Energy Demand Forecasting Research Articles

Forecasting Energy Consumption Based on SVR and Markov Model: A Case Study of China

Forecasting energy demand in emerging nations is a critical policy tool utilized by decision makers worldwide. However, as estimated economic and demographic characteristics frequently diverge from realizations, precise forecast results are difficult to get due to the economic system’s intrinsic complexity. This work proposed a machine learning model for estimating energy consumption in China using the support vector regression model (SVR). Additionally, Markov Chain (MC) is employed to forecast and analyze the evolving energy consumption structure. The results demonstrate that SVR model is more accurate (98.4%) than the linear model (Moving Average model), the nonlinear model (Grey model), and past research in predicting energy usage. Under the current rate of energy consumption, China’s total energy consumption will break through six billion in the next 4 years. Furthermore, it is expected that China’s energy consumption structure will be more rational in 2025, with increased non-fossil energy consumption and decreased coal consumption, while natural gas consumption continues to grow at a low rate. It provides scientific basis for the implementation of carbon emission peak action, energy security and energy development plan during the 14th Five-Year Plan period.

The Implementation of the European Green Deal Strategy as a Challenge for Energy Management in the Face of the COVID-19 Pandemic

This article aims to present changes in the use of electricity by service companies, resulting from regulations within the framework of increasing energy efficiency from the perspective of the implementation of the European Green Deal strategy. To achieve the above goal, the following research question was formulated: to what extent did the COVID-19 pandemic affect the implementation of energy transformation and electricity consumption among the surveyed group of recipients? It should be noted that, so far in the global environment, more and more electricity has been used every year, and this tendency is still continuous and growing. Therefore, in European Union countries, measures have been taken to balance demand and its rational use, resulting from the implementation of the European Green Deal strategy. According to the strategic goal of the indicated policy, EU countries are obliged to implement a sequence of actions enabling their transformation into a modern, resource-efficient, and competitive economy. In particular, the strategy aims to achieve three main goals: 1. Achieving climate neutrality by reducing net greenhouse gas emissions to zero in 2050; 2. Decoupling economic growth from the consumption of natural resources; and 3. Striving for an equal standard of living in all countries. Therefore, the behavior of individual countries should focus on providing the required amount of energy to ensure socioeconomic growth while reducing pollution and environmental devastation caused by traditional methods of energy production and use. There are numerous talks and debates about the defined tasks and mechanisms leading to the achievement of the indicated goals, in which hardly any mention is made of the methods of monitoring the progress and evaluation of individual projects at the stage of building a new green deal. This study aims to fill the research gap observed in the current state of knowledge on energy management in EU countries from the perspective of the European Green Deal strategy and changes in its management resulting from the conditions of the COVID-19 pandemic. Therefore, in line with the aim of the article: 1. Factors influencing electricity consumption in service enterprises operating in EU countries were indicated; 2. Energy consumption variability in these entities was determined; and 3. The correlation between electricity consumption and GDP growth in the service sector located in EU countries was indicated. Therefore, a hybrid research method was used to analyze the data obtained from the databases of Eurostat and Statistics Poland, which consisted of the following analyses: diagnostic-descriptive, main components, and wavelet transform. Based on the conducted research, it should be concluded that energy consumption among service companies operating in the EU market is conditioned by three sources of causes along with the relevant consumption factors. The first group includes energy and technical premises, along with technological determinants. The second is shaped by financial and economic motives, which include socioeconomic factors. The third group is related to environmental sources represented by the natural environment and geographic and meteorological factors. The increase in electricity consumption in service enterprises is related to the average GDP growth of EU countries following a two-way cause-and-effect relationship implemented until 2019. Nevertheless, since 2020, there has been a noticeable decrease in energy consumption by 14.01% by service entities, which results from the limitations caused by the COVID-19 pandemic and the implementation of the European Green Deal strategy. Meanwhile, the structure of electricity consumption growth was dominated by industrial enterprises (increased by 37.7%) and individual consumers (increased by 7.8%). The results of the research may motivate the managers of EU countries and enterprises to analyze the factors of energy consumption, its variability, and dependence on economic growth, which contribute to determining the forecast of future energy demand, in connection with the ongoing energy transformation resulting from the implementation of the European Green Deal strategy, and economic slowdown caused by the COVID-19 pandemic. The issues presented in this article are an attempt to fill the gap indicating practical experience related to the process of electricity management and management in the political, economic, and technological dimensions from the perspective of implementing the European Green Deal strategy and the conditions resulting from the COVID-19 pandemic.

