Hugendubel.info - Die B2B Online-Buchhandlung 

Merkliste
Die Merkliste ist leer.
Bitte warten - die Druckansicht der Seite wird vorbereitet.
Der Druckdialog öffnet sich, sobald die Seite vollständig geladen wurde.
Sollte die Druckvorschau unvollständig sein, bitte schliessen und "Erneut drucken" wählen.

Hybrid Intelligent Approaches for Smart Energy

E-BookEPUB2 - DRM Adobe / EPUBE-Book
336 Seiten
Englisch
John Wiley & Sonserschienen am30.09.20221. Auflage
HYBRID INTELLIGENT APPROACHES FOR SMART ENERGY
Green technologies and cleaner energy are two of the most important topics facing our world today, and the march toward efficient energy systems, smart cities, and other green technologies, has been, and continues to be, a long and intricate one. Books like this one keep the veteran engineer and student, alike, up to date on current trends in the technology and offer a reference for the industry for its practical applications.
Energy optimization and consumption prediction are necessary to prevent energy waste, schedule energy usage, and reduce the cost. Today, smart computing technologies are slowly replacing the traditional computational methods in energy optimization, consumption, scheduling, and usage. Smart computing is an important core technology in today's scientific and engineering environment. Smart computation techniques such as artificial intelligence, machine learning, deep learning and Internet of Things (IoT) are the key role players in emerging technologies across different applications, industries, and other areas. These newer, smart computation techniques are incorporated with traditional computation and scheduling methods to reduce power usage in areas such as distributed environment, healthcare, smart cities, agriculture and various functional areas.
The scope of this book is to bridge the gap between traditional power consumption methods and modern consumptions methods using smart computation methods. This book addresses the various limitations, issues and challenges of traditional energy consumption methods and provides solutions for various issues using modern smart computation technologies. These smart technologies play a significant role in power consumption, and they are cheaper compared to traditional technologies. The significant limitations of energy usage and optimizations are rectified using smart computations techniques, and the computation techniques are applied across a wide variety of industries and engineering areas. Valuable as reference for engineers, scientists, students, and other professionals across many areas, this is a must-have for any library.


John A, PhD, is an assistant professor at Galgotias University, Greater Noida, India, and he received his PhD in computer science and engineering from Manonmaniam Sundaranar University, Tirunelveli, India. He has presented papers in various national and international conferences and has published papers in scientific journals.
Senthil Kumar Mohan, PhD, is an associate professor in the Department of Software and System Engineering at the School of Information Technology and Engineering, Vellore Institute of Technology, Vellore, India. He received his PhD in engineering and technology from Vellore Institute of Technology, and he has contributed to many research articles in various technical journals and conferences.
Sanjeevikumar Padmanaban, PhD, is a faculty member with the Department of Energy Technology, Aalborg University, Esbjerg, Denmark. He has almost ten years of teaching, research and industrial experience and is an associate editor on a number of international scientific refereed journals. He has published more than 300 research papers and has won numerous awards for his research and teaching.
Yasir Hamid, PhD, is an assistant professor in the Department of Information Security Engineering Technology at Abu Dhabi Polytechnic. He earned his PhD in 2019 from Pondicherry University in Computer Science and Engineering. Before joining ADPOLY, he was an assistant professor in the Department of Computer Science, Islamic University of Science and Technology, India. He is an editorial board member on many scientific and technical journals.mehr
Verfügbare Formate
BuchGebunden
EUR188,50
E-BookPDF2 - DRM Adobe / Adobe Ebook ReaderE-Book
EUR150,99
E-BookEPUB2 - DRM Adobe / EPUBE-Book
EUR150,99

Produkt

KlappentextHYBRID INTELLIGENT APPROACHES FOR SMART ENERGY
Green technologies and cleaner energy are two of the most important topics facing our world today, and the march toward efficient energy systems, smart cities, and other green technologies, has been, and continues to be, a long and intricate one. Books like this one keep the veteran engineer and student, alike, up to date on current trends in the technology and offer a reference for the industry for its practical applications.
Energy optimization and consumption prediction are necessary to prevent energy waste, schedule energy usage, and reduce the cost. Today, smart computing technologies are slowly replacing the traditional computational methods in energy optimization, consumption, scheduling, and usage. Smart computing is an important core technology in today's scientific and engineering environment. Smart computation techniques such as artificial intelligence, machine learning, deep learning and Internet of Things (IoT) are the key role players in emerging technologies across different applications, industries, and other areas. These newer, smart computation techniques are incorporated with traditional computation and scheduling methods to reduce power usage in areas such as distributed environment, healthcare, smart cities, agriculture and various functional areas.
The scope of this book is to bridge the gap between traditional power consumption methods and modern consumptions methods using smart computation methods. This book addresses the various limitations, issues and challenges of traditional energy consumption methods and provides solutions for various issues using modern smart computation technologies. These smart technologies play a significant role in power consumption, and they are cheaper compared to traditional technologies. The significant limitations of energy usage and optimizations are rectified using smart computations techniques, and the computation techniques are applied across a wide variety of industries and engineering areas. Valuable as reference for engineers, scientists, students, and other professionals across many areas, this is a must-have for any library.


