Introduction to Machine Learning Methods in Energy Engineering

Nowadays, the increment of energy demand in the world as well as the development of smart grids and the combination of different types of energy systems have led to the complexity of power systems. On the other hand, ever-expanding energy consumption, development of industry and technology systems, and high penetration of renewable energies have made electricity networks operating in more complex and uncertain conditions. Also, consumers and especially users of sensitive load tend to have access to a reliable and sustainable power supply. Therefore, power producers need a variety of long- and short-term planning methodologies for attaining to sustainable investment, production, and operation. Analysis of traditional power and energy systems requires physical modeling and extensive numerical computation. To analyze the behavior of these systems, advanced metering, and condition monitoring devices and systems are utilized, which generate a huge amount of data. Evaluation of these data is approximately impossible by conventional or numerical methods, and it requires powerful data mining procedures. Regression, classification, and clustering applications of machine learning and deep learning methods are powerful tools to use for dealing with such issues. These procedures can be utilized for load/demand forecasting, renewable energy generation forecasting, demand response evaluation, and power system analysis. Understanding the problem and functioning of each learning methods is, therefore, one of the most important issues in the application of such approaches to solve power system problems. Accordingly, in this chapter, the authors will introduce and discuss selected applications of machine learning and deep learning based on their learning, structure, mode of operation, and application in the load forecasting of power systems. Literature review on machine learning and deep learning applications in load forecasting will be presented in this chapter.

Guest Editorial: Situational awareness of integrated energy systems

In this study, an effective Web-enabled generalized architectural model is proposed. The proposed model makes the power system analysis to be completely secured, distributed, platform-independent and language-independent. Although significant progress has been made for on-line power system analysis in distributed environment, still efficient distributed environment has not yet been exploited. The Web-enabled architectural models are emerging as basic methodologies to support the integration of different power system applications using open Internet standard. The existing RMI models for solving multi-area power system problems in distributed environment, has its limitations to use in internet-based applets. The main objective of this study is to convert existing RMI model for power system applications into Web based model for on-line monitoring of multi-area power systems in distributed environment. Test bench has been created to carry out the performance analysis of the proposed model. Using the proposed model, load flow analysis is carried out for various bus systems. The results obtained are reported.

Direct stability analysis using an energy function has recently been widely used for angular and voltage stability analysis in electric power systems. Considerable efforts have concentrated on seeking energy functions for lossy power systems, accomplished only for the simple two-bus system. A new approach is presented to derive an energy function by using the complex integral of bus current equations with respect to bus voltages. Under an assumption that all transmission lines have a uniform R/X ratio, it is shown that the energy function can be developed for multimachine power systems with losses. The proposed energy function is tested on sample systems by comparing the time simulation method, which shows the validity of the proposed method.

Voltage instability has caused a great deal of concern in the existing power systems. Despite best efforts, it is still not uncommon a phenomenon and has caused some of the major blackouts and catastrophic failures in the recent past, resulting in huge social and economic losses. The advent of synchrophasor technology has made possible wide-area measurements in real-time, and has found huge applications in power systems. The identification of subregions in power systems that experience a unique voltage instability problem is one of the most important steps of voltage stability analysis. This paper presents a method to identify voltage control area (VCA), based on coherent groups of buses, using system states obtained from synchronized phasor measurements. The coherency is identified by applying hierarchical clustering, a machine learning technique. The coherent buses are identified by applying the machine learning technique on the angles obtained from bus voltage phasors. The results so obtained using the data from Phasor Measurement Units (PMUs) on 10-machine, 39-bus New England power system model are presented.