Using neural network modelling for estimation and forecasting of transport sector energy demand in developing countries

In developing countries, national-level institutions are often limited by key transportation energy efficiency indicators. A transportation model based on 40 artificial neural networks was developed to fill this gap. Data on energy efficiency indicators for 28 European countries have been collected to train a model for predicting these indicators using socio-economic variables. A bottom-up approach is then used to compare the predicted data to the total energy consumption. Morocco is used as a case study because of the absence of its energy efficiency indicators. The model's outstanding performance was proved after calculating energy demand at a highly disaggregated level. The model was used to forecast energy consumption up to 2050, considering a variety of alternative hypotheses. Four long-term energy demand scenarios were evaluated: frozen efficiency, implementation of EU legislation, cars electrification, and modal shift. The redistribution of passenger kilometres and tonne-kilometres as a way of rising average occupancy and average load revealed a significantly greater potential for energy savings. Switching from diesel to biofuel for buses and light cars was also examined as a solution to minimize GHG emissions. The developed model supplies decision-making institutions with the necessary tools for identifying critical issues, implementing policies, and redistributing infrastructure.

Research on Energy Supply and Demand Forecast and Carbon Neutralization Path Based on Grey-Monarch Butterfly Optimization Model

Energy and electricity are the key areas for China to achieve the double carbon target, and accurate forecasting of future energy supply and demand and carbon emissions is beneficial to develop a feasible path for low carbon transition. The gray prediction model GM (1, 1) is one of the most widely used dynamic prediction models in the field of energy forecasting, but it requires high raw data and the model may fail when the development coefficient of GM (1, 1) is large. On the other hand, the gray action of GM (1, 1) directly determines the model prediction accuracy, this paper introduces a novel population intelligence algorithm monarch butterfly optimization (MBO), which has excellent performance in practical optimization problems, into the optimization process of gray action of GM (1, 1), and proposes a new Gray-Monarch Butterfly optimization prediction model to realize the prediction. By comparing the prediction data with the classical literature, the effectiveness and superiority of the proposed Gray-monarch butterfly optimization prediction model are confirmed.Finally a carbon neutral pathway is given for Tianjin based on the prediction results.

On the Prediction of One-Year Ahead Energy Demand in Turkey using Metaheuristic Algorithms

On the Prediction of One-Year Ahead Energy Demand in Turkey using Metaheuristic Algorithms

Systematic Review of Deep Learning and Machine Learning for Building Energy

The building energy (BE) management plays an essential role in urban sustainability and smart cities. Recently, the novel data science and data-driven technologies have shown significant progress in analyzing the energy consumption and energy demand datasets for a smarter energy management. The machine learning (ML) and deep learning (DL) methods and applications, in particular, have been promising for the advancement of accurate and high-performance energy models. The present study provides a comprehensive review of ML- and DL-based techniques applied for handling BE systems, and it further evaluates the performance of these techniques. Through a systematic review and a comprehensive taxonomy, the advances of ML and DL-based techniques are carefully investigated, and the promising models are introduced. According to the results obtained for energy demand forecasting, the hybrid and ensemble methods are located in the high-robustness range, SVM-based methods are located in good robustness limitation, ANN-based methods are located in medium-robustness limitation, and linear regression models are located in low-robustness limitations. On the other hand, for energy consumption forecasting, DL-based, hybrid, and ensemble-based models provided the highest robustness score. ANN, SVM, and single ML models provided good and medium robustness, and LR-based models provided a lower robustness score. In addition, for energy load forecasting, LR-based models provided the lower robustness score. The hybrid and ensemble-based models provided a higher robustness score. The DL-based and SVM-based techniques provided a good robustness score, and ANN-based techniques provided a medium robustness score.

Multi‐hour ahead forecasting of building energy through a new integrated model

AbstractThe increase in electricity demand requires improved energy planning programs, which involve better energy distribution. Therefore, precise energy demand forecasting is of great importance for optimizing energy distribution. In this respect, a new hybrid forecasting model was proposed in this study for multi‐hour forecasting of total energy requirement (cooling and heating load of buildings) in three different regions in Algeria with different climate conditions. The proposed models are based on two main steps: In the first step, time‐series data were decomposed using Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN) into a different intrinsic function (IMFs), and then extreme learning machine (ELM) is employed as essence predictors. The proposed CEEMDAN‐ELM model uses the decomposed IMFs as input data, and the total energy concern as a desired output. The integrated CEEMDAN‐ELM model was evaluated and validated on three different databases each of which had 2 years of measurement on an hourly scale. Experimental results show that the hybridization mechanism CEEMDAN‐ELM outperform the stand‐alone model (ELM) in terms of forecasting errors over the entire forecasting horizon. Forecasting results of CEEMDAN‐ELM led to a normalized root mean square error (nRMSE) in the range of [0.71–4.66] for all studied regions and horizons whereas conventional ELM provides an nRMSE in the range of [0.93–16.89].

