No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Energy forecasting tools and services
Abstract
The increasing complexity of the power grid and the continuous integration of volatile renewable energy systems on all aspects of it have made more precise forecasts of both energy supply and demand necessary for the future Smart Grid. Yet, the ever increasing volume of tools and services makes it difficult for users (e.g., energy utility companies) and researchers to obtain even a general sense of what each tool or service offers. The present contribution provides an overview and categorization of several energy‐related forecasting tools and services (specifically for load and volatile renewable power), as well as general information regarding principles of time series, load, and volatile renewable power forecasting. WIREs Data Mining Knowl Discov 2018, 8:e1235. doi: 10.1002/widm.1235
This article is categorized under: Application Areas > Business and Industry
Application Areas > Data Mining Software Tools
Technologies > Prediction
... Yet, it is crucial to manage and quantify the uncertainty surrounding forecasts. Moreover, the ongoing trend towards increasing automation in distribution systems and active coordination of distribution and transmission systems induces the need for time series forecasting on lower voltage levels, see, e.g., [1]. ...
... The chosen copula models are (rotated) Joe, (rotated) BB8, t, (rotated) Gumbel, Independence, (rotated) Clayton, Frank, (rotated) Tawn type 1, (rotated) Tawn type 2, Gaussian, i.e., we mainly observe tail dependencies here, which means that high or low values in active power at different times often are related. 1 Using inverse transform sampling we now derive the cumulative distribution function (c.d.f.) corresponding to the kernel 1 Due to space limitations of this note, we cannot detail the properties of these copula models here, we refer to [26] for an introduction. density estimate for each period. ...
... The chosen copula models are (rotated) Joe, (rotated) BB8, t, (rotated) Gumbel, Independence, (rotated) Clayton, Frank, (rotated) Tawn type 1, (rotated) Tawn type 2, Gaussian, i.e., we mainly observe tail dependencies here, which means that high or low values in active power at different times often are related. 1 Using inverse transform sampling we now derive the cumulative distribution function (c.d.f.) corresponding to the kernel 1 Due to space limitations of this note, we cannot detail the properties of these copula models here, we refer to [26] for an introduction. density estimate for each period. ...
Predicting the time series of future evolutions of renewable injections and demands is of utmost importance for the operation of power systems. However, the current state of the art is mostly focused on mean-value time series predictions and only very few methods provide probabilistic forecasts. In this paper, we rely on kernel density estimation and vine copulas to construct probabilistic models for individual load profiles of private households. Our approach allows the quantification of variability of individual energy consumption in general and of daily peak loads in particular. We draw upon an Australian distribution grid dataset to illustrate our findings. We generate synthetic loads that follow the distribution of the real data.
... , K} reflects a set of K ∈ N >0 observations typically measured at equidistant points in time [20]. A time series forecasting model f (·) estimates future valuesŷ for one or more time points -the forecast horizon H ∈ N >0 -using current and past values [21]. It is defined aŝ ...
... where H 1 ∈ N >0 indicates the horizon for past values k − H 1 , the vector w contains the model's parameters, the vector u denotes values from exogenous time series, the vectorû indicates that the exogenous values originate from another forecast, and y represents values of the target time series [21]. 6 non-stationary, and the requirements for applying ARMA are not fulfilled. ...
... The random search may performs better than the grid search if some hyperparameters are much more important than others. 21 Figure 3 shows a comparison of both methods with two hyperparameters and an equal number of computations B. Hyndman et al. [87] generalize the formulation of ES forecasting methods with the Error Trend Seasonality (ETS) method and suggest a grid search that automatically selects the hyperparameter configuration with the lowest in-sample AIC. Utilizing grid search for HPO is also applied to statistical forecasting methods based on AR and Moving Averages (MAs) averages. ...
Time series forecasting is fundamental for various use cases in different domains such as energy systems and economics. Creating a forecasting model for a specific use case requires an iterative and complex design process. The typical design process includes the five sections (1) data pre-processing, (2) feature engineering, (3) hyperparameter optimization, (4) forecasting method selection, and (5) forecast ensembling, which are commonly organized in a pipeline structure. One promising approach to handle the ever-growing demand for time series forecasts is automating this design process. The present paper, thus, analyzes the existing literature on automated time series forecasting pipelines to investigate how to automate the design process of forecasting models. Thereby, we consider both Automated Machine Learning (AutoML) and automated statistical forecasting methods in a single forecasting pipeline. For this purpose, we firstly present and compare the proposed automation methods for each pipeline section. Secondly, we analyze the automation methods regarding their interaction, combination, and coverage of the five pipeline sections. For both, we discuss the literature, identify problems, give recommendations, and suggest future research. This review reveals that the majority of papers only cover two or three of the five pipeline sections. We conclude that future research has to holistically consider the automation of the forecasting pipeline to enable the large-scale application of time series forecasting.
... A forecasting model based on autoregression (·) estimates future expected valuesˆat the origin for the forecast horizon ∈ N 1 using past and current values [42]. Formally, such a model is defined aŝ ...
... where y represents values of the WP turbine's power generation, the matrix X represents covariates,X indicates that the respective covariates are estimates, 1 ∈ N 1 is the horizon of past values 1 , and the vector p includes the model's trainable parameters [42]. ...
Renewable energies and their operation are becoming increasingly vital for the stability of electrical power grids since conventional power plants are progressively being displaced, and their contribution to redispatch interventions is thereby diminishing. In order to consider renewable energies like Wind Power (WP) for such interventions as a substitute, day-ahead forecasts are necessary to communicate their availability for redispatch planning. In this context, automated and scalable forecasting models are required for the deployment to thousands of locally-distributed onshore WP turbines. Furthermore, the irregular interventions into the WP generation capabilities due to redispatch shutdowns pose challenges in the design and operation of WP forecasting models. Since state-of-the-art forecasting methods consider past WP generation values alongside day-ahead weather forecasts, redispatch shutdowns may impact the forecast. Therefore, the present paper highlights these challenges and analyzes state-of-the-art forecasting methods on data sets with both regular and irregular shutdowns. Specifically, we compare the forecasting accuracy of three autoregressive Deep Learning (DL) methods to methods based on WP curve modeling. Interestingly, the latter achieve lower forecasting errors, have fewer requirements for data cleaning during modeling and operation while being computationally more efficient, suggesting their advantages in practical applications.
... In fact, it is widely seen as the cornerstone that underpins stable grid operation by matching production and consumption in real time [1]. Over time, forecasting the future has gained in importance with energy market liberalization and increasing proliferation of intermittent renewable energy sources [2]. ...
... In this regard, one can find both white-and black-box forecasting methods. A comprehensive review of different forecasting techniques deployed in the electric power sector can be found in [2]. In the remainder of this study, we focus exclusively on short-term, day-ahead forecasts. ...
Transmission system operators (TSOs) forecast load and renewable energy generation to maintain smooth functioning of the grid by contracting sufficient generation and reserve capacity. These forecasts are also utilized by third parties, such as energy generators and demand aggregators, in their own forecasting and decision-making pipelines e.g. to determine suitable trading strategies. Inaccurate forecasts by the TSOs can therefore lead to increased balancing needs as well as elevated societal and market costs. The situation is further exacerbated by the challenges arising due to rapidly increasing renewable generation and the effects of the post-Covid era. In this paper, we analyse five years of TSO forecasts for load, wind and solar generation for 16 European countries. More concretely, using a comprehensive set of metrics, we explore relevant questions such as whether there are TSO specific differences in forecast accuracy, and how forecast errors have changed over time and if they can be reduced further. Our results show that while errors tend to increase linearly with demand or renewable generation, most TSOs still have considerable room for improvement in terms of accuracy. The paper concludes with a set of recommendations for TSOs to improve their forecasts, as well as the ENTSO-E transparency platform where we obtained the data used in this study.
... Still, the pipeline requires manual tailoring by the data scientist to meet the specific requirements. Most published literature on energy forecasts range between Automation level 0 and 1, see reference [7]. Across the literature, standard procedures have emerged that can be used to create automated pipeline templates for specific tasks. ...
... For grid search, we adopt the configuration space Λ of reference [37] with the 1C-SVM implementation of the Scikit-learn library 7 We implemented the automated design process in Python [42] and adapted the grid search from the implementation of the Scikit-learn library 7 [41]. SVR, RF, and LR are based on the Scikit-learn library as well. ...
Undoubtedly, the increase of available data and competitive machine learning algorithms has boosted the popularity of data-driven modeling in energy systems. Applications are forecasts for renewable energy generation and energy consumption. Forecasts are elementary for sector coupling, where energy-consuming sectors are interconnected with the power-generating sector to address electricity storage challenges by adding flexibility to the power system. However, the large-scale application of machine learning methods in energy systems is impaired by the need for expert knowledge, which covers machine learning expertise and a profound understanding of the application's process. The process knowledge is required for the problem formalization, as well as the model validation and application. The machine learning skills include the processing steps of i) data pre-processing, ii) feature engineering, extraction, and selection, iii) algorithm selection, iv) hyperparameter optimization, and possibly v) post-processing of the model's output. Tailoring a model for a particular application requires selecting the data, designing various candidate models and organizing the data flow between the processing steps, selecting the most suitable model, and monitoring the model during operation - an iterative and time-consuming procedure. Automated design and operation of machine learning aim to reduce the human effort to address the increasing demand for data-driven models. We define five levels of automation for forecasting in alignment with the SAE standard for autonomous vehicles, where manual design and application reflect Automation level 0.
... Network operators, consequently, rely heavily on demand forecasts for scheduling the power supply to match the demand. This need for demand forecasts has led to various forecasting approaches being introduced as presented in e. g. [1,6,8,9,22,23]. ...
... Traditionally, statistical forecasting models are used, as they are easy to interpret while having a great forecasting performance [9]. However, in many areas, neural networks and especially deep neural networks have shown great success in learning difficult dependencies and uncover hidden knowledge in large data sets [16]. ...