John A, PhD, is an assistant professor at Galgotias University, Greater Noida, India, and he received his PhD in computer science and engineering from Manonmaniam Sundaranar University, Tirunelveli, India. He has presented papers in various national and international conferences and has published papers in scientific journals.
Senthil Kumar Mohan, PhD, is an associate professor in the Department of Software and System Engineering at the School of Information Technology and Engineering, Vellore Institute of Technology, Vellore, India. He received his PhD in engineering and technology from Vellore Institute of Technology, and he has contributed to many research articles in various technical journals and conferences.
Sanjeevikumar Padmanaban, PhD, is a faculty member with the Department of Energy Technology, Aalborg University, Esbjerg, Denmark. He has almost ten years of teaching, research and industrial experience and is an associate editor on a number of international scientific refereed journals. He has published more than 300 research papers and has won numerous awards for his research and teaching.
Yasir Hamid, PhD, is an assistant professor in the Department of Information Security Engineering Technology at Abu Dhabi Polytechnic. He earned his PhD in 2019 from Pondicherry University in Computer Science and Engineering. Before joining ADPOLY, he was an assistant professor in the Department of Computer Science, Islamic University of Science and Technology, India. He is an editorial board member on many scientific and technical journals.
Details
Weitere ISBN/GTIN9781119821854
ProduktartE-Book
EinbandartE-Book
FormatEPUB
Format Hinweis2 - DRM Adobe / EPUB
FormatFormat mit automatischem Seitenumbruch (reflowable)
Erscheinungsjahr2022
Erscheinungsdatum30.09.2022
Auflage1. Auflage
ReiheNext Generation Computing and Communication Engineering
Seiten336 Seiten
SpracheEnglisch
Dateigrösse31024 Kbytes
Artikel-Nr.9929440
Rubriken
Genre9201

Inhalt/Kritik

Leseprobe

1
Review and Analysis of Machine Learning Based Techniques for Load Forecasting in Smart Grid System

Shihabudheen KV1 and Sheik Mohammed S2*

1Electrical Engineering Department, National Institute of Technology, Calicut, Kerala, India

2Electrical and Electronic Engineering Programme Area, Faculty of Engineering, Universiti Teknologi Brunei, Gadong, Brunei Darussalam
Abstract

Electrical load forecasting is an important process that can improve the efficiency and economy of the utility grid especially in the smart grid environment. Load forecasting plays a significant role in making decisions such as planning, generation scheduling, operation, pricing customer satisfaction, and system security. However, load forecasting is a tedious and difficult task due to the intermittent nature of Renewable Energy Systems (RES) that varies depending on the seasons and parameters such as change in temperature and humidity. Moreover, the connect loads are also complex in nature as they vary from season to season. Artificial Intelligent (AI) techniques are a promising approach for better load forecasting having chaotic and random variation of both load and generation. In this chapter, a load-forecasting algorithm for time series loads using AI techniques with supervised methods is presented and discussed. A comparative assessment of load forecasting based on supervised artificial intelligent algorithms, such as Artificial Neural Networks (ANN), Support Vector Machine (SVM), and Extreme Learning Machine (ELM), is performed on smart meter data. The results are presented and performance of the selected algorithms are analysed.

Keywords: Electricity load forecasting, Artificial Intelligence (AI), Artificial Neural Networks (ANN), Support Vector Machine (SVM), extreme learning machine (ELM), smart grid, smart meter
1.1 Introduction

Traditionally, electric grids are a network of electric power generation, transmission, and distribution systems controlled by a centralized system. The conventional electric grid system has one way power flow and a communication approach from the generating station until the end customer. However, the power generation, transmission, and distribution systems had a noteworthy revolution in recent times. The progress and development in power generation using renewable energy sources is one of the important reasons behind this transformation. The restructured power system has made distribution systems bi-directional, controllable grids called Smart Grids. The Smart Grid consists of a number of RES with loads, ESS, sensors, and communications networks connected in a well-arranged fashion so that it has the potential to improve overall system performance. The coordinated and controlled operation of this integrated structure makes the grid smarter by managing generation, distribution, customers, and the market in both an efficient and effective manner. Figure 1.1 shows the different domains and stakeholders of the Smart Grid.

Figure 1.1 Structure of smart grid [1].