Prediction of transportation energy demand by novel hybrid meta-heuristic ANN

Road automobiles are deemed one of the major resources of energy consumption throughout cities. To realize and design sustainable urban transport, it is essential to comprehend as well as evaluate interactions among a set of elements, which form transport impacts and behaviors. The goal of the current research was to propose a hybrid algorithm, Artificial Neural Network (ANN)-Genetic Algorithm (ANN-GA), ANN-Simulated Annealing (ANN-SA), and Particle Swarm Optimization (ANN-PSO) to better optimize the coefficients for predicting the energy demand based on the several predictor variables (1975–2019) i.e., GDP, year, vehicle-km, population, oil price, passenger-km, and ton-km in Turkey. Eleven combinations of all predictor variables were selected and then compared with real data. The outcomes exposed that the proposed ANN-PSO technique based on the GDP, population, ton-km outperforms the other two models. It is anticipated that this research can be useful for developing extremely productive and applicable planning regarding transportation energy policies.

Prediction of Energy Demand by Socio-Economic and Climatic Indicators Using ANN optimized by Developed Search and Rescue (DSAR) Algorithm

With the highest energy growth rate and the fastest economic growth, China has more serious concerns about supplying energy to its economy. Therefore, one of the methods of energy management in this country is to forecast energy demand for the coming years. The main purpose of this study is to predict energy demand under the influence of socio-economic indicators by Artificial Neural Network (A = NN). More accurate forecasts make more reliable decisions to manage energy. For this purpose, in this study, innovation in the ANN method has been presented. This innovation is to use a developed version of a newly introduced metaheuristic. This developed algorithm increases the accuracy of prediction by solving optimization problems. This algorithm is a newly developed version of the Search and Rescue (SAR) algorithm. The results of predictions showed that the parameters population and GDP (per capita) have the highest correlation in train and test steps. The correlation between population and GDP (per capita) in training are R2 = 0.675 and R2 = 0.652, respectively. Also, the correlation between population and GDP in testing are R2= 0.624 and R2 = 0.519, respectively. The results also showed that in the coming years, society will face an upward trend in demand. Therefore, with the knowledge of this increase in demand, it is possible to manage energy demand in the coming years by managing energy in the current years.

An improved version of Search and Rescue algorithm for electrical power energy production and the demand of China

A new technique is suggested in this paper for optimal electrical power energy production and demand forecasting in China. The paper provides a new improved design of the Search and Rescue algorithm for better optimization of the forecasting models. This study is performed based on three scenarios on the electrical power energy production and demand in the period 2020–2050. The study utilizes linear and quadratic models for electricity energy production and demand prediction of the analyzed scenarios. The improved Search and Rescue algorithm in the study is designed based on import and export, gross domestic product (GDP), and population. All of the models based on the scenarios are applied for optimal prediction of electrical power energy generation and demand. The results show that the proposed optimal quadratic models based on the Improved Search and Rescue algorithm provides the fittest results with economic indicators.

FederatedGrids: Federated Learning and Blockchain-Assisted P2P Energy Sharing

Peer-to-Peer (P2P) energy trading platforms envisioned energy sectors to satisfy the increasing demand for energy. The vision of this paper is not only to trade energy but also to have part of it being shared. Therefore, this paper presents <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">FederatedGrids</i> which is a P2P energy trading and sharing platform inside and across microgrids. Energy sharing allows exchanging energy between the categories of consumers and prosumers in return for future benefits. <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">FederatedGrids</i> platform uses blockchain and federated learning to enable autonomous activities while providing trust and privacy among all participants. Indeed, based on various smart contracts using federated learning, <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">FederatedGrids</i> calculates a prediction of the future energy production and demand allowing the system to autonomously switch between trading and sharing, and enabling the prosumers to make decisions related to their participation in the energy sharing process. Up to our knowledge, this work is the first attempt to create a hybrid energy trading and sharing platform, with the real sharing meaning, and that uses federated learning over the smart contract for energy demand prediction. The experimental results showed a 17.8% decrease in energy cost for consumers and a 76.4% decrease in load over utility grids.