... The primary areas of attention in energy research include neural networks, expert systems, pattern recognition, and fuzzy logic models (Demirci et al., 2019;Tyralis et al., 2019). These domains encompass energy production and distribution, operations, and maintenance, which have been of significant interest in the field of energy (Wang & Srinivasan, 2017;Ordiano et al., 2018;Merizalde et al., 2019;Nishant et al., 2020). Machine learning algorithms are utilized for predicting future outcomes (Olowu et al., 2018), whereas NC algorithms are employed to address multi-objective issues (Li et al., 2018). ...
Artificial Intelligence plays a crucial role in addressing various environmental sustainability challenges through technological innovations in the fields of energy, transportation, biodiversity, and water management. Thus, the present study aims to present a concise review of Artificial Intelligence (AI) toward achieving environmental sustainability. The main areas of concentration in innovations in the field of energy encompass neural networks, expert systems, pattern recognition, and fuzzy logic models. Artificial Intelligence enables the creation of advanced prediction models for renewable energy production, enhancing the allocation of resources and management of the power grid. Moreover, computer vision and decision assistance have been used in the field of transportation. Additionally, the use of Artificial Intelligence and machine learning is growing to predict and enhance water resource conservation. Besides, machine learning and natural language processing techniques are being used in biodiversity research to predict ecological services. However, regular monitoring of initiatives is necessary to enhance environmental sustainability.
... To overcome the manual feature engineering, new methods based on deep learning were developed. González Ordiano et al. [13] give an overview on existing energy time-series forecasting methods, including linear regression and multi-layer perceptrons, which we are going to use as baseline models (see Section 5). More sophisticated methods were developed in the meantime, such as profile neural networks [14]. ...
Accurate forecasts of the electrical load are needed to stabilize the electrical grid and maximize the use of renewable energies. Many good forecasting methods exist, including neural networks, and we compare them to the recently developed Transformers, which are the state-of-the-art machine learning technique for many sequence-related tasks. We apply different types of Transformers, namely the Time-Series Transformer, the Convolutional Self- Attention Transformer and the Informer, to electrical load data from Baden- Württemberg. Our results show that the Transformes give up to 11% better forecasts than multi-layer perceptrons for long prediction horizons. Furthermore, we analyze the Transformers’ attention scores to get insights into the model.
... Renewable energy sources are essential to mitigate climate change (Clarke et al. 2022), and being able to forecast their supply is necessary to balance any energy system that includes them (González Ordiano et al. 2018). However, renewable energy sources and demand are weather dependent; thus, weather information plays a crucial role in renewable energy-related forecasting (Bloomfield et al. 2021;Harish et al. 2020;Vanting et al. 2021). ...
Renewable energy systems depend on the weather, and weather information, thus, plays a crucial role in forecasting time series within such renewable energy systems. However, while weather data are commonly used to improve forecast accuracy, it still has to be determined in which input shape this weather data benefits the forecasting models the most. In the present paper, we investigate how transformations for weather data inputs, i. e., station-based and grid-based weather data, influence the accuracy of energy time series forecasts. The selected weather data transformations are based on statistical features, dimensionality reduction, clustering, autoencoders, and interpolation. We evaluate the performance of these weather data transformations when forecasting three energy time series: electrical demand, solar power, and wind power. Additionally, we compare the best-performing weather data transformations for station-based and grid-based weather data. We show that transforming station-based or grid-based weather data improves the forecast accuracy compared to using the raw weather data between 3.7 and 5.2%, depending on the target energy time series, where statistical and dimensionality reduction data transformations are among the best.
... The main research areas are pattern recognition, expert systems, neural networks, and fuzzy logic (Nishant et al. 2020), which are relevant to energy research (Tyralis et al. 2019). This includes the distribution and production of energy and maintenance and operations, which are the major research areas in energy (González Ordiano et al. 2017). Computer-aided learning is employed to forecast (Olowu et al. 2018), and NC algorithms can also be used in solving complex problems (Li et al. 2018). ...
Environmental issues have continued to spur discussions, debates, public outrages, and awareness campaigns, inciting interest in emerging technologies such as Artificial Intelligence. Its usage is spread across many environmental industries, including wildlife protection, natural resource conservation, clean energy, agriculture, energy management, pollution control, and waste management. In 2017, at the United Nations Artificial Intelligence Summit in Geneva, the UN acknowledged that AI could be an enabler in the sustainable development process towards peace, prosperity, and dignified life for humankind and proposed to refocus on the application of AI in assisting global efforts on sustainable development to eradicate poverty, hunger and to protect the environment as well as to conserve natural resources. It is vital to address environmental sustainability concerns; however, with the advent of AI, most common environmental issues are now solvable by prioritizing human interests. Sustainability encompasses the interrelated areas of the environment, society, and economy. According to the United Nations’ “Our Common Future,” also known as the “Brundtland Report,” it is defined as “development that satisfies current needs without compromising the ability of future generations to meet their own needs.” Unfortunately, the Earth is currently facing serious consequences from global warming and climate change, and immediate action is required to encourage the use of environmentally friendly and sustainable products to address these issues. Environmental degradation and climate change are numerous environmental concerns requiring novel and intelligent artificial intelligence solutions. The literature on AI and environmental sustainability encompasses various domains. Notably, AI is being used to address the bulk of regional and global environmental concerns, including energy, water, biodiversity, and transportation, even though many of these sectors have permeated and evolved. However, there is a need to combine current literature on the application of AI, particularly in relation to environmental sustainability in areas such as energy, water, biodiversity, and transportation. There is a significant lack of research on how AI can promote environmental sustainability. This research aims to explore how AI can be applied to address environmental issues in various sectors to achieve the Sustainable Development Goals (SDGs).
... • DNV GL Synergi Electric [137,138]: A commercial software, which is commonly used for power system analysis and planning, including energy storage modeling. It allows users to simulate the performance of energy storage systems in distribution and transmission networks and optimize their design for different applications, such as peak shaving, load balancing, and voltage support. ...
As the world’s population continues to grow and the demand for energy increases, there is an urgent need for sustainable and efficient energy systems. Renewable energy sources, such as wind and solar power, have the potential to play a significant role in meeting this demand, but their intermittency can make integration into existing energy systems a challenge. Moreover, the development of sustainable energy systems has become even more critical in recent years, due to a confluence of events, including the decline in fuel prices, geopolitical conflicts, and the recent COVID-19 pandemic. The decrease in fuel prices has led to a decline in investment in renewable energy and has slowed the transition to sustainable energy systems. Additionally, geopolitical conflicts and pandemics have highlighted the need for resilient and self-sufficient energy systems that can operate independently of external factors. Also, energy storage technologies play a critical role in achieving this goal by providing reliable backup power and enabling microgrids to operate independently of the larger power grid. As such, developing efficient and effective energy storage technologies is essential for creating sustainable energy systems that can meet the demands of modern society while mitigating the impact of external factors. In this regard, this work provides an overview of microgrids’ latest energy storage technologies, including their applications, types, integration strategies, optimization algorithms, software, and uncertainty analysis. Energy storage technologies have a wide range of applications in microgrids, including providing backup power and balancing the supply and demand of energy. Different energy storage techniques have been discussed, including batteries, flywheels, supercapacitors, pumped hydro energy storage, and others. Moreover, integration strategies of energy storage in microgrids, models, assessment indices, and optimization algorithms used in the design of energy storage systems are presented in detail. The capabilities of software used in energy storage sizing are explored. Further, uncertainty analysis in modeling energy storage devices is presented and discussed. This state-of-the-art technology has been prepared to demonstrate the effectiveness of energy storage technologies in microgrids, providing valuable insights for future developments in the field.
... González Ordiano et al. [3] give an overview on existing energy time-series forecasting methods, including linear regression and multi-layer perceptrons, which we use as baseline methods (see Section 5.3). Haben et al. [4] review applications and methods for low-voltage level forecasting, but do not cover Transformer neural networks, which were applied to time-series forecasting [5][6][7][8][9] and electrical load forecasting [10][11][12][13][14][15][16] only recently. ...
Accurate electrical load forecasts of buildings are needed to optimize local energy storage and to make use of demand-side flexibility. We study the usage of Transformer neural networks for short-term electrical load forecasting of 296 buildings from a public dataset. Transformer neural networks trained on many buildings give the best forecasts on 115 buildings, and multi-layer perceptrons trained on a single building are better on 161 buildings. In addition, we evaluate the models on buildings that were not used for training, and find that Transformer neural networks generalize better than multi-layer perceptrons and our statistical baselines. This shows that the usage of Transformer neural networks for building load forecasting could reduce training resources due to the good generalization to unseen buildings, and they could be useful for cold-start scenarios.
... To overcome the manual feature engineering, new methods based on deep learning were developed. González Ordiano et al. [13] give an overview on existing energy time-series forecasting methods, including linear regression and multi-layer perceptrons, which we are going to use as baseline models (see Section 5). More sophisticated methods were developed in the meantime, such as profile neural networks [14]. ...
Accurate forecasts of the electrical load are needed to stabilize the electrical grid and maximize the use of renewable energies. Many good forecasting methods exist, including neural networks, and we compare them to the recently developed Transformers, which are the state-of-the-art machine learning technique for many sequence-related tasks. We apply different types of Transformers, namely the Time-Series Transformer, the Convolutional Self-Attention Transformer and the Informer, to electrical load data from Baden-Württemberg. Our results show that the Transformers give up to 11% better forecasts than multi-layer perceptrons for long prediction horizons. Furthermore, we analyze the Transformers' attention scores to get insights into the model.
... Disruptive technologies, such as AI and big data, facilitate data flow and closed-loop business models with the ability to analyze data and understand the relationships between data (Jabbour et al., 2019). Energy consumption of buildings, energy production, energy consumption, energy distribution, maintenance and operation, and renewable energy are some of the applied topics in the field of AI (Gonzalez Ordiano et al., 2018;Merizalde et al., 2019;Wang & Srinivasan, 2017). Also, by using intelligent processes and expert information systems in the organization, it is possible to track goods and thus manage assets and products with high accuracy. ...