Electricity load forecasting is a projection of the load demand that electricity users are expected to have in the future. Load forecasts enable the utilities to manage supply and demand and also ensure the stability of power grids. Load forecasting is the key element for smart grid operation, as it plays a vital role in decision-making such as planning, scheduling, operation, capacity addition, pricing, generation planning, and system security. Another major advantage of load forecasting is that it helps both the utility and consumers to optimize their energy usage.

Load forecasting is classified based on horizon and scale. Scale level is the unit size at which the forecasting is performed. Scale level forecasting ranges from individual forecasting (meter-level) in homes and building levels (multi-meters level) to region, district, state, and up to country (integrated) level load forecasting. On the other hand, horizon defines the time range of load forecasting. Horizon level forecasting is classified as very short-term load forecasting (VSTLF), short-term load forecasting (STLF), medium-term load forecasting (MTLF) and long-term load forecasting (LTLF). In VSTLF, minutes to hour ahead prediction is carried out and SLTF is day ahead to weekly forecasting. MTLF deals with one week to three years prediction and more than three years prediction is known as LTLF. Each type of forecasting serves different purposes in the power system for scheduling, economic dispatch, operation planning, maintenance, capacity expansion planning, fuel economy, sales, etc. [2-4].

Traditionally, expected demand is forecast using the information about past use and other related data with the aid of charts and graphs by applying an engineering approach. Later, data driven approaches are predominantly used to build prediction algorithms to improve the efficiency and accuracy of forecasting. Statistical based techniques and intelligent computing are the two main categories of data driven approaches applied for electricity load forecasting [5, 6]. Statistical approaches use historical data to compare energy consumption with the most relevant variables as inputs. More high-quality historical data therefore plays a vital role in the efficacy of statistical models. Conventional approaches like Regression Models, Conditional Demand Analysis (CDA), Auto Regressive Moving Average (ARMA), and ARIMA are the most commonly adopted statistical methods for time series prediction. However, many researchers have investigated forecasting using AI based techniques and deep learning and those techniques have become the widely accepted technology over the past decade [7]. In addition to that, machine learning techniques like Classification and Regression Trees (CART), as well as Support Vector Machine (SVM) techniques are also used for time series prediction. Fuzzy Logic Systems, Artificial Neural Networks (ANN), Evolutionary Programming, and expert systems are some of the AI based approaches. Among them, ANN has widespread acceptance for time series forecasting [8-12]. Many attempts are made to solve the load-forecasting problems using AI based hybrid approaches. A comprehensive review on all such types of forecasting techniques are discussed in [13-16]. A review based on different categorization of the various forecasting, including the hybrid method, is less attempted in most of the literature.

In this book chapter, an extensive review of different supervised AI based load forecasting methodology is discussed. The review includes different categories of prediction such as single prediction and hybrid prediction methods. The details of hybrid prediction such as combined AI based prediction and signal decomposition based prediction techniques are included. Moreover, a comparative simulation study is performed on smart meter data. Methodology of forecasting is described in Section 1.2. A comprehensive review of various AI based prediction strategies applied for load forecasting is presented in Section 1.3. Comparative assessment of single and hybrid predictions is performed on smart meter data and the results are presented and discussed in Section 1.4. AI techniques such as Back Propagation Based Neural Network (BPNN), Support Vector Machine (SVM), and Extreme Learning Machine (ELM) are used for single prediction. For hybrid prediction, Empirical Model Decomposition (EMD) based BPNN and SVM (EMD-BPNN and EMD-SVM) prediction are performed. The chapter is concluded in Section 1.5.
1.2 Forecasting Methodology

The forecasting methodology for prediction of time series loads consists of two steps, as shown in Figure 1.2. The first step of methodology is feature extraction. Initially, a sufficient quantity of features is extracted form load time series data. Feature extraction procures the features which aid in the prediction of time series data. Transferring the collected features into a more informative analysis domain helps in sensing the hidden characteristics of future loads. The second step is the implementation of a predictor for accurate forecasting.

Figure 1.2 Overall steps for load time series forecasting methodology.

Let y(t) represent a load time series data. The prediction equation can be mathematically represented by

(1.1)

where g(t) indicates the extracted features, f represents the predictive function to be approximated by predictor and ytË( + k) is k step ahead of predicted values of y(t).
1.3 AI-Based Prediction Methods

Many AI-based prediction methods are proposed in literatures for time series based load forecasting. The classification of AI based prediction techniques is shown in Figure 1.3. An overview some of commonly used prediction methods are outlined in this section.
1.3.1 Single Prediction Methods

Single prediction indicates the prediction of time series, which is formed using a single AI technique. Some of the AI techniques used for single prediction of time series data are Linear Regression, Artificial Neural Networks, Support Vector Regression, Extreme Learning Machine, and Neuro-Fuzzy Methods.
1.3.1.1 Linear Regression
This approach is used to predict the dependent variable using several independent variables or features. It uses the assumption that a linear relation may exist between the features and output signals. A linear...
mehr

Autor