An improved grey Verhulst model to forecast energy demand in the USA and Turkey

The importance of accurate energy demand modelling has increased to support the decision making of policymakers for ensuring a safe energy supply. However, forecasting energy demand has several difficulties due to the complexity of the supply line, demand increase, non-linearity of data and volatility of energy usage. In this study, an improved grey Verhulst model with a constant term (GVMCT), which is based on the grey model, is introduced for improving the accuracy of energy demand prediction models. Within this context, the total residential electricity demand of both the USA and Turkey is modelled by way of linear and quadratic trend models, as well as three grey models, including the proposed GVMCT model. The effectiveness of the models is assessed based on the mean absolute error (MAE), mean squared error and root mean square error. The results show that the linear trend is the best-performing model, with an MAE of 34 564.81844, for the US data, whereas the proposed GVMCT, with an MAE of 4130.086917, outperforms all models for the data of Turkey.

Validation of weather reanalysis datasets and geospatial and techno-economic viability and potential assessment of concentrated solar power plants

Accelerating the decarbonization of power sector requires strategic planning and multi-aspectual feasibility analysis of renewable energy systems including, but not limited to, meteorological, land-use, techno-economic and environmental parameters. In this study, at first, the suitability of reanalysis datasets of Direct Normal Irradiance, air temperature, wind speed, relative humidity and air pressure was investigated at different timescales and locations over Pakistan. High correlation (R > 0.9) and low bias is found for Surface Solar Radiation Data records – Heliosat East, after preprocessing, among radiation datasets in consideration. Likewise, Modern-Era Retrospective analysis for Research and Applications version 2 datasets of temperature and air pressure strongly correlate with ground measurements whereas wind speed and relative humidity show peculiarity. Afterwards, resource maps from long-term timeseries are developed for each of the aforementioned parameters. In the second part, spatial feasibility and Analytical Hierarchy Process based site suitability is assessed utilizing site-specific meteorological, socio-economic and environmental parameters. Moreover, multi-parametric evaluation of techno-economic potential is performed using bias-corrected datasets. For technical potential, System Advisor Model’s Physical Trough Collector model is used for exclusively selected points. Additionally, cost of water usage is introduced in estimation of Levelized Cost of Electricity besides other parameters. Cumulative power capacity potential of 10,035 GW is identified for country, with net cost estimated below 0.096 USD/kWh. Nine sites are found to have generation potential greater than 1200 TWh/year, the individual capacity to fulfil total predicted energy demand of the country for year 2030. Around 6.5% of viable regions are found to be highly suitable for plant deployment. Land cover and lack of access to transmission network act as limiting barriers in installation of concentrated solar power plants at many potential sites across the country.

Building Thermal-Network Models: A Comparative Analysis, Recommendations, and Perspectives

The development of smart buildings, as well as the great need for energy demand reduction, has renewed interest in building energy demand prediction. Intelligent controllers are a solution for optimizing building energy consumption while maintaining indoor comfort. The controller efficiency on the other hand, is mainly determined by the prediction of thermal behavior from building models. Due to the development complexity of the models, these intelligent controllers are not yet implemented on an industrial scale. There are primarily three types of building models studied in the literature: white-box, black-box, and gray-box. The gray-box models are found to be robust, efficient, of low cost computationally, and of moderate modeling complexity. Furthermore, there is no standard model configuration, development method, or operation conditions. These parameters have a significant influence on the model performance accuracy. This motivates the need for this review paper, in which we examined various gray-box models, their configurations, parametric identification techniques, and influential parameters.

Development of climatic zoning and energy demand prediction for Ethiopian cities in degree days

The energy need for comfort rises as the living standard and urbanization escalate. This study intends to regionalize the climatic zones of Ethiopia and estimate the energy needs of respective regions based on the prevalent method of degree days. For this temperature data for 952 National Center for Environmental Prediction (NCEP) stations from 1979 to 2013 were used to compute the heating degree days (HDD) and cooling degree days (CDD) with base temperature of 18.3 °C and to regionalize the country into 5 ASHRAE standard zones. Within each zone, five reference cities were identified for their population size and urbanization level, namely, Goba Robe, Robe, Addis Ababa, Dire Dawa and Gode. The result has shown that the eastern and western peripheries of the country (402,552 km2) attained high CDD (2659 DD) and low HDD. The central spine of the country attained high HDD (29,981 km2) and low CDD. Between the two there are intermediate zones (zone 3 and 4) of 706,178 km2 coverage. The yearly total energy demand for comfort in Ethiopia is estimated to be 0.068 to 0.168 kJ (kJ)/°C in a unit cube of room space. That means, if all households are furnished with AC and standard of living raised, with the population size of 143,028,465 in 2030, Ethiopia demands to produce 4,689,205.7 kJ of energy.