Considering the increase in the stakeholders’ supervision and the change of production processes, sustainable development plays a crucial role in the survival of businesses. In order to achieve sustainable development, the circular economy (CE) seeks to manage the flow of materials and energy to closed-loop systems. Circular economy has led to the formation of sustainable business models. Artificial intelligence (AI) capabilities change work activities, data flows, and organizational processes. The purpose of this study is to identify the impact of adoption of AI on circular economy practices in the organization. The research questions include: What are the factors affecting the adoption of AI in manufacturing companies? What effect does the adoption of AI have on the CE practices in the organization? In the first phase, research constructs are identified and a conceptual model is developed based on previous studies. In the second phase, the research model is evaluated among 97 manufacturing companies in the Middle East. Structural equation model and Smart PLS software have been used for data analysis. The findings show that the technology characteristics, organizational capabilities and external task environment have an effect on adoption of AI, and adoption of AI has a positive effect on circular economy practices. Based on the results, AI technology can be a solution to change the production process and reduce the destructive effects of industry on the environment. Managers of manufacturing companies can use the capabilities of machine learning, intelligence and neural networks to manage resources and optimize product production.
... load shifting-typically require accurate short-term load forecasts. Therefore, in recent years, a large number of machine learning models for short-term load forecasting (STLF) and load forecasting in general have been developed (Upadhaya et al. 2019;González Ordiano et al. 2018). Although these models provide an improved forecasting accuracy, their increasing complexity is also associated with a growing need for training data (Hippert et al. 2001;Hastie et al. 2009), often initially multiple years, e.g., 1 year (Wu and Shahidehpour 2010), 2 years (Yona et al. 2008), or 3 years (Mabel and Fernandez 2008). ...
Sustainable energy systems are characterised by an increased integration of renewable energy sources, which magnifies the fluctuations in energy supply. Methods to to cope with these magnified fluctuations, such as load shifting, typically require accurate short-term load forecasts. Although numerous machine learning models have been developed to improve short-term load forecasting (STLF), these models often require large amounts of training data. Unfortunately, such data is usually not available, for example, due to new users or privacy concerns. Therefore, obtaining accurate short-term load forecasts with little data is a major challenge. The present paper thus proposes the latent space-based forecast enhancer (LSFE), a method which combines transfer learning and data augmentation to enhance STLF when training data is limited. The LSFE first trains a generative model on source data similar to the target data before using the latent space data representation of the target data to generate seed noise. Finally, we use this seed noise to generate synthetic data, which we combine with real data to enhance STLF. We evaluate the LSFE on real-world electricity data by examining the influence of its components, analysing its influence on obtained forecasts, and comparing its performance to benchmark models. We show that the Latent Space-based Forecast Enhancer is generally capable of improving the forecast accuracy and thus helps to successfully meet the challenge of limited available training data.
... A time series y k ½ ; k ¼ 1, 2…, K f g reflects a set of K ℕ > 0 observations typically measured at equidistant points in time (Brockwell & Davis, 2016). A time series forecasting model f Á ð Þ estimates future values b y for one or more time points-the forecast horizon H ℕ > 0 -using current and past values (Gonz alez Ordiano et al., 2018). It is defined as F I G U R E 2 Related to this review are the fields of pipeline creation, data preprocessing, feature engineering, hyperparameter optimization (HPO) and combined algorithm selection and hyperparameter optimization (CASH), and forecast ensembling b y k þ H ½ ¼f y k ½ , …,y k À H 1 ½ ð ...
Time series forecasting is fundamental for various use cases in different domains such as energy systems and economics. Creating a forecasting model for a specific use case requires an iterative and complex design process. The typical design process includes five sections (1) data preprocessing, (2) feature engineering, (3) hyperparameter optimization, (4) forecasting method selection, and (5) forecast ensembling, which are commonly organized in a pipeline structure. One promising approach to handle the ever‐growing demand for time series forecasts is automating this design process. The article, thus, reviews existing literature on automated time series forecasting pipelines and analyzes how the design process of forecasting models is currently automated. Thereby, we consider both automated machine learning (AutoML) and automated statistical forecasting methods in a single forecasting pipeline. For this purpose, we first present and compare the identified automation methods for each pipeline section. Second, we analyze these automation methods regarding their interaction, combination, and coverage of the five pipeline sections. For both, we discuss the reviewed literature that contributes toward automating the design process, identify problems, give recommendations, and suggest future research. This review reveals that the majority of the reviewed literature only covers two or three of the five pipeline sections. We conclude that future research has to holistically consider the automation of the forecasting pipeline to enable the large‐scale application of time series forecasting.
This article is categorized under: Technologies > Machine Learning
Technologies > Prediction
Algorithmic Development > Spatial and Temporal Data Mining
... Probabilistic forecasting (PF) can be modeled using supervised/unsupervised ML algorithms. Although time series PF is a mature field, compared to point forecasting, its literature on power systems is fairly limited (González Ordiano, 2018). PF has been fostered only in the past decade as system operators identified the necessity to cope with the increasing uncertainty of renewables, loads, and electricity prices, which has been encouraged by energy forecasting competitions (Tao et al., 2019). ...
Behind-the-meter (BTM) resources is being recognized as a viable solution to offer grid services including flexibility procurement which is required for volatile renewable power systems. This paper brings an overview of current and esteemed frameworks and respective challenges revolving around BTM flexibility notion and mechanisms. To begin with, we review grid architectures, e.g., microgrids and virtual power plants, capable of accommodating BTM flexibility and desirable flexibility market designs, including peer-to-peer trading. The role of machine learning initiatives, including reinforcement learning and probabilistic forecasting, in designing reliable energy management systems is extensively deliberated. Last but not least, supplementary discussions in making this concept a reality, which can be regarded as future research, are given.
... In the opinion of Nishant et al. [5], artificial intelligence (AI) was argued to have supported the reduction of natural resource, as well as energy demands of human activities. Area neural network, expert systems, pattern recognition and fuzzy logic models are the main focus areas when it comes to energy research [15,29] including energy production and distribution, operations and maintenance, which have been core areas of interest in energy [4,5,30,31]. Machine learning algorithms are used for forecasting [32] and NC algorithms are also used to solve multi-objective problems [33]. ...
Artificial Intelligence (AI) has become an important area to tackle most environmental sustainability issues such as biodiversity, energy, transportation and water management. Biodiversity research has developed machine learning or natural language processing solutions to predict ecosystem services. Artificial intelligence applications and machine learning models have been increasingly used for predicting and optimizing water resource conservation. Area neural network, expert systems, pattern recognition, and fuzzy logic models are the main focus areas in energy. Applications of computer vision and decision support were found in transportation. Timely monitoring of interventions is required to improve environmental sustainability.
... In this paper, we analyze the impact of data representations on the performance of DNNs at the commonly investigated example of energy time series forecasting (see e.g. [14,15,16,13]). For this purpose, we investigate the time series in its original form and the derivative of the time series, and both reshaped as an image. For the analysis, we use two different architectures, namely a Fully Connected Network (FCN) and a Convolutional Neural Network (CNN). ...
Deep Neural Networks are able to solve many complex tasks with less engineering effort and better performance. However, these networks often use data for training and evaluation without investigating its representation, i.e.~the form of the used data. In the present paper, we analyze the impact of data representations on the performance of Deep Neural Networks using energy time series forecasting. Based on an overview of exemplary data representations, we select four exemplary data representations and evaluate them using two different Deep Neural Network architectures and three forecasting horizons on real-world energy time series. The results show that, depending on the forecast horizon, the same data representations can have a positive or negative impact on the accuracy of Deep Neural Networks.
... While first research on load forecasting dates back to the 1960s, selecting the most appropriate model for a specific forecasting scenario has since become a much more difficult task [5]. The majority of research in the subject area is conducted on increasing the accuracy of load forecasts, while the resulting (economic) value of forecasts is rarely addressed [6,7]. ...
Power system operators are confronted with a multitude of new forecasting tasks to ensure a constant supply security despite the decreasing number of fully controllable energy producers. With this paper, we aim to facilitate the selection of suitable forecasting approaches for the load forecasting problem. First, we provide a classification of load forecasting cases in two dimensions: temporal and hierarchical. Then, we identify typical features and models for forecasting and compare their applicability in a structured manner depending on six previously defined cases. These models are compared against real data in terms of their computational effort and accuracy during development and testing. From this comparative analysis, we derive a generic guide for the selection of the best prediction models and features per case.
... To further improve the imputation, it is a common approach to focus on time series from a particular domain and to consider their characteristics as additional information. In the context of smart meters, the recorded time series of electricity consumption or generation are typically influenced by factors such as weather, human routines, social norms (e.g., weekends or holidays) and many others [16], [17]. These factors often lead to the commonly known characteristic patterns with daily, weekly, and yearly periodicities, which can be utilized by imputation methods. ...
A cornerstone of the worldwide transition to smart grids are smart meters. Smart meters typically collect and provide energy time series that are vital for various applications, such as grid simulations, fault-detection, load forecasting, load analysis, and load management. Unfortunately, these time series are often characterized by missing values that must be handled before the data can be used. A common approach to handle missing values in time series is imputation. However, existing imputation methods are designed for
power
time series and do not take into account the total energy of gaps, resulting in jumps or constant shifts when imputing
energy
time series. In order to overcome these issues, the present paper introduces the new Copy-Paste Imputation (CPI) method for
energy
time series. The
CPI
method copies data blocks with similar characteristics and pastes them into gaps of the time series while preserving the total energy of each gap. The new method is evaluated on a real-world dataset that contains six shares of artificially inserted missing values between 1 and 30%. It outperforms the three benchmark imputation methods selected for comparison. The comparison furthermore shows that the
CPI
method uses matching patterns and preserves the total energy of each gap while requiring only a moderate run-time.
