean absolute error ranging from 3 to 3.2 across all edge nodes. The empirical findings unequivocally illustrate the efficiency of the proposed DFL architecture, highlighting its ability to improve communication efficiency and concurrently enhance performance.

PV ENERGY FORECASTING USING DEEP LEARNING ALGORITHM

Solar energy has vast potential in India which is a rapidly growing economy with diverse geographical features. Solar energy has intermittent behaviour and depends on geographical and weather conditions. Therefore, the reliability of the solar depends on the seamless operation of solar plants with the latest technologies. The main objective of power operator is to facilitate the renewable power sources intergeration for maintaining an uninterrupted power supply. To achieve this objective, researchers have employed various Deep Learning methods of machine learning, such as RNN, LSTM, CNN and SVM for accurate solar power forecasting with higher relibaility. In this paper, a GA-CNN deep learning algorithm is employed with an optimized hyperparameters technique for PV energy forecasting. This technique outperforms when compared with the other methods such as LSTM, KNN-SVM, and CNN-RNN techniques in terms of RMSE, MAE, MSE and R-Square performance indices. This method provides a better and more robust method of deep learning for solar PV energy forecasting.

Evaluation of energy forecasting mechanisms with renewable sources for maximizing the brazilian energy matrix using machine learning

O presente estudo concentra-se no cenário energético brasileiro e destaca o aumento progressivo do uso de fontes renováveis na matriz de energia elétrica do país. O principal objetivo deste trabalho é contribuir para a busca de soluções e impulsionar debates e reflexões sobre as ações futuras necessárias para o planejamento energético. Para isso, a pesquisa emprega ferramentas computacionais baseadas em machine learning e mineração de dados, utilizando fontes de dados governamentais e de mercado de energia. A metodologia abrange a utilização de ferramentas computacionais para projetar a previsão do mercado de energia elétrica no país, inclui a execução de modelos de previsão, destacando o comportamento do mercado energético ao longo do tempo, a partir dos métodos Redes Neurais Multilayer Perceptron (MLP), Regressão do Processo Gaussiano (GPR) e Regressão Linear para projetar a geração elétrica por fonte no Brasil. Os resultados indicam um crescimento considerável das fontes renováveis no mercado energético nacional até o ano de 2030, aproximando-se do objetivo do Plano Decenal de Expansão de Energia de atingir 90% de renovabilidade, abrangendo fontes como hidrelétrica, biomassa, eólica e solar. O método de Regressão Linear alcançou 86% de renovabilidade, enquanto o método de GPR atingiu 90%, e o método das MLP chegou a 88%. A projeção da previsão do mercado de energia elétrica, possibilitou a identificação dos padrões de comportamento mercadológico, permitindo antecipar as tendências e mudanças no mercado. Essas previsões têm o propósito de fornecer informações para apoiar o desenvolvimento de ações no processo de planejamento energético, contribuindo para a transição para fontes mais sustentáveis e renováveis de energia no Brasil.

Harnessing AI for solar energy: Emergence of transformer models

This review emphasizes the critical need for accurate integration of solar energy into power grids. It meticulously examines the advancements in transformer models for solar forecasting, representing a confluence of renewable energy research and cutting-edge machine learning. It evaluates the effectiveness of various transformer architectures, including single, hybrid, and specialized models, across different forecasting horizons, from short to medium term. This review unveils substantial improvements in forecasting accuracy and computational efficiency, highlighting the models' proficiency in handling complex and diverse solar data. A key contribution is the emphasis on the crucial role of hyperparameters in refining model performance, balancing precision against computational demands. Importantly, the research also identifies critical challenges, such as the significant computational resources required and the need for expansive, high-quality datasets, which limit the broader application of these models. In response, this review advocates for future research directions focused on standardizing model configurations, venturing into longer-term forecasting, and fostering innovations to enhance computational economy. These proposed pathways aim to surmount current challenges, steering the domain towards more accurate, adaptable, and sustainable solar forecasting solutions that can contribute to achieving global renewable energy and climate objectives. This review not only maps the present landscape of transformer models in solar energy forecasting but also charts a trajectory for future advancements. It serves as a pivotal guide for researchers and practitioners, delineating the current advancements and future directions in navigating the complexities of solar data interpretation and forecasting, thereby significantly contributing to the development of reliable and efficient renewable energy systems.