Optimal Energy Demand Forecasting for Systems by Improved Manta Ray Foraging Optimizer: A Case Study in Taiwan

Optimal Energy Demand Forecasting for Systems by Improved Manta Ray Foraging Optimizer: A Case Study in Taiwan

Optimal Energy Demand Forecasting for Systems by Improved Manta Ray Foraging Optimizer: A Case Study in Taiwan

Optimal Energy Demand Forecasting for Systems by Improved Manta Ray Foraging Optimizer: A Case Study in Taiwan

Forecasting Method of Energy Demand of Integrated Energy System Considering Seasonal Catastrophe

The change of season will cause a variety of factors affecting energy demand to change, resulting in severe fluctuations in energy demand. Accurate prediction is of great value for energy management. Therefore, a prediction method of energy demand of integrated energy system considering seasonal mutation is proposed. Based on the analysis of the basic concepts, attributes and influencing factors of energy and energy demand, a seasonal energy demand impact decomposition model is constructed by lmdi-i to decompose the impact of season on energy demand. By dividing the energy indicators of the integrated energy system, a fuzzy neural network with error output and correction mechanism is established to predict energy demand. The test results show that the maximum relative error of the prediction results of the design method is 5.89%, the minimum relative error is 1.03%, the average absolute error is 3.21%, the root mean square error is 0.019, and the hill inequality coefficient is 0.020. Are better than the comparison method.

China’s Energy Demand Forecasting Based on the Hybrid PSO-LSSVR Model

Forecasting energy demand accurately is the basis for the formulation and implementation of energy planning. In this paper, energy demand influencing factors are mainly decomposed into scale economy effect, population size effect, energy structure effect, and residential consumption effect based on the Logarithmic Mean Divisia Index (LMDI). Then, the Cointegration and Granger Causality tests are used to discover the influencing factors of energy demand in China. On this basis, a hybrid optimization algorithm, the least-squares support-vector regression optimized by particle swarm optimization (PSO-LSSVR), is proposed to forecast the energy demand of China. Then, three scenarios are set up to analyze the further development of drive factors of energy demand. Finally, in accordance with the forecasting results, some suggestions related to China’s energy development policy are given. The main results are as follows. First, gross domestic product (GDP), the total population at the end of the year (POP), the coal consumption ratio in energy (CCR), and residential consumption levels (RCLs) are dominant indicators of energy demand in China. Second, the improved PSO-LSSVR model has significant superiority than other models in energy demand forecasting, a complex and nonlinear system with small samples. Third, China’s energy demand will peak in 2022, which is 4.9 million tce in all scenarios.

Modelling of Turkey’s Energy Consumption Using Artificial Neural Networks

Energy is one of the most important inputs to maintain social and economic improvement in the countries. It is necessary that energy demand should be performed at the right time economically and be of good quality and respectful if increasing environmental consciousness in order to preserve national development and a high standard of living. Turkey's energy use is expected to increase by 50% over the next decade. Turkey's installed capacity has exceeded 88 GW as of January 2019, representing a threefold increase in 15 years. For this reason, an accurate prediction of the consumed energy is critical. Predictions of energy demand in developing countries show more deviations than in developed countries. The essential scope of this study is to develop a new electricity prediction model for Turkey, which has not been used in the literature before. In the study, the global system for mobile communications (gsm) subscribers, fertility rate and cultivated land per capita have been used for the first time in the literature as variables. Factors resulting from health-ecological problems as well as cultural, social and economic changes and differences in Turkey were included in the model to obtain more realistic results. The model was developed between 1975 and 2016, and 73 different economic and social variables were evaluated using artificial neural networks (ANN). The model was established by reducing the number of variables according to the weight ratio. Then, two different cases have been created and tested. Turkey’s electricity consumption has been predicted accurately until 2023 using SPSS Clementine software.