... Solar power forecasts can be categorized into deterministic and probabilistic forecasts [3]. Some examples of deterministic forecasting methods present in literature can be found in [1,7,13,23,24]. While deterministic forecasts predict only the expected future generation, probabilistic forecasts offer a description of the forecast uncertainty. This additional information helps in managing resources, as well as, in calculating risks associated with future decisions [4,11]. ...
The high penetration of volatile renewable energy sources such as solar make methods for coping with the uncertainty associated with them of paramount importance. Probabilistic forecasts are an example of these methods, as they assist energy planners in their decision-making process by providing them with information about the uncertainty of future power generation. Currently, there is a trend towards the use of deep learning probabilistic forecasting methods. However, the point at which the more complex deep learning methods should be preferred over more simple approaches is not yet clear. Therefore, the current article presents a simple comparison between a long short-term memory neural network and other more simple approaches. The comparison consists of training and comparing models able to provide one-day-ahead probabilistic forecasts for a solar power system. Moreover, the current paper makes use of an open-source dataset provided during the Global Energy Forecasting Competition of 2014 (GEFCom14).
... In the context of smart meters, the time series of recorded electricity consumption or generation typically depend on factors such as weather, human routines, social norms (e.g. weekends or holidays) and many others [16], [17]. These factors often lead to the commonly known patterns with different periodicitymostly daily, weekly, and yearly. ...
A cornerstone of the worldwide transition to smart grids are smart meters. Smart meters typically collect and provide energy time series that are vital for various applications, such as grid simulations, fault-detection, load forecasting, load analysis, and load management. Unfortunately, these time series are often characterized by missing values that must be handled before the data can be used. A common approach to handle missing values in time series is imputation. However, existing imputation methods are designed for power time series and do not take into account the total energy of gaps, resulting in jumps or constant shifts when imputing energy time series. In order to overcome these issues, the present paper introduces the new Copy-Paste Imputation (CPI) method for energy time series. The CPI method copies data blocks with similar properties and pastes them into gaps of the time series while preserving the total energy of each gap. The new method is evaluated on a real-world dataset that contains six shares of artificially inserted missing values between 1 and 30%. It outperforms by far the three benchmark imputation methods selected for comparison. The comparison furthermore shows that the CPI method uses matching patterns and preserves the total energy of each gap while requiring only a moderate run-time.
... ey also examined the status of relationships between the input and the output data and the preprocessing of input data [68]. In addition, González Ordiano et al. appended a contribution in the PV power forecasting topic that consisted of the time-series forecasting techniques, probabilistic forecasting techniques of point forecast, and an outline of time horizons [85]. Moreover, Sobri et al. added a clustering of PV power forecasting methods, in which three main categories were distinguished in this paper. ...
The management of clean energy is usually the key for environmental, economic, and sustainable developments. In the meantime, the energy management system (EMS) ensures the clean energy which includes many sources grouped in a small power plant such as microgrid (MG). In this case, the forecasting methods are used for helping the EMS and allow the high efficiency to the clean energy. The aim of this review paper is providing the necessary data about the basic principles and standards of photovoltaic (PV) power forecasting by stating numerous research studies carried out on the PV power forecasting topic specifically in the short-term time horizon which is advantageous for the EMS and grid operator. At the same time, this contribution can offer a state of the art in different methods and approaches used for PV power forecasting along with a careful study of different time and spatial horizons. Furthermore, this current review paper can support the tenders in the PV power forecasting.
... Generally, any kind of forecasting model [3] could serve as a basis for detecting these anomalies by considering large deviations from the forecasting model as anomalies. This paper examines three methods, namely a Deep Neural Network Regression (DNNR), an Autoencoder with reconstruction (AER), and the encoder of the Autoencoder (EAE), to detect the two aforementioned anomalies. ...
... Forecasting, especially probabilistic forecasting, is essential for decision makers in power systems in order to optimally operate and maintain the grid [1,2]. With the push towards energy systems with high shares of renewable energy sources, forecasting renewable generation, such as wind power, becomes increasingly important. ...
Capturing the uncertainty in probabilistic wind power forecasts is challenging, especially when uncertain input variables, such as the weather play a role. Since ensemble weather predictions aim to capture this uncertainty in the weather system accurately, they can be used to propagate this uncertainty through to subsequent wind power forecasting models. However, as weather ensemble systems are known to be biased and underdispersed, meteorologists post-process the ensembles. This post-processing can successfully correct the biases in the weather variables but has not been evaluated thoroughly in the context of subsequent forecasts, such as wind power generation. The present paper evaluates multiple strategies for applying ensemble post-processing to probabilistic wind power forecasts. We use Ensemble Model Output Statistics (EMOS) as the post-processing method and evaluate four possible strategies: only using the raw ensembles without post-processing, a one-step strategy where only the weather ensembles are post-processed, a one-step strategy where we only post-process the power ensembles, and a two-step strategy where we post-process both the weather and power ensembles. Results show that post-processing improves the probabilistic forecasting accuracy and that the post-processing of the final power ensemble forecast is the crucial step.
... Schätzung der zukünftigen Fallzahlen auf der Basis des gleichen Wochentags der Vorwoche. Solche Ansätze sind aus vielen Anwendungsbereichen mit typischen Wochenrhythmen bekannt, z.B. für die Lastprognose in Energiezeitreihen[4]. Erweiterungen wie die Kompensation von Feiertagen können bei Bedarf vorgenommen werden. ...
Zusammenfassung: Der Beitrag analysiert die Auswirkungen von wöchent-lichen Periodizitäten und zeitlichen Korrekturen auf die Schätzung einer zeitabhängigen Reproduktionszahl R bei Infektionskrankheiten. Zur Reduktion dieser Schwankungen wird eine einfache Methode vorgeschlagen, die auf einem akausalen Filter der Filterlänge 7 und optionalen Schätzungen zukünftiger Fall-zahlen beruht. Dabei werden die gleichen Tage der Vorwoche als Basis für die Schätzungen verwendet, weil sich das in einer anderen Domäne mit wöchent-lichen Periodizitäten-der Lastprognose in Energiezeitreihen-bewährt hat. Akausale Filter vermeiden unerwünschte Zeitverzögerungen, die bei kausalen Filtern auftreten. Die Ergebnisse werden anhand der Fallzahlen von SARS-CoV-2-Infektionen und COVID-19 (Coronavirus Disease 19) in Deutschland mit exi-stierenden Ansätzen des Robert-Koch-Instituts in Deutschland verglichen. Die vorgeschlagene Methode kompensiert wöchentliche Periodizitäten besser und reduziert Phasen mit einer scheinbarenÜberschreitung von R > 1, die oftmals eine besondereöffentliche Aufmerksamkeit hervorrufen. Darüberhinaus werden die Potenziale und Grenzen von verschiedenen Nowcasting-Modellen aufgezeigt, die Fallzahlen auf ein Erkrankungsdatum projizieren.
To address the majority of environmental sustainability issues, like transportation, energy, water management, and biodiversity, artificial intelligence (AI) has grown in importance. The primary focus areas in energy are area neural networks, expert systems, fuzzy logic models, and pattern recognition. This chapter assesses how the use of AI could improve the environment by reducing the effects of agriculture, climate change, water resources, weather forecasting, ocean health, and disaster resilience. The impact of AI on the environment and its sustainable applications are examined in this chapter using qualitative analysis. With an emphasis on AI's role in advancing environmental sustainability, the chapter's main goal is to identify applications of AI for environmentally friendly practices. It can help stakeholders understand international initiatives to improve environmental sustainability by utilizing AI.
In this chapter, we'll take a deep dive into how artificial intelligence (AI) is stepping up to tackle some of our biggest environmental problems. With AI and data technology advancing rapidly, we now have an incredible opportunity to use these tools to protect our planet. We'll explore how AI can gather, analyze, and understand massive amounts of environmental data, giving us valuable insights to make smarter decisions. Through real-world examples and stories, we'll see how AI is being used to model climate change, track wildlife, detect pollution, and manage our precious natural resources. By showing the power of combining data with smart choices, this chapter aims to highlight how AI is playing a crucial role in building a better, greener future for us all.
Energy software plays a crucial role in the energy transition, contributing to the sustainability of the world. This paper presents a systematic summary and review of various software products. Different types of software tools for energy generation, transmission, distribution, storage, and consumption and their features, limitations, and principles of each software tool, along with examples of their application in design and optimization were summarized. Recent development history of energy software, highlighting research gaps, working principles, and underlying architectures were also discussed. Only a few software covers the entire energy system process, and no software focuses solely on energy trading. The United States retains the largest market share (60 % still work). Among 34 surviving software products, only 16 software products have achieved low modeling error, high precision, long-term stability, full process coverage, high compatibility, and strong visualization. In zero-carbon park quantitative tests, the quantitative assessment results indicated software with high accuracy in short-term or transient analysis may require further consideration and comparison when simulating longer time scales. Finally, we discuss emerging insights and future directions, providing recommendations for energy software development. This review aims to assist researchers in better utilizing energy software and expanding the scope of “energy” and “software” issues.
