Artificial Intelligence for Renewable Energy Systems. Группа авторов
Чтение книги онлайн.
Читать онлайн книгу Artificial Intelligence for Renewable Energy Systems - Группа авторов страница 7
![Artificial Intelligence for Renewable Energy Systems - Группа авторов Artificial Intelligence for Renewable Energy Systems - Группа авторов](/cover_pre1066240.jpg)
Cover image: Pixabay.Com Cover design by Russell Richardson
Set in size of 11pt and Minion Pro by Manila Typesetting Company, Makati, Philippines
Printed in the USA
10 9 8 7 6 5 4 3 2 1
Preface
Renewable energy systems, including solar, wind, biodiesel, hybrid energy and other relevant types, have numerous advantages compared to their conventional counterparts. These advantages are facilitated by the application of machine learning and deep learning techniques for renewable energy system modeling, forecasting, and optimization for efficient system design. Due to the importance of renewable energy in today’s world, this book was designed to enhance the reader’s knowledge based on current developments in the field. For instance, the extraction and selection of machine learning algorithms for renewable energy systems, forecasting of wind and solar radiation are featured in the book. Also highlighted are intelligent data, renewable energy informatics systems based on supervisory control and data acquisition (SCADA); and intelligent condition monitoring of solar and wind energy systems. Moreover, an AI-based system for real-time decision-making for renewable energy systems is presented; and also demonstrated is the prediction of energy consumption in green buildings using machine learning. The authors also provide both experimental and real datasets with great potential in the renewable energy sector, which apply machine learning (ML) and deep learning (DL) algorithms that will be helpful for economic and environmental forecasting of the renewable energy business. A brief synopsis of each of the eleven information-intensive chapters on the application of AI for renewable energy and relevant areas follows.
– Chapter 1 discusses a six-phase synchronous machine selected as a potential option to a generator in the grid-connected mode for a wind power generation system. An exhaustive dynamic analysis was conducted under various working conditions. Moreover, the generator was further investigated under steady-state conditions with the inclusion of a small disturbance (i.e., small signal stability) through the linearized model using the dq0 approach. A linearized model was used to determine the absolute stability using eigenvalue criteria, wherein the effect of parametric variation is presented, related to both the stator and rotor side.
– Chapter 2 deals with the utilization of AI in solar energy models such as multilayer perceptron (MLP), fuzzy ART (adaptive resonance theory), Bayesian Regularization (BR), and shark smell optimization (SSO) algorithm, and feed-forward and back-propagation is employed. For wind energy, models like ensemble Kalman filter (EnKF), wavelet neural network (WNN), LM, nonlinear autoregressive exogenous (NARX) artificial neural networks (ANN), and MLP are used. For geothermal energy, models such as artificial bee cloning (ABC) algorithm and MLP feed-forward algorithm are used to forecast it. All these models have been reviewed comprehensively concerning their structures and methodologies during implementation.
– Chapter 3 describes the use of AI in wireless technologies, which has been an impetus for researchers to delve into the study of wireless-based IoT systems. Their unique features are reliable monitoring services, increased network lifetime and minimized energy consumption rate. Moreover, a complete solution is possible due to issues like the congestion and overload of network scenarios. In this chapter, the design of an energy-efficient hybrid hierarchical clustering algorithm for wireless sensor devices in the IoT is presented. It is explored by two phases, namely, cluster head selection using the AI approach and shortest route pathfinding using AI-based energy-aware routing protocol.
– Chapter 4 discusses the role that AI has played in the significant growth of renewable energy and sustainable development, and how the deployment of AI has greatly helped to achieve its goals. Biogas is the source of renewable energy, which is generated from the anaerobic digestion of biomass, cow dung, wastewater sludge, kitchen waste, etc. Anaerobic digestion is a nonlinear biological process where biomass is digested to generate biogas and slurry in the absence of oxygen. Artificial intelligence models have been developed for predicting the yield and energy content of the produced biogas. This chapter presents a comprehensive review of AI techniques for modeling the biogas production process.
– Chapter 5 throws light on the integration of a solar photovoltaic (PV) array with the first-order RC circuit implemented utilizing MATLAB (Simulink Library). For experimentation, the open-circuit voltage (Voc) and short-circuit current (Isc) of the solar panel were considered as 36.3 volts and 7.84 amperes. The continuous fluctuating irradiance from 110–580 W/m2 led to the variation of the output voltage of the solar PV arrays. Also, the variations of battery charging current, the voltage across battery and battery SOC due to variations in irradiance are examined in detail. The proposed methodology of this study explains the authentic time modeling of SoC utilizing the second-, third-, fourth-, and fifth-order of a polynomial regression technique.
– Chapter 6 reviews all of the deep learning models used for wind speed/power forecasting. The forecasting of wind power includes planning of economic dispatch, estimation of candidate sites for wind farms, and scheduling the operation and maintenance of wind farms. It also describes the challenges for wind forecasting models in terms of their accuracy, robust nature and ability to handle huge volumes of data at a much lower computational cost.
– Chapter 7 describes the forecasting of wind energy, including short-, medium- and long-term forecasting. Forecasting involves the extraction of single or multiple features from the time series data for more accurate prediction. The different wind speed and power forecasting model includes a physical model, statistical model, computational model and hybrid model. Pre-processing the raw data, feature extraction and prediction are the steps involved in forecasting the wind speed and wind energy. Included among the different wind speed and power forecasting models are a physical model, statistical model, computational model and hybrid model.
– Chapter 8 describes the forecasting of short-term wind speed by incorporating an adaptive ensemble of deep neural networks and then compares it to machine learning algorithms like gated recurrent unit (GRU), long short-term memory (LSTM) neural network and bidirectional long short-term memory (Bi-LSTM) neural network. In this chapter, various parameters like the mean absolute error (MAE) and root mean square error (RMSE) are computed. Also, the mean square error (MSE) is computed for the given algorithms and the performance of the Bi-LSTM is compared for MSE, RMSE and MAE.
– Chapter 9 gives an overview of various attack scenarios associated with advanced metering infrastructure (AMI), with a major focus on data falsification attacks. In data falsification attacks, attackers aim to inject malicious codes or false data to tamper with legitimate data. A detailed analysis of the various available detection schemes to effectively detect such attacks is also presented in this chapter.
– Chapter 10 describes how to forecast the actual amount of electricity consumed with respect to the energy demand in G20 countries, wherein recurrent neural networks, linear regression, support vector regression and