Overviews on the applications of the Kuramoto model in modern power system analysis

This paper presents how power electronics concepts are utilized to teach the topics in power system analysis course. Power system analysis is one of the multiple core courses on power engineering in almost all Electrical Engineering programs around the world. At Grand Valley State University, it is the only course in power engineering offered to Electrical Engineering students during a twelve-week summer semester in junior year. This course covers the fundamental concepts and their applications in all three sections of a complex power system namely generation, transmission, and distribution systems. In recent years, there has been a significant change in the role and application of power electronics in modern power systems. However, at GVSU power electronics is a graduate course and it is also taken by some senior students with deep interest in it as an elective. Hence, attempts have been made to get all Electrical Engineering students, not only the willing seniors, introduced to power electronics and its applications in all three sections of a power system. These include teaching some power system concepts with applications of power electronics and administering power converter based course projects. This paper presents the design of the course to implement power electronics concepts and the implementation of course projects in power system analysis course. The pedagogical challenge is to enhance students' learning of modern power system. This paper shows how it enhanced and changed students' perception about the scope of power engineering.

Power system transfer learning to identify event types with or without event labels

Nowadays, in power systems, we still lack the existence of standardized test systems that can be used to benchmark the performance and solution quality of proposed optimization techniques. Several authors report that the electric load pattern is very complex. It is therefore necessary to develop new methods for design of test cases for economic analysis in power systems. Therefore, we compared two methods to generate test systems: time series model and a method simulating stable random variables based on the use of Chambers-Mallows-Stuck. This paper describes a method for simulating stable random variables in the generation of test systems for economic analysis in power systems. A study focused on generating test electrical systems through stable distribution to model for unit commitment problem in electrical power systems. Usually, the instances of test systems in unit commitment are generated using normal distribution, but the behavior of electrical demand does not follow a normal distribution; in this work, simulation data are based on a new method. For empirical analysis, we used three original systems to obtain the demand behavior and thermal production costs. Numerical results illustrate the applicability of the proposed method by solving several unit commitment problems directly and through the Lagrangian relaxation of the original problem. All Rights Reserved © 2015 Universidad Nacional Autonoma de Mexico, Centro de Ciencias Aplicadas y Desarrollo Tecnologico. This is an open access item distributed under the Creative Commons CC License BY-NC-ND 4.0.

The increasing share of variable renewable energy (VRE) generation poses challenges to power systems. Possible challenges include adequacy of reserves, planning and operation of power systems, and interconnection expansion studies in future power systems with very different generation patterns compared to today. To meet these challenges, there is a need to develop models and tools to analyze the variability and uncertainty in VRE generation. To address the varied needs, the tools should be versatile and applicable to different geographical and temporal scales. Time series simulation tools can be used to model both today and future scenarios with varying VRE installations. Correlations in Renewable Energy Sources (CorRES) is a simulation tool developed at Technical University of Denmark, Department of Wind Energy capable of simulating both wind and solar generation. It uses a unique combination of meteorological time series and stochastic simulations to provide consistent VRE generation and forecast error time series with temporal resolution in the minute scale. Such simulated VRE time series can be used in addressing the challenges posed by the increasing share of VRE generation. These capabilities will be demonstrated through three case studies: one about the use of large‐scale VRE generation simulations in energy system analysis, and two about the use of the simulations in power system operation, planning, and analysis.This article is categorized under: Wind Power > Systems and Infrastructure Energy Infrastructure > Systems and Infrastructure Energy Systems Economics > Systems and Infrastructure

This paper presents evolution of traditional power systems from its smaller size to the current state that incorporate renewables, distributed energy resources, demand-side management, smart grid, micro grid and electric market deregulation and restructuring. The resulting complexity and challenges in terms operational aspects, analysis of various important features such management and control have been presented together with challenges that are faced in the process of analysis due to huge amount data involved. Further, computational techniques used in power systems such as parallel processing, distributed computing and grid computing are discussed together with the state-of-the-art cloud computing and its applications for efficient and reliable management of the power systems along with an appropriate conclusion on the current scenario and future requirements.