The recent advancements made in the realms of Artificial Intelligence (AI) and Artificial Intelligence of Things (AIoT) have unveiled transformative prospects and opportunities to enhance and optimize the environmental performance and efficiency of smart cities. These strides have, in turn, impacted smart eco-cities, catalyzing ongoing improvements and driving solutions to address complex environmental challenges. This aligns with the visionary concept of smarter eco-cities, an emerging paradigm of urbanism characterized by the seamless integration of advanced technologies and environmental strategies. However, there remains a significant gap in thoroughly understanding this new paradigm and the intricate spectrum of its multifaceted underlying dimensions. To bridge this gap, this study provides a comprehensive systematic review of the burgeoning landscape of smarter eco-cities and their leading-edge AI and AIoT solutions for environmental sustainability. To ensure thoroughness, the study employs a unified evidence synthesis framework integrating aggregative, configurative, and narrative synthesis approaches. At the core of this study lie these subsequent research inquiries: What are the foundational underpinnings of emerging smarter eco-cities, and how do they intricately interrelate, particularly urbanism paradigms, environmental solutions, and data-driven technologies? What are the key drivers and enablers propelling the materialization of smarter eco-cities? What are the primary AI and AIoT solutions that can be harnessed in the development of smarter eco-cities? In what ways do AI and AIoT technologies contribute to fostering environmental sustainability practices, and what potential benefits and opportunities do they offer for smarter eco-cities? What challenges and barriers arise in the implementation of AI and AIoT solutions for the development of smarter eco-cities? The findings significantly deepen and broaden our understanding of both the significant potential of AI and AIoT technologies to enhance sustainable urban development practices, as well as the formidable nature of the challenges they pose. Beyond theoretical enrichment, these findings offer invaluable insights and new perspectives poised to empower policymakers, practitioners, and researchers to advance the integration of eco-urbanism and AI- and AIoT-driven urbanism. Through an insightful exploration of the contemporary urban landscape and the identification of successfully applied AI and AIoT solutions, stakeholders gain the necessary groundwork for making well-informed decisions, implementing effective strategies, and designing policies that prioritize environmental well-being.
In the smart grid of the future, accurate load forecasts on the level of individual clients can help to balance supply and demand locally and to prevent grid outages. While the number of monitored clients will increase with the ongoing smart meter rollout, the amount of data per client will always be limited. We evaluate whether a Transformer load forecasting model benefits from a transfer learning strategy, where a global univariate model is trained on the load time series from multiple clients. In experiments with two datasets containing load time series from several hundred clients, we find that the global training strategy is superior to the multivariate and local training strategies used in related work. On average, the global training strategy results in 21.8% and 12.8% lower forecasting errors than the two other strategies, measured across forecasting horizons from one day to one month into the future. A comparison to linear models, multi-layer perceptrons and LSTMs shows that Transformers are effective for load forecasting when they are trained with the global training strategy.
Accurately predicting the prices of financial time series is essential and challenging for the financial sector. Owing to recent advancements in deep learning techniques, deep learning models are gradually replacing traditional statistical and machine learning models as the first choice for price forecasting tasks. This shift in model selection has led to a notable rise in research related to applying deep learning models to price forecasting, resulting in a rapid accumulation of new knowledge. Therefore, we conducted a literature review of relevant studies over the past 3 years with a view to aiding researchers and practitioners in the field. This review delves deeply into deep learning‐based forecasting models, presenting information on model architectures, practical applications, and their respective advantages and disadvantages. In particular, detailed information is provided on advanced models for price forecasting, such as Transformers, generative adversarial networks (GANs), graph neural networks (GNNs), and deep quantum neural networks (DQNNs). The present contribution also includes potential directions for future research, such as examining the effectiveness of deep learning models with complex structures for price forecasting, extending from point prediction to interval prediction using deep learning models, scrutinizing the reliability and validity of decomposition ensembles, and exploring the influence of data volume on model performance.
This article is categorized under: Technologies > Prediction
Technologies > Artificial Intelligence
Deep learning is a subset of the Machine learning method, which is a vital factor in the Energy Industrial sector, which helps optimize the contracting methods and minimize the overall cost in the expenses to produce energy. This research paper aims to identify the approaches that can be used to identify the valuable depending factors in the data sets and correct the missing data value, which can be later on used to analyze the data better. The data set, which we worked on within this research, is a data set of power plant productions of a decade in the US. Then deep learning will help us to be able to visualize the shortcoming in the approach and come up with an appropriate solution. Inspiration for this research was an old project, which included a smart power control system and a solar‐wind hybrid power source, so by this research, we can identify the areas where better and smarter systems can be implemented replacing conventional methods and how to tackle the old data sets.
Capturing the uncertainty in probabilistic wind power forecasts is challenging, especially when uncertain input variables, such as the weather, play a role. Since ensemble weather predictions aim to capture the uncertainty in the weather system, they can be used to propagate this uncertainty through to subsequent wind power forecasting models. However, as weather ensemble systems are known to be biassed and underdispersed, meteorologists post‐process the ensembles. This post‐processing can successfully correct the biasses in the weather variables but has not been evaluated thoroughly in the context of subsequent forecasts, such as wind power generation forecasts. The present paper evaluates multiple strategies for applying ensemble post‐processing to probabilistic wind power forecasts. We use Ensemble Model Output Statistics (EMOS) as the post‐processing method and evaluate four possible strategies: only using the raw ensembles without post‐processing, a one‐step strategy where only the weather ensembles are post‐processed, a one‐step strategy where we only post‐process the power ensembles and a two‐step strategy where we post‐process both the weather and power ensembles. Results show that post‐processing the final wind power ensemble improves forecast performance regarding both calibration and sharpness whilst only post‐processing the weather ensembles does not necessarily lead to increased forecast performance.
Data mining technology is more and more widely used in the daily load forecasting of natural gas systems. It is still difficult to carry out high-precision, timely intraday load forecasting and intraday load dynamic characteristics clustering for natural gas systems. Based on data mining technology, this paper proposes a stable intraday load forecasting method for the natural gas flow state-space model. The load sensitivity under the current operating conditions of the system is obtained by calculation; the sample space of the state space is established through data processing; the partitions under different clustering radii are calculated; and the best intraday load flow is obtained through the state space effectiveness evaluation method. The experimental results show that the model load forecasting accuracy and relative error reached 98.5% and 0.026, respectively, which solved the problem of processing the long-term accumulated historical data of gas intra-day load. At the same time, the amount of data calculation was reduced by 33.6%, which effectively promoted the quantification of intraday load influencing factors and qualitative analysis.
Forecasting is of an immense importance in energy markets, it aims to build accurate forecasting models to inspect future scenarios. It can also be used to monitor energy profiles to detect faults and other security concerns. Many tools and services have been introduced to automate the forecasting process. However, the ‘ideal’ tool highly depends on how much it perfectly fulfills the desired behavior in the targeted application. In this paper, we introduce a configurable energy forecasting tool, which extends a set of machine learning models to provide dynamic energy forecasting services. The developed tool aims mainly to predict energy generation/consumption, build forecasting models, compare predictions, and fine-tune prediction models. Then, we utilize this tool to conduct two case studies. The first aims to compare the performance of different prediction models in residential and office buildings, using multiple experiments with variations of input fields. The second one investigates the role of the energy forecasting tool to raise awareness regarding security incidents in a shop floor. Results from both case studies emphasized the prospective role of the developed tool in energy forecasting and security awareness.
This paper addresses the cyber resilience issues of multi-vector energy distribution systems (MEDS) caused by false data injection FDI, considering the uncertainty from renewable resources. A novel two-stage distributionally robust optimization (DRO) is proposed to realize the day-ahead and real-time resilience improvement. The first stage determines an initial plan for day-ahead reserve preparation and the second stage makes adjustment and takes resilience-based actions after potential load redistribution (LR) attacks and renewable output deviation. The ambiguity set is based on both the Wasserstein distance and moment information. Compared to robust optimization which considers the worst case, DRO yields less-conservative solutions and thus provides more economic operation schemes. The Wasserstein-metric based ambiguity set enables to provide additional flexibility hedging against renewable uncertainty. Case studies are demonstrated on two representative MEDS networked with energy hubs, i.e., a 33-bus-20-node MEDS and a 69-bus-20-node-MEDS, illustrating the effectiveness of the proposed cyber-secured model.
Storage is a key concept to cope with a growing supply of volatile renewable energy. Compared with genuine grid storage (powerX), domestic heating and cooling, operated as flexible loads (power to heat), promises to lower infrastructure costs and offers potential for a short-term demand response. Before designing new service markets, the true technical potential must be assessed. Herein, storage capacity is estimated by exploiting the entire thermally usable building mass as back-end storage. Balancing power is determined from the steady-state coincidence factor of buildings seen as thermostatically controlled load (TCL) populations. Results are reported for the residential and tertiary sectors in Germany by considering six classes of heating and cooling equipment. One advantage of the method is simplicity, which results in closed-form estimates derived top-down from nationwide and publicly available data on the building stock.
The present contribution offers evidence regarding the possibility of obtaining reasonable photovoltaic power forecasts without using weather data and with simple data-driven models. The lack of weather data as input stems from the fact that the constant obtainment of forecast weather data might become too expensive or that communication with weather services might fail, but still accurate planning and scheduling decisions have to be conducted. Therefore, accurate one-day ahead forecasting models with only information of past generated power as input for offline photovoltaic systems or as backup in case of communication failures are of interest. The results contained in the present contribution, obtained using a freely available dataset, provide a baseline with which more complex forecasting models can be compared. Additionally, it will also be shown that the presented weather-free data-driven models provide better forecasts than a trivial persistence technique for different forecast horizons. The methodology used in the present work for the data preprocessing and the creation and validation of forecasting models has a generalization capacity and thus can be used for different types of time series as well as different data mining techniques.
Electricity production via solar energy is tackled via short-term forecasts and risk management. Our main tool is a new setting on time series. It allows the definition of "confidence bands" where the Gaussian assumption, which is not satisfied by our concrete data, may be abandoned. Those bands are quite convenient and easily implementable. Numerous computer simulations are presented.
Additional flexibilities on the demand side can be obtained by using price signals to change the consumption behavior of household electricity customers. The present contribution proposes a new theoretically motivated demand response model type called virtual storage. First, the basic model structure of several virtual storage models is introduced. All of these models are based on a system of difference equations that describe load reductions/increases in response to price signals. The virtual storage models differ thereafter in how past or prognosis-based future price information is considered. After a description of a proposed model validation strategy, the model behavior of several virtual storage models is compared with some of the common demand response model types and with real customer responses to a price signal. Thus, a model comparison is performed on the basis of a real smart meter data set (Olympic Peninsula Project).
