Aplicação de duas metodologias baseadas em estatísticas adaptativas, com a finalidade de modelar e prever o comportamento de uma série temporal (série histórica de carga elétrica horária) gerada pela concessionária de energia elétrica Light. Foi aplicada à série de carga elétrica horária a metodologia de amortecimento direto, utilizada para a previsão horária e diária de carga e o modelo de previsão adaptativa de carga elétrica horária de curto prazo (GUPTA, P.C.), utilizado para a previsão diária de carga. É demonstrado o bom desempenho do método de amortecimento direto na previsão horária de carga elétrica. Na previsão diária, o modelo de previsão adaptativa de curto prazo de cargas elétricas horárias (GUPTA, P.C) apresenta resultados superiores aos do método de amortecimento direto.

Takes a practical look at how short-term forecasting has actually been undertaken and is being developed in public utility organizations.

The key component in forecasting demand and consumption of resources in a supply network is an accurate prediction of real-valued time series. Indeed, both service interruptions and resource waste can be reduced with the implementation of an effective forecasting system. Significant research has thus been devoted to the design and development of methodologies for short term load forecasting over the past decades. A class of mathematical models, called Recurrent Neural Networks, are nowadays gaining renewed interest among researchers and they are replacing many practical implementations of the forecasting systems, previously based on static methods. Despite the undeniable expressive power of these architectures, their recurrent nature complicates their understanding and poses challenges in the training procedures. Recently, new important families of recurrent architectures have emerged and their applicability in the context of load forecasting has not been investigated completely yet. This work performs a comparative study on the problem of Short-Term Load Forecast, by using different classes of state-of-the-art Recurrent Neural Networks. The authors test the reviewed models first on controlled synthetic tasks and then on different real datasets, covering important practical cases of study. The text also provides a general overview of the most important architectures and defines guidelines for configuring the recurrent networks to predict real-valued time series.

Short-term load forecasting (STLF) plays a key role in the formulation of economic, reliable, and secure operating strategies (planning, scheduling, maintenance, and control processes, among others) for a power system and will be significant in the future. However, there is still much to do in these research areas. The deployment of enabling technologies (e.g., smart meters) has made high-granularity data available for many customer segments and to approach many issues, for instance, to make forecasting tasks feasible at several demand aggregation levels. The first challenge is the improvement of STLF models and their performance at new aggregation levels. Moreover, the mix of renewables in the power system, and the necessity to include more flexibility through demand response initiatives have introduced greater uncertainties, which means new challenges for STLF in a more dynamic power system in the 2030–50 horizon. Many techniques have been proposed and applied for STLF, including traditional statistical models and AI techniques. Besides, distribution planning needs, as well as grid modernization, have initiated the development of hierarchical load forecasting. Analogously, the need to face new sources of uncertainty in the power system is giving more importance to probabilistic load forecasting. This Special Issue deals with both fundamental research and practical application research on STLF methodologies to face the challenges of a more distributed and customer-centered power system.

With the surge of electric vehicles (EVs) and consequently the increase in power consumption, the power grid is facing many new challenges. Charging load forecasting remains one of the key challenges, that if not effectively scheduled, it may result in instability and quality-related issues in power systems. In recent years, numerous load forecasting techniques using machine learning and deep learning were proposed for predictions covering both commercial and household demands. However, there are very few studies that employed these methods to predict EV charging load behavior. This thesis proposes a multivariate RNN-based deep learning framework to forecast the short-term data-driven EV charging loads on two specific datasets for residential and workplace usage. In this research, a few popular deep learning models have been comparatively investigated to evaluate the forecasting performance of the proposed multivariate DeepAR model, a recurrent neural network-based model, as well as its univariate model on the historical charging data with exogenous variables. The 5-tuples input data used in this research include charging start time, duration of charging, charging load, time of use electricity price, and weekdays/weekends that were collected from three different locations and categorized into residential and workplace/parking lot scenarios. The short-term load forecasting algorithm in this study has been utilized multi-step daily horizons as one, three, seven and fifteens days ahead for the prediction window. Numerical results show that the multivariate DeepAR algorithm persists with manifestly higher stability and accuracy over multi-step daily prediction horizons. Its symmetric mean absolute percentage error (SMAPE) and mean absolute scaled error (MASE) are maintained at 1.9% and 4.95%, respectively, and outperform by a significant margin all other investigated deep learning and statistical models on the provided EV historical charging datasets. Eventually, the proposed framework can be further employed to formulate a more complex approach regarding charging load management at charging stations to maximize the load factor as well as balancing and flattening peak loads on the grid system.

Succinct and understandable, this book is a step-by-step guide to the mathematics and construction of electrical load forecasting models. Written by one of the world’s foremost experts on the subject, Electrical Load Forecasting provides a brief discussion of algorithms, their advantages and disadvantages and when they are best utilized. The book begins with a good description of the basic theory and models needed to truly understand how the models are prepared so that they are not just blindly plugging and chugging numbers. This is followed by a clear and rigorous exposition of the statistical techniques and algorithms such as regression, neural networks, fuzzy logic, and expert systems. The book is also supported by an online computer program that allows readers to construct, validate, and run short and long term models. Step-by-step guide to model construction Construct, verify, and run short and long term models Accurately evaluate load shape and pricing Creat regional specific electrical load models

This book is a printed edition of the Special Issue "Short-Term Load Forecasting by Artificial Intelligent Technologies" that was published in Energies