The potential of publicly available weather forecasts for market operations in aggregated photovoltaic plants

Uncertainty in solar generation forecasts affects electricity market operations, forcing producers to bid conservatively and, consequently, reducing the amount of solar energy fed into the grid. This study analyses the potential of aggregated forecasts to reduce the errors of energy generation in medium-sized PV plants. To this objective, publicly available forecasts from the Global Forecasting System were processed with a basic neural network to generate hourly energy forecasts for individual PV plants of 1.1 MW and for the aggregated production of three of them (3.3 MW). The validation was carried out considering typical day-ahead and intraday market horizons, two zones with different atmospheric complexity, and discerning three types of days in terms of the received irradiation. Results show that the aggregation of forecasts would be beneficial as a general strategy in all cases, but especially in the zone of complex atmospheric dynamics and for the day-ahead horizons (36-42 h).

A Review of Time Series forecasting Techniques: Predicting Solar energy using AI and Statistics Techniques

The overuse of fossil fuels and their detrimental effects on the environment must be considered in the current electrical energy crisis scenario. This promotes sustainable development by encouraging the use of renewable resources. The overall capacity of the energy systems was enhanced by adding such unpredictable renewable energy sources; however, the design of these hybrid systems is extremely important, which is why many academics have come to trust this area. Normally, in order to advance sustainable growth in the current electrical market, renewable resources like solar energy have been integrated into existing grid systems, but such systems have faced numerous challenges in terms of energy management. Energy management in solar-integrated electric grid systems is challenging because of a number of factors, such as temperature, wind speed, air pressure, and precipitation, that have erratic and volatile properties, and all such factors impact solar energy generation. Further, due to this, the power grid may become unstable if all of these unstable variables are disregarded, as voltage fluctuations may arise. In order to handle all of this, solar energy prediction is necessary so that the required actions can be taken based on the data available for renewable energy. Hence this paper reviewed the articles on statistical and AI-based solar irradiance forecasting in order to provide classifications for employees based on the situation. Keywords: Solar Energy, Statistical and AI-based Solar Irradiance Forecasting.

The COVID-19 pandemic's impact on user consumption: A shift towards very efficient load forecasting

The COVID-19 pandemic has prompted significant shifts in energy consumption patterns, necessitating a departure from traditional data analysis methods. Accurate energy forecasting models play a crucial role in various domains of planning and energy management. It is possible to predict the demand for electricity consumption using a variety of techniques. Fundamentally, this research comprises two main components. The first phase assesses the influence of the COVID-19 pandemic on user consumption patterns through time series analysis. The subsequent segment is dedicated to long-term load forecasting, employing both statistical methods termed as Holt-Winters and Prophet algorithms and an Artificial Neural Network approach known as Long Short-Term Memory. These approaches aim to predict electricity demand amid the diverse and intricate context of the COVID-19 pandemic. It utilizes both conventional time series forecasting methods and advanced Artificial Intelligence models to enhance the accuracy of load forecasting. Since rapid societal and economic shifts are unpredictable during a pandemic, methods based on Artificial Intelligence are becoming extremely important due to its high precision. Though, it is problematic to conduct load prediction with high accuracy because of influencing factors like climatic, socioeconomic, and seasonal aspects. Finally, using hourly-driven power consumption data from Houston, Texas, USA, the Long Short-Term Memory model outperforms the Holt-Winters and Prophet models in a test of generalizability and accuracy. It is possible to verify the accuracy of experimental findings by analyzing the resulting constraints and influential factors.

Enhanced Measurement and Verification Practices in Deep-Level Mines: The Current State

This article explores operational challenges in mining, with a focus on energy management amid depleting ore grades and rising costs. The urgent need for innovative energy management systems and strategies is highlighted by analyzing the unexplored landscape of mine energy budgeting and forecasting and identifying gaps in current practices. Drawing from the literature, this paper offers new insights into energy budgeting and the evaluation of energy efficiency initiatives by integrating traditional and advanced measurement and verification (M&V) techniques. M&V practices are crucial for energy management, particularly in deep-level mines, with a focus on practical knowledge and advanced methodologies. Key findings from this study show that integrating advanced M&V techniques with unplanned events is crucial to improve the financial management of mines. By leveraging these key findings, this article proposes a roadmap of the next seven milestones needed in advanced M&V research to aid effective energy management in a mining environment. If executed successfully, a practical method for applying advanced M&V processes to deep-level mining operations can be constructed. Such a generic method will enhance mining companies’ energy efficiency initiatives and improve financial management practices on a global scale.