Short-term load forecasting at the distribution level predicts the load of substations, feeders, transformers, and possibly customers from half an hour to one week ahead. Effective forecasting is important for the planning and operation of distribution systems. The problem, however, is difficult in view of complicated load features, the large number of distribution-level nodes, and possible switching operations. In this paper, a new forecasting approach within the hierarchical structure is presented to solve these difficulties. Load of the root node at any user-defined subtree is first forecast by a wavelet neural network with appropriate inputs. Child nodes categorized as “regular” and “irregular” based on load pattern similarities are then forecast separately. Load of a regular child node is simply forecast as the proportion from the parent node load forecast while the load of an irregular child node is forecast by an individual neural network model. Switching operation detection and follow-up adjustments are also performed to capture abnormal changes and improve the forecasting accuracy. This new approach captures load characteristics of nodes at different levels, takes advantage of pattern similarities between a parent node and its child nodes, detects abnormalities, and provides high quality forecasts as demonstrated by two practical datasets.
Electrical load forecasting plays a vital role in order to achieve the concept of next generation power system such as smart grid, efficient energy management and better power system planning. As a result, high forecast accuracy is required for multiple time horizons that are associated with regulation, dispatching, scheduling and unit commitment of power grid. Artificial Intelligence (AI) based techniques are being developed and deployed worldwide in on Varity of applications, because of its superior capability to handle the complex input and output relationship. This paper provides the comprehensive and systematic literature review of Artificial Intelligence based short term load forecasting techniques. The major objective of this study is to review, identify, evaluate and analyze the performance of Artificial Intelligence (AI) based load forecast models and research gaps. The accuracy of ANN based forecast model is found to be dependent on number of parameters such as forecast model architecture, input combination, activation functions and training algorithm of the network and other exogenous variables affecting on forecast model inputs. Published literature presented in this paper show the potential of AI techniques for effective load forecasting in order to achieve the concept of smart grid and buildings.
The accurate forecasting of energy production from renewable sources represents an important topic also looking at different national authorities that are starting to stimulate a greater responsibility towards plants using non-programmable renewables. In this paper the authors use advanced hybrid evolutionary techniques of computational intelligence applied to photovoltaic systems forecasting, analyzing the predictions obtained by comparing different definitions of the forecasting error.
Rapid growth in wind power, as well as increase on wind generation, requires serious research in various fields. Because wind power is weather dependent, it is variable and intermittent over various time-scales. Thus accurate forecasting of wind power is recognized as a major contribution for reliable large-scale wind power integration. Wind power forecasting methods can be used to plan unit commitment, scheduling and dispatch by system operators, and maximize profit by electricity traders. In addition, a number of wind power models have been developed internationally, such as WPMS, WPPT, Prediktor, ARMINES, Previento, WPFS Ver1.0 etc. This paper provides a review on comparative analysis on the foremost forecasting models, associated with wind speed and power, based on physical methods, statistical methods, hybrid methods over different time-scales. Furthermore, this paper gives emphasis on the accuracy of these models and the source of major errors, thus problems and challenges associated with wind power prediction are presented. (C) 2011 Published by Elsevier Ltd. Selection and/or peer-review under responsibility of University of Electronic Science and Technology of China (UESTC)
A variety of methods and ideas have been tried for electricity price forecasting (EPF) over the last 15 years, with varying degrees of success. This review article aims to explain the complexity of available solutions, their strengths and weaknesses, and the opportunities and threats that the forecasting tools offer or that may be encountered. The paper also looks ahead and speculates on the directions EPF will or should take in the next decade or so. In particular, it postulates the need for objective comparative EPF studies involving (i) the same datasets, (ii) the same robust error evaluation procedures, and (iii) statistical testing of the significance of one model’s outperformance of another.
The Smart Grid, regarded as the next generation power grid, uses two-way flows of electricity and information to create a widely distributed automated energy delivery network. In this article, we survey the literature till 2011 on the enabling technologies for the Smart Grid. We explore three major systems, namely the smart infrastructure system, the smart management system, and the smart protection system. We also propose possible future directions in each system. colorred{Specifically, for the smart infrastructure system, we explore the smart energy subsystem, the smart information subsystem, and the smart communication subsystem.} For the smart management system, we explore various management objectives, such as improving energy efficiency, profiling demand, maximizing utility, reducing cost, and controlling emission. We also explore various management methods to achieve these objectives. For the smart protection system, we explore various failure protection mechanisms which improve the reliability of the Smart Grid, and explore the security and privacy issues in the Smart Grid.
The increasing amount of power generation from weather-dependent renewable sources in Germany is projected to lead to a considerable number of hours in which power generation exceeds power demand. One possibility to take advantage of this power surplus is through the Power-to-Heat technology. As combined heat and power (CHP)-plants can be upgraded relatively easily with a Power-to-Heat facility, a huge potential can be developed in German district heating grids which are mainly served by CHP-plants. In this paper the potential of the Power-to-Heat technology in district heating grids in Germany is evaluated for the years 2015 to 2030 under different assumptions.
The fourth edition of this popular graduate textbook, like its predecessors, presents a balanced and comprehensive treatment of both time and frequency domain methods with accompanying theory. Numerous examples using nontrivial data illustrate solutions to problems such as discovering natural and anthropogenic climate change, evaluating pain perception experiments using functional magnetic resonance imaging, and monitoring a nuclear test ban treaty.
The book is designed as a textbook for graduate level students in the physical, biological, and social sciences and as a graduate level text in statistics. Some parts may also serve as an undergraduate introductory course. Theory and methodology are separated to allow presentations on different levels. In addition to coverage of classical methods of time series regression, ARIMA models, spectral analysis and state-space models, the text includes modern developments including categorical time series analysis, multivariate spectral methods, long memory series, nonlinear models, resampling techniques, GARCH models, ARMAX models, stochastic volatility, wavelets, and Markov chain Monte Carlo integration methods.
This edition includes R code for each numerical example in addition to Appendix R, which provides a reference for the data sets and R scripts used in the text in addition to a tutorial on basic R commands and R time series. An additional file is available on the book’s website for download, making all the data sets and scripts easy to load into R.
• Student-tested and improved
• Accessible and complete treatment of modern time series analysis
• Promotes understanding of theoretical concepts by bringing them into a more practical context
• Comprehensive appendices covering the necessities of understanding the mathematics of time series analysis
• Instructor's Manual available for adopters
New to this edition:
• Introductions to each chapter replaced with one-page abstracts
• All graphics and plots redone and made uniform in style
• Bayesian section completely rewritten, covering linear Gaussian state space models only
• R code for each example provided directly in the text for ease of data analysis replication
• Expanded appendices with tutorials containing basic R and R time series commands
• Data sets and additional R scripts available for download on Springer.com
• Internal online links to every reference (equations, examples, chapters, etc.)
•
As crude oil price is influenced by numerous factors, capturing its behavior precisely is quite challenging, and thus leads to the difficulty of forecasting. In this study, a deep learning ensemble approach is proposed to deal with this problem. In our approach, two techniques are utilized. One is an advanced deep neural network model named stacked denoising autoencoders (SDAE) which is used to model the nonlinear and complex relationships of oil price with its factors. The other is a powerful ensemble method named bootstrap aggregation (bagging) which generates multiple data sets for training a set of base models (SDAEs). Our approach combines the merits of these two techniques and is especially suitable for oil price forecasting. In the empirical study, the WTI crude oil price series are investigated and 198 economic series are used as exogenous variables. Our approach is tested against some competing approaches and shows superior forecasting ability that is statistically proved by three tests.
In recent years, estimating the power output of inherently intermittent and potentially distributed renewable energy sources has become a major scientific and societal concern. In this paper, we provide an algorithmic framework, along with an interactive web-based tool, to enable short-to-middle term forecasts of photovoltaic (PV) systems and wind generators output. Importantly, we propose a generic PV output estimation method, the backbone of which is a solar irradiance approximation model that incorporates free-to-use, readily available meteorological data coming from online weather stations. The model utilizes non-linear approximation components for turning cloud-coverage into radiation forecasts, such as an MLP neural network with one hidden layer. We present a thorough evaluation of the proposed techniques, and show that they can be successfully employed within a broad geographical region (the Mediterranean belt) and come with specific performance guarantees. Crucially, our methods do not rely on complex and expensive weather models and data, and our web-based tool can be of immediate use to the community as a simulation data acquisition platform.
Variability of solar resource poses difficulties in grid management as solar penetration rates rise continuously. Thus, the task of solar power forecasting becomes crucial to ensure grid stability and to enable an optimal unit commitment and economical dispatch. Several forecast horizons can be identified, spanning from a few seconds to days or weeks ahead, as well as spatial horizons, from single site to regional forecasts. New techniques and approaches arise worldwide each year to improve accuracy of models with the ultimate goal of reducing uncertainty in the predictions. This paper appears with the aim of compiling a large part of the knowledge about solar power forecasting, focusing on the latest advancements and future trends. Firstly, the motivation to achieve an accurate forecast is presented with the analysis of the economic implications it may have. It is followed by a summary of the main techniques used to issue the predictions. Then, the benefits of point/regional forecasts and deterministic/probabilistic forecasts are discussed. It has been observed that most recent papers highlight the importance of probabilistic predictions and they incorporate an economic assessment of the impact of the accuracy of the forecasts on the grid. Later on, a classification of authors according to forecast horizons and origin of inputs is presented, which represents the most up-to-date compilation of solar power forecasting studies. Finally, all the different metrics used by the researchers have been collected and some remarks for enabling a fair comparison among studies have been stated.
This paper proposes a generic framework for probabilistic energy forecasting, and discusses the application of the method to several tracks in the 2014 Global Energy Forecasting Competition (GEFCom2014). The proposed method uses a multiple quantile regression approach to predict a full distribution over possible future energy outcomes, uses the alternating direction method of multipliers to solve the optimization problems resulting from this quantile regression formulation efficiently, and uses a radial basis function network to capture the non-linear dependencies on the input data. For the GEFCom2014 competition, the approach proved general enough to obtain one of the top five ranks in three tracks, solar, wind, and price forecasting, and it was also ranked seventh in the final load forecasting track.