This research explores the resurgence of DC applications in power systems, focusing on the integration of DC microgrids and the effectiveness of control strategies in managing such systems. The research aims to investigate how artificial intelligence (AI) can enhance the efficiency of DC microgrids, especially in the context of renewable energy sources (RES) integration. The methodology involves a comprehensive review of existing literature on the evolution of DC and AC power systems, with a particular emphasis on voltage regulation. The research examines various DC microgrid control methods, including centralized, decentralized, distributed, and hierarchical strategies. Special attention is given to the role of AI, machine learning, and the Internet of Things (IoT) in the tertiary level of hierarchical control, which aims to optimize system performance. The findings highlight the advantages of DC microgrids, particularly their compatibility with RES like Solar Photovoltaics (PV) and energy storage systems, as well as modern electrical appliances. Hierarchical control, with its primary, secondary, and tertiary levels, effectively manages voltage, power sharing, and operational efficiency. The tertiary level of control, when combined with AI technologies, significantly enhances the microgrid's performance through real-time data analysis and predictive decision-making, improving responsiveness and the integration of renewable energy. The research concludes that hierarchical control methods, augmented by AI, enable reliable, stable, and efficient operation of DC microgrids. These strategies improve the coordination of distributed generation, reduce operational costs, and enable participation in energy markets, promoting the sustainability of power systems. Future research should focus on advanced AI algorithms for optimization, exploring scalable control methods for larger microgrids and hybrid AC-DC systems to enhance energy transfer and storage capabilities.

Building a new power system with new energy as the main body is one of the most important measures to achieve “carbon peak and carbon neutral” in China, which also puts forward higher requirements on the data analysis and processing capacity of the current power system. In recent years, artificial intelligence algorithms represented by deep learning have provided strong support for the realization and development of new energy power system. The paper summarizes the basic characteristics of deep learning algorithm and its five mainstream models. From the aspects of image recognition, power load prediction, fault diagnosis and system optimization scheduling, the application of deep learning in new energy power system is deeply analyzed. The application prospect of deep learning in new energy power system is prospected in order to provide reference for the research and construction of new power system.

This study addresses current situations and challenges surrounding the full implementation of the Smart Grid and presents wholistic ways going forward. Considering the challenges of environmental degradation, ageing infrastructures, increasing energy demands, high energy costs, sustainability, transparency and partial participation in the vertical power systems, a shift in the energy paradigm has been looked upon by the various stakeholders. An intelligent-based grid (Smart Grid) development is considered as sustainable alternative - with the confidence that the challenges concerned with traditional power systems will alleviate, and by including large renewable energy resources, energy cost will reduce, and system operation will become more secure and reliable by leveraging the various technologies that support full implementation of the new grid. Despite these promises and the availability of the important technologies for its realization, Smart Grid (SG) development has been slow and full implementation has not been realized in most countries where it is adopted. This paper presents a way forward by expositing the roles that Customers, Utilities and Companies could take as pacesetters in accelerating the implementation of the new technology solution.

The general focus of this study is to design a multilevel deep learning model that provides big data analytics and emergency management knowledge. A big data covariance analysis approach has been used to find multilevel representations of data based on prior knowledge from large scale power systems. For purpose of meeting requirements of incremental knowledge discovery, an adaptive regression algorithm is presented. Given the multilevel operating status and development trend of power system, the emergency management techniques are then proposed to produce intelligent decision making support. In this paper, a multilevel clustered hidden Markov model based global optimization approach is considered for power system emergency management problem, which is an extension of the conventional optimal power flow problem. The objective is defined to generate operation mode that minimizes multilevel cost while satisfying different constraints. To demonstrate the effectiveness of the presented approach, this paper carefully compared the discriminatory power of knowledge discovery models that utilize deep learning with dimensionality reduction based method and machine learning without dimensionality reduction based method. The experimental results showed that the proposed multilevel deep learning approach consistently outperformed the traditional machine learning method. The emergency management of large scale power system may also benefit from the modified hidden Markov model and global optimization.