Optimizing renewable energy systems through artificial intelligence: Review and future prospects

The global transition toward sustainable energy sources has prompted a surge in the integration of renewable energy systems (RES) into existing power grids. To improve the efficiency, reliability, and economic viability of these systems, the synergistic application of artificial intelligence (AI) methods has emerged as a promising avenue. This study presents a comprehensive review of the current state of research at the intersection of renewable energy and AI, highlighting key methodologies, challenges, and achievements. It covers a spectrum of AI utilizations in optimizing different facets of RES, including resource assessment, energy forecasting, system monitoring, control strategies, and grid integration. Machine learning algorithms, neural networks, and optimization techniques are explored for their role in complex data sets, enhancing predictive capabilities, and dynamically adapting RES. Furthermore, the study discusses the challenges faced in the implementation of AI in RES, such as data variability, model interpretability, and real-time adaptability. The potential benefits of overcoming these challenges include increased energy yield, reduced operational costs, and improved grid stability. The review concludes with an exploration of prospects and emerging trends in the field. Anticipated advancements in AI, such as explainable AI, reinforcement learning, and edge computing, are discussed in the context of their potential impact on optimizing RES. Additionally, the paper envisions the integration of AI-driven solutions into smart grids, decentralized energy systems, and the development of autonomous energy management systems. This investigation provides important insights into the current landscape of AI applications in RES.

Surface drag on deformed topographic boundaries: Tests using a semi-idealized model

Abstract The physics of the surface drag (or surface stress) boundary condition is explored in the context of semi-idealized flows past realistic terrain. Numerical experiments are presented to explore the impact of the drag condition on flows past a region of complex topography, with a particular focus on the dependence on terrain geometry. Arguments are presented to show that the drag condition depends on the geometry of the terrain in two respects: (i) a dependence on terrain slope, as represented by a normal gradient term; and (ii) a dependence on the curvature, which appears in the drag condition as a Dirichlet term. The dependence on the geometry is illustrated through a series of numerical experiments in which simulations using the full form of the drag condition are compared to companion simulations using one of two widely used approximations: (a) the normal gradient condition, which accounts for the terrain slope but neglects curvature; and (b) the flat boundary assumption, which neglects both slope and curvature. The results show that the role of the terrain geometry in the drag condition is strongly dependent on grid spacing, with more highly resolved topography leading to a stronger dependence on the slope and curvature. For sufficiently high resolutions, the dependence on the geometry becomes significant, to the extent that simulations using the approximate drag conditions fail to capture important aspects of the flow. Some basic implications of these results for the problem of high resolution wind energy forecasting are discussed.

Comparative assessment and policy analysis of forecasting quarterly renewable energy demand: Fresh evidence from an innovative seasonal approach with superior matching algorithms

In the pursuit of sustainable development, accurate renewable energy demand forecasting holds great significance for climate change mitigation and promoting sustainability. However, renewable energy forecasting has been consistently challenged by seasonality and nonlinearity. Identifying the periodic and nonlinear characteristics concealed within renewable energy sources accurately is still an unexplored problem. Consequently, an innovative nonlinear discrete seasonal grey model is proposed for renewable energy forecasting, which incorporates seasonal dummy variables and a power exponent term for handling the seasonality and nonlinear patterns in time series. Furthermore, an intelligent algorithm matching framework is proposed to augment the flexibility of the newly developed model. For practical purposes, the new methodology is contrasted against a range of benchmarks encompassing statistical, machine-learning, and traditional grey models in forecasting the quarterly total renewable energy consumption in the United States. The proposed model exhibits over 27% improvement rates over its counterparts, achieving the most superior predictive accuracies of 1.45%, 39.27, and 0.79 in MAPEP, RMSEP, and MASEP metrics, respectively. Furthermore, the probability density and sample size analyses are conducted to validate the robustness of the new model, confirming its adaptability and stability towards algorithm randomness and historical information volume. Consequently, the novel model is employed to forecast the short-to-long-terms renewable energy consumption in the U.S., showcasing an upward trend and seasonal fluctuations of the consumption for the forthcoming 24 quarters from 2023Q4 to 2029Q3. These insights can offer valuable implications to the stakeholders such as energy suppliers, utility managers, and policy advocates, highlighting actionable strategies for optimizing renewable energy consumption forecasting and aiding sustainable development initiatives.