The energy industry has been going through a significant modernization process over the last decade. Its infrastructure is being upgraded rapidly. The supply, demand and prices are becoming more volatile and less predictable than ever before. Even its business model is being challenged fundamentally. In this competitive and dynamic environment, many decision-making processes rely on probabilistic forecasts to quantify the uncertain future. Although most of the papers in the energy forecasting literature focus on point or single-valued forecasts, the research interest in probabilistic energy forecasting research has taken off rapidly in recent years. In this paper, we summarize the recent research progress on probabilistic energy forecasting. A major portion of the paper is devoted to introducing the Global Energy Forecasting Competition 2014 (GEFCom2014), a probabilistic energy forecasting competition with four tracks on load, price, wind and solar forecasting, which attracted 581 participants from 61 countries. We conclude the paper with 12 predictions for the next decade of energy forecasting.
Lars Dannecker developed a novel online forecasting process that significantly improves how forecasts are calculated. It increases forecasting efficiency and accuracy, as well as allowing the process to adapt to different situations and applications. Improving the forecasting efficiency is a key pre-requisite for ensuring stable electricity grids in the face of an increasing amount of renewable energy sources. It is also important to facilitate the move from static day ahead electricity trading towards more dynamic real-time marketplaces. The online forecasting process is realized by a number of approaches on the logical as well as on the physical layer that we introduce in the course of this book. Nominated for the Georg-Helm-Preis 2015 awarded by the Technische Universität Dresden.
Probabilistic forecasting provides quantitative information of energy uncertainty, which is very essential for making better decisions in power system operation with increasing penetration of wind power and solar power. On the basis of k-nearest neighbor and kernel density estimator method, this paper presents a general framework of probabilistic forecasts for renewable energy generation. Firstly, the k-nearest neighbor algorithm is modified to find the days with similar weather conditions in historical dataset. Then, kernel density estimator method is applied to derive the probability density from k nearest neighbors. This approach is demonstrated by an application in probabilistic solar power forecasting. The effectiveness of our proposed approach is validated with the real data provided by Global Energy Forecasting Competition 2014.
The share of renewable generation (RG) in the energy mix has seen constant growth in recent years. RG is volatile and not (fully) controllable. Consequently, the alignment of stochastic demand with supply, which is fundamental for ensuring grid stability, becomes more difficult. The utilization of demand side flexibility as well as RG portfolio design are attractive opportunities to avoid excessive investments in conventional power plants and costs for balancing power. This paper provides a comprehensive centralized scheduling model to exploit demand flexibility from residential devices. We analyze the monetary value of households for demand response (DR) by determining the potential of various current and possible available future end consumer devices to reduce generation costs of a flexibility aggregator in a microgrid with a large share of RG. Furthermore, we identify key characteristics affecting the value of demand flexibility and derive recommendations for an aggregator's RG portfolio structure. Our simulation results indicate that electric vehicles, stationary batteries, and storage heaters are the most promising devices for residential DR. Furthermore, we show that the potential of a device to directly utilize intermittent RG is largely influenced by the composition of the renewable energy source portfolio.
Load forecasting has been a fundamental business problem since the inception of the electric power industry. Over the past 100 plus years, both research efforts and industry practices in this area have focused primarily on point load forecasting. In the most recent decade, though, the increased market competition, aging infrastructure and renewable integration requirements mean that probabilistic load forecasting has become more and more important to energy systems planning and operations. This paper offers a tutorial review of probabilistic electric load forecasting, including notable techniques, methodologies and evaluation methods, and common misunderstandings. We also underline the need to invest in additional research, such as reproducible case studies, probabilistic load forecast evaluation and valuation, and a consideration of emerging technologies and energy policies in the probabilistic load forecasting process.
In Energy Lab 2.0, the interplay of different forms of energy on different value chains is investigated. Novel concepts to stabilize the volatile energy supply of renewables by the use of storage systems and mainly by applying to-be-developed tools and algorithms of the information and communication technology sector are sought. Hence, a key element of Energy Lab 2.0 is the smart energies system simulation and control center. This consists of three parts: a power-hardware-in-the-loop experimental field, an energy grid simulation and analysis laboratory, and a control, monitoring, and visualization center. For these three labs, big data technologies, advanced control methods, and reliable, safe, and secure software structures are of equal importance. As an example, a data processing pipeline to create power flow simulation models from raw Open Street Map data, statistical databases, and geodata is presented and discussed.
Due to the challenge of climate and energy crisis, renewable energy generation including solar generation has experienced significant growth. Increasingly high penetration level of photovoltaic (PV) generation arises in smart grid. Solar power is intermittent and variable, as the solar source at the ground level is highly dependent on cloud cover variability, atmospheric aerosol levels, and other atmosphere parameters. The inherent variability of large-scale solar generation introduces significant challenges to smart grid energy management. Accurate forecasting of solar power/irradiance is critical to secure economic operation of the smart grid. This paper provides a comprehensive review of the theoretical forecasting methodologies for both solar resource and PV power. Applications of solar forecasting in energy management of smart grid are also investigated in detail.
Day-ahead electricity prices are generally used as reference prices for decisions done in energy trading, e.g. purchase and sale strategies are typically based on the day-ahead spot prices. Therefore, wellperforming forecast methods for day-ahead electricity prices are essential for energy traders and supply companies. In this paper, a methodology based on artificial neuronal networks (ANN) is presented to forecast electricity prices. As the performance of an ANN forecast model depends on appropriate input parameter sets, the focus is set on the selection and preparation of fundamental data that has a noticeable impact on electricity prices. This is done with the help of different cluster algorithms, but also by comparing the results of the pre-selected model configurations in combination with different input parameter settings. After the determination of the optimal input parameters, affecting day-ahead electricity prices, and wellperforming ANN configuration, the developed ANN model is applied for in-sample and out-of-sample analyses. The results show that the overall methodology leads to well-fitting electricity price forecasts, whereas forecast errors are as low as or even lower than other forecast models for electricity prices known from the literature.
Despite the rapid uptake of small-scale solar photovoltaic (PV) systems in recent years, public availability of generation and load data at the household level remains very limited. Moreover, such data are typically measured using bi-directional meters recording only PV generation in excess of residential load rather than recording generation and load separately. In this paper, we report a publicly available dataset consisting of load and rooftop PV generation for 300 de-identified residential customers in an Australian distribution network, with load centres covering metropolitan Sydney and surrounding regional areas. The dataset spans a 3-year period, with separately reported measurements of load and PV generation at 30-min intervals. Following a detailed description of the dataset, we identify several means by which anomalous records (e.g. due to inverter failure) are identified and excised. With the resulting ‘clean’ dataset, we identify key customer-specific and aggregated characteristics of rooftop PV generation and residential load.
By influencing the demand side by means of price signals (Demand Response) additional flexibility potential in electric supply systems can be provided. However, by influencing the demand side typical consumption patterns of previously unaffected consumers are changed. This will lead to increasing uncertainty in load forecasting. This paper deals with the forecast of load time series in consideration of price-based consumption influence. Additional requirements for load forecasting methods resulting from the price elastic consumption behaviour are analysed in this paper. Furthermore, the model residuals of established model approaches will be analysed to explain the disturbance characteristic caused by the price elasticity. Finally, the impact of the model residuals on the load forecast was investigated.
Renewable power output is an important factor in scheduling and for improving balanced area control performance. This investigation develops an evolutionary seasonal decomposition least-square support vector regression (ESDLS-SVR) to forecast monthly solar power output. The construction of the ESDLS-SVR uses seasonal decomposition and least-square support vector regression (LS-SVR). Genetic algorithms (GA) are used simultaneously to select the parameters of the LS-SVR. Monthly solar power output data from Taiwan Power Company are used. Empirical results indicate that the proposed forecasting system demonstrates a superior performance in terms of forecasting accuracy. A comparative study has been introduced showing that the ESDLS-SVR model performance is better than autoregressive integrated moving average (ARIMA), seasonal autoregressive integrated moving average (SARIMA), generalized regression neural network (GRNN) and LS-SVR models.
The rapid increase in energy demand requires effective measures to plan and optimize resources for efficient energy production within a smart grid environment. This paper presents a data driven approach to forecasting heat load for multi-family apartment buildings in a District Heating System (DHS). The forecasting model is built using six and eleven weeks of data from five building substations. The external factors and internal factors influencing the heat load in substations are parameters used as our model's input. Short-term forecast models are generated using four supervised Machine Learning (ML) techniques: Support Vector Regression (SVR), Regression Tree, Feed Forwards Neural Network (FFNN) and Multiple Linear Regression (MLR). Performance comparison among these ML methods was carried out. The effects of combining the internal and external factors influencing heat load at substations was studied. The models are evaluated with varying horizon up to 24-hours ahead. The results show that SVR has the best accuracy of 5.6% MAPE for the best-case scenario.
This chapter is devoted to hybrid wind systems. It describes the different configurations and the different combinations of hybrid wind systems. Different synoptic schemes and simulation applications are also presented. Different synoptic schemes and simulation applications are also presented. We have presented a study of an hybrid wind/PV//Diesel system with battery storage. The application is under Matlab/Simulink during three different profiles of insolation and wind speed (low, medium and high conditions). This study seems interesting and can be applied to electrification or a pumping system for example.
The increasing installation of volatile renewable energy sources like photovoltaics and wind enforces the need for flexibility options to match the renewable generation with the demand. One of these options is Demand Side Management (DSM) in the context of building energy systems combined with thermal storage systems. This paper discusses such concepts for DSM. A method for analyzing the flexibility that is needed to maintain the stability of the electrical grid is presented followed by the restrictions that are caused by meeting the heat demand and satisfying the comfort criteria of the residents. Approaches for simultaneously fulfilling these constraints as well as matching the flexibility needs of the electrical grid and the flexibility provided by the local building energy systems are discussed. To enhance the analysis options for the shown systems, a simulation platform that covers the electrical grid simulation, the building systems’ simulation and the control strategies is presented. This platform can be used to analyze different scenarios of building energy systems with different penetrations of renewable energy sources and different building types.
The issue of limited fossil fuels combined with the vast technological improvements in recent years has initiated numerous installations of renewable power production, particularly in form of photovoltaic cells and wind turbines. Since the volatile character of wind and solar radiation leads to a fluctuating power production, these renewables are incapable of providing reliable base load power. To enable the transition to a renewable energy system, large-scale energy storage is required to compensate for short-term and seasonal imbalances and to save temporary excess power. Due to the order of magnitude involved, this can best be achieved by converting electricity into hydrogen via electrolysis, a process that is also called “power to gas”. Hereby, hydrogen can serve as a link combining the electricity, traffic and heating sector into one energy market. This paper presents the process chains of different power-to-gas paths, including different transformation technologies, which it evaluates with regard to their suitability for applications, the optional methanation step including the necessary production of CO2, distribution options and geological storage options as well as end-user applications. Finally, the use of hydrogen and methane in transportation and reconversion to power are compared from the economic point of view.
A probabilistic forecast takes the form of a predictive probability distribution over future quantities or events of interest. Probabilistic forecasting aims to maximize the sharpness of the predictive distributions, subject to calibration, on the basis of the available information set. We formalize and study notions of calibration in a prediction space setting. In practice, probabilistic calibration can be checked by examining probability integral transform (PIT) histograms. Proper scoring rules such as the logarithmic score and the continuous ranked probability score serve to assess calibration and sharpness simultaneously. As a special case, consistent scoring functions provide decision-theoretically coherent tools for evaluating point forecasts. We emphasize methodological links to parametric and nonparametric distributional regression techniques, which attempt to model and to estimate conditional distribution functions; we use the context of statistically postprocessed ensemble forecasts in numerical weather prediction as an example. Throughout, we illustrate concepts and methodologies in data examples.
In recent years, estimating the power output of in-herently intermittent and potentially distributed renewable energy sources has become a major scientific and societal concern. In this paper, we provide an algorithmic framework, along with an inter-active web-based tool, to enable short-to-middle term forecasts of photovoltaic (PV) systems and wind generators output. Importantly, we propose a generic PV output estimation method, the backbone of which is a solar irradiance approximation model that incorpo-rates free-to-use, readily available meteorological data coming from online weather stations. The model utilizes non-linear approxima-tion components for turning cloud-coverage into radiation forecasts, such as an MLP neural network with one hidden layer. We present a thorough evaluation of the proposed techniques, and show that they can be successfully employed within a broad geographical region (the Mediterranean belt) and come with specific performance guar-antees. Crucially, our methods do not rely on complex and expensive weather models and data, and our web-based tool can be of immedi-ate use to the community as a simulation data acquisition platform.
This paper proposes an appropriate combinational approach which is based on improved BP neural network for short-term gas load forecasting, and the network is optimized by the real-coded genetic algorithm. Firstly, several kinds of modifications are carried out on the standard neural network to accelerate the convergence speed of network, including improved additional momentum factor, improved self-adaptive learning rate and improved momentum and self-adaptive learning rate. Then, it is available to use the global search capability of optimized genetic algorithm to determine the initial weights and thresholds of BP neural network to avoid being trapped in local minima. The ability of GA is enhanced by cat chaotic mapping. In light of the characteristic of natural gas load for Shanghai, a series of data preprocessing methods are adopted and more comprehensive load factors are taken into account to improve the prediction accuracy. Such improvements facilitate forecasting efficiency and exert maximum performance of the model. As a result, the integration model improved by modified additional momentum factor gets more ideal solutions for short-term gas load forecasting, through analyses and comparisons of the above several different combinational algorithms.
To improve real-time control performance and reduce possible negative impacts of photovoltaic (PV) systems, an accurate forecasting of PV output is required, which is an important function in the operation of an energy management system (EMS) for distributed energy resources. In this paper, a weather-based hybrid method for 1-day ahead hourly forecasting of PV power output is presented. The proposed approach comprises classification, training, and forecasting stages. In the classification stage, the self-organizing map (SOM) and learning vector quantization (LVQ) networks are used to classify the collected historical data of PV power output. The training stage employs the support vector regression (SVR) to train the input/output data sets for temperature, probability of precipitation, and solar irradiance of defined similar hours. In the forecasting stage, the fuzzy inference method is used to select an adequate trained model for accurate forecast, according to the weather information collected from Taiwan Central Weather Bureau (TCWB). The proposed approach is applied to a practical PV power generation system. Numerical results show that the proposed approach achieves better prediction accuracy than the simple SVR and traditional ANN methods.
In this paper, a simple but accurate approach for short-term forecasting of the power produced by a Large-Scale Grid Connected Photovoltaic Plant (LS-GCPV) is presented. A 1-year database of solar irradiance, cell temperature and power output produced by a 1-MWp photovoltaic plant located in Southern Italy is used for developing three distinct artificial neural network (ANN) models, to be applied to three typical types of day (sunny, partly cloudy and overcast). The possibility of obtaining accurate results by using solely the monitored data rather than knowing the actual architecture and details of the plant is a notable advantage; in particular, the method’s reliability gives to operation and maintenance and to grid operators excellent confidence in the evaluation of the performance of the plant and in the programming of the dispatching plans, respectively.
The share of renewables in the primary energy demand of Germany has increased from 1.3% in 1990 to 11.7% in 2013. The plans are to increase it further, by the year 2050, to 50% or even 100%, by carrying out the national Energiewende (energy change/switch) programme. In this paper we analyze the current energy scenario and recall goals of the Energiewende programme. We also identify infrastructural and technological bottlenecks which are likely to affect the programme execution. By doing so, the enormous scale of this national venture becomes apparent and it is clear that its execution will last for several generations and perhaps even for a century or two. Additionally, we introduce an idea of Exclusively Green Energy Communities (EGECs) as an Energiewende mile-stone. Due to the dimensions of the tasks, the energy change/switch programme will last for several generations and perhaps even for a century or two. Realization of this overwhelming national venture has to proceed through milestones and demonstration projects. We propose demo-projects, called Green Energy Communities (GECs), to examine with what efforts and under what constraints such communities can operate using exclusively renewable forms of energy.
Electric load profiles are useful for accurate load forecasting, network planning and optimal generation capacity. They represent electricity demand patterns and are to a large extent predictable. However, new and heavier loads (heat pumps and electric vehicles), distributed generation, and home energy management technologies will change future energy consumption pattern of residential customers. This article analyses future residential load profiles via modelling and simulation of residential loads and distributed generations. The household base loads are represented by synthetic load profiles. Mathematical models are implemented for heat pumps, micro combined heat and power units, electric vehicles and photovoltaic systems. Scenario-based simulations are performed with different combination and penetration levels of load and generation technologies for different seasons. The results of the analyses show that with varying penetration levels of distributed generation and heavy loads, future residential load profiles will be more dynamic and dependent on multiple factors deviating from the classical demand pattern.
The rapid development of human population, buildings and technology application currently has caused electric consumption to grow rapidly. Therefore, efficient energy management and forecasting energy consumption for buildings are important in decision-making for effective energy saving and development in particular places. This paper reviews the building electrical energy forecasting method using artificial intelligence (AI) methods such as support vector machine (SVM) and artificial neural networks (ANN). Both methods are widely used in the field of forecasting and their aim on finding the most accurate approach is ever continuing. Besides the already existing single method of forecasting, the hybridization of the two forecasting methods has the potential to be applied for more accurate results. Further research works are currently ongoing, regarding the potential of hybrid method of Group Method of Data Handling (GMDH) and Least Square Support Vector Machine (LSSVM), or known as GLSSVM, to forecast building electrical energy consumption.
Streamflow forecasting plays a critical role in nearly all aspects of water resources planning and management. In this work, Gaussian Process Regression (GPR), an effective kernel-based machine learning algorithm, is applied to probabilistic streamflow forecasting. GPR is built on Gaussian process, which is a stochastic process that generalizes multivariate Gaussian distribution to infinite-dimensional space such that distributions over function values can be defined. The GPR algorithm provides a tractable and flexible hierarchical Bayesian framework for inferring the posterior distribution of streamflows. The prediction skill of the algorithm is tested for one-month-ahead prediction using the MOPEX database, which includes long-term hydrometeorological time series collected from 438 basins across the U.S. from 1948 to 2003. Comparisons with linear regression and artificial neural network models indicate that GPR outperforms both regression methods in most cases. The GPR prediction of MOPEX basins is further examined using the Budyko framework, which helps to reveal the close relationships among water-energy partitions, hydrologic similarity, and predictability. Flow regime modification and the resulting loss of predictability have been a major concern in recent years because of climate change and anthropogenic activities. The persistence of streamflow predictability is thus examined by extending the original MOPEX data records to 2012. Results indicate relatively strong persistence of streamflow predictability in the extended period, although the low-predictability basins tend to show more variations. Because many low-predictability basins are located in regions experiencing fast growth of human activities, the significance of sustainable development and water resources management can be even greater for those regions.
The Global Energy Forecasting Competition (GEFCom2012) attracted hundreds of participants worldwide, who contributed many novel ideas to the energy forecasting field. This paper introduces both tracks of GEFCom2012, hierarchical load forecasting and wind power forecasting, with details on the aspects of the problem, the data, and a summary of the methods used by selected top entries. We also discuss the lessons learned from this competition from the organizers’ perspective. The complete data set, including the solution data, is published along with this paper, in an effort to establish a benchmark data pool for the community.
In this work, a new hybrid model for short-term power forecasting of a grid-connected photovoltaic plant is introduced. The new model combines two well-known methods: the seasonal auto-regressive integrated moving average method (SARIMA) and the support vector machines method (SVMs). An experimental database of the power produced by a small-scale 20 kWp GCPV plant is used to develop and verify the effectiveness of the proposed model in short-term forecasting. Hourly forecasts of the power produced by the plant were carried out for a few days showing a quite good accuracy. A comparative study has also been introduced showing that the developed hybrid model performs better than both the SARIMA and the SVM model.