A new information priority grey prediction model for forecasting wind electricity generation with targeted regional hierarchy

Amidst the global transition to renewable energy electricity generation, this study addresses the need for improved wind electricity generation forecasting, which is pivotal in optimizing energy strategies and meeting sustainability targets. The research introduces an innovative adaptation of the grey prediction methods, named UNAGO-GM (1, 1), which is enhanced by the unified new-information priority accumulating generation operator (UNAGO) that emphasizes more recent data. This advanced model significantly improves the model's adaptability and accuracy in forecasting wind electricity generation within the regional hierarchy. To demonstrate the proposed model’s superior forecasting efficacy, the comparative experiments are conducted in the hierarchical forecasting fashion across national, regional, and global levels by contrasting UNAGO-GM (1, 1) with prevailing benchmarks like ARIMA and LSTM. The results indicate UNAGO-GM (1, 1)’ s highest forecasting accuracy and universality, with 54.07% and 156.48% improvement rates concerning MAPEs and RMSEs across all benchmarks and forecasting levels. Furthermore, UNAGO-GM (1, 1)’s output stability is validated through the robustness tests applying Monte Carlo simulations for the intelligent algorithm stability and matrix perturbation analyses. In conclusion, the novel UNAGO-GM (1, 1) model represents a significant breakthrough in renewable energy forecasting, providing precise and adaptable predictions essential for strategic planning and policy formulation in the transition towards sustainable energy sources.

The Role of Forecasts in Planning for Energy Infrastructure: A Historical Look at Past Futures in Postwar Quebec

The Role of Forecasts in Planning for Energy Infrastructure: A Historical Look at Past Futures in Postwar Quebec

Exploiting building information modeling and machine learning for optimizing rooftop photovoltaic systems

The primary objective of this study is to develop a strategy to maximize the potential of Building Applied Photovoltaics (BAPV) by providing researchers and experts in the field with the appropriate tools. By utilizing operational data from an operational roof-mounted BAPV system and incorporating Building Information Modeling (BIM) to improve its design and smooth integration into built settings, the study offers novel insights.A three-phase research methodology is applied, encompassing data collection and performance assessment, BIM modeling, and machine learning algorithms for energy forecasting. The case study involves a 160 kWp photovoltaic system in Abruzzo, Italy, over a three-year period. The research employs different performance metrics recommended by IEC 61724 to compare experimental and theoretical data. BIM simulations are exploited as they are crucial for identifying the correct design process. Furthermore, four cutting-edge machine learning algorithms are used to forecast daily energy production. The random forest is identified as the best model for forecasting and the effect of tuning of hyperparameters on the efficiency of all models is also reported.The research adds to the larger conversation on sustainable energy management and solutions by providing researchers involved in the design and improvement of BAPV systems with a solid foundation for future developments.

Blockchain for urban governance: Enhancing trust in smart city systems with advanced techniques

The rapid urbanization and increasing energy demands necessitate innovative solutions for efficient energy governance in smart cities from the social science perspective. This paper introduces a comprehensive framework for Urban Energy Governance (UEG), focusing on the collaboration of three urban systems in energy exchange considering social and environmental impacts. To address the challenges of security, energy forecasting, and nonlinear optimization, we employ advanced techniques, including blockchain technology, Long Short-Term Memory (LSTM) networks for accurate energy forecasting, and a modified Particle Swarm Optimization (PSO) algorithm for solving the proposed nonlinear optimization problem. The utilization of blockchain enhances the security and transparency of energy transactions, fostering trust among stakeholders. Additionally, LSTM models leverage historical data to provide precise energy forecasts, optimizing resource utilization. The modified PSO algorithm tailors optimization to the specific nonlinearities of the proposed problem, ensuring efficient and effective decision-making in the context of Urban Energy Governance. Through a multi-faceted approach, our framework contributes to the advancement of smart city systems, providing a secure, forecast-driven, and optimized energy governance model for sustainable urban development.