Energy Price Forecasting Research Articles

A meta-analysis of the schematic design process of deep retrofit projects

Deep retrofits of the existing building stock will be necessary to meet global emissions reductions targets. One building archetype, low-rise MURBs have been neglected in terms of research and funding for deep retrofits. A meta-analysis was conducted that compares and contrasts the schematic design approach taken for six such buildings in British Columbia which are scheduled to undergo deep retrofits with the goal of reducing GHG emissions by 80%. The analysis showed that design teams had converged toward common solutions for each building while achieving the GHG reduction target. The recommended measures include electrification of space and domestic hot water heating, adding insulation through overcladding, air sealing, ventilators for each unit, and double pane windows. A life cycle cost analysis showed that the economic viability of deep retrofits were dependent on energy price forecasts, capital cost reductions through market forces and transformation, or incentives cover the non-monetizable co-benefits of deep retrofits such as improved resiliency to climate-change or reducing overheating and air quality risks. The meta-analysis can help to streamline the early-stage and schematic design process for such buildings, which is critical to increasing the retrofit rate. This process could be replicated for other building types and construction archetypes.

Extension of a low-tech Model Predictive Control (MPC) algorithm for grid-supportive heat pump operation

This paper presents the further development of a low-tech Model Predictive Control (MPC) for the grid-friendly operation of heat pumps. The developed algorithm is already available in the literature [1] and successfully implemented in multiple buildings. The overall goal is, to predictively optimize comfort parameters taking into account weather forecast data for 48 h. A simple mathematical building model and an optimization function are used to calculate the heating (or cooling) requirement and determine the optimum heating (cooling) output curve over time.Based on the initial results, the algorithm is extended to include an energy price forecast. This modification enables the simultaneous optimization of comfort and heat supply costs. This forecast-based cost & comfort function is designed for the use of heat pumps in buildings with thermally activated components (TAB). The extension enables the MPC to process day-ahead electricity prices or other price signals in addition to weather forecast data. This means that the operation of heat pumps can be shifted to periods that are beneficial to the grid.The algorithm is analysed and validated in a Matlab/Simulink building simulation for a heating period of one sample month. A case study with different price scenarios is the main part of this paper. Depending on the assumed price fluctuation, the simulation leads to cost savings between 6.65 % and 12.5 % for the observation period of one month. The simulation results thus show that heat generation can be shifted to times with lower electricity prices, which leads to a significant reduction in heating costs without any significant loss of comfort. In conclusion, the developed low-tech MPC enables grid-supportive operation of heat pumps.

Price responsiveness of solar and wind capacity demands

Accurate estimates of the price responsiveness of residential, commercial, and industrial electricity demands are essential for energy policy modelling, integrated resource planning, and determining a competitive wholesale electricity market's generation levels, prices, and capacity investments. Hence, we estimate the own-price elasticities of solar and wind capacity demands of a load serving entity (LSE) that provides retail electricity service, thereby answering two interrelated research questions: (1) does solar capacity demand far exceed wind capacity demand? and (2) are solar and wind capacity demands price-elastic? Inspired by the theory of input demand under input price uncertainty, our innovative methodology integrates (a) wholesale spot energy price forecasts by time of day; (b) pseudo data found by minimizing a LSE's annual risk-adjusted budget for procuring solar and wind capacities; and (c) econometric analysis of (b) to estimate the extent of substitutability between solar and wind capacities and the own-price elasticities of solar and wind capacity demands. Using Texas as an illustrative example, we find that when solar and wind power purchase agreements have similar energy prices, solar capacity demand is approximately four times wind capacity demand. Further, the own-price elasticity estimates are −5.34 for solar capacity demand and −5.65 for wind capacity demand. As a result, solar and wind capacity demands tend to substantially grow (shrink) in response to declining (rising) solar and wind energy prices. This lends support to proposals to raise solar and wind energy prices for mitigating the adverse effects of large-scale variable renewable energy development on an electric grid's efficient operation and system reliability. However, adopting such proposals also slows the grid's pace of decarbonization, thus underscoring the policy and regulatory challenges in the quest for a clean and sustainable electricity future. Hence, our policy recommendation of price managing solar and wind capacity demands is a topic of policy debate that deserves the attention of an electric grid's stakeholders.

Seeing is believing: Forecasting crude oil price trend from the perspective of images

AbstractIn this paper, we propose a novel imaging method to forecast the daily price data of West Texas Intermediate (WTI) crude oil futures. We use convolutional neural networks (CNNs) for future price trend prediction and obtain higher prediction accuracy than other benchmark forecasting methods. The results show that images can contain more nonlinear information, which is beneficial for energy price forecasting. Nonlinear factors also have a strong influence during drastic fluctuations in crude oil prices. In the robustness tests, we find that the image‐based CNN is the most stable approach and can be applied in various futures forecasting scenarios. In the prediction of low‐frequency models for high‐frequency data, the CNN method still retains considerable predictive power, indicating the possibility of transfer learning of our novel approach. By unleashing the power of the picture, we open up a whole new perspective for forecasting future energy trends.

Powering Electricity Forecasting with Transfer Learning

Accurate forecasting is one of the keys to the efficient use of the limited existing energy resources and plays an important role in sustainable development. While most of the current research has focused on energy price forecasting, very few studies have considered medium-term (monthly) electricity generation. This research aims to fill this gap by proposing a novel forecasting approach based on zero-shot transfer learning. Specifically, we train a Neural Basis Expansion Analysis for Time Series (NBEATS) model on a vast dataset comprising diverse time series data. Then, the trained model is applied to forecast electric power generation using zero-shot learning. The results show that the proposed method achieves a lower error than the benchmark deep learning and statistical methods, especially in backtesting. Furthermore, the proposed method provides vastly superior execution time as it does not require problem-specific training.

Optimum exploitation of multiple energy system using IGDT approach and risk aversion strategy and considering compressed air storage with solar energy

The concept of an energy hub within a microgrid has garnered significant attention from investors due to its potential for efficiently managing diverse energy sources and accurately forecasting energy prices. This study presents a model for an energy hub that facilitates the optimal operation of a microgrid with multiple energy carrier infrastructures under future-day scenarios. The primary objective of the optimization process is to minimize operating costs, environmental impacts, and technical limitations. The proposed energy hub model integrates both distributable generation, such as a Combined Cooling, Heating, and Power (CCHP) unit, and non-distributable generation sources, including wind turbines and photovoltaic systems. Additionally, it incorporates various energy storage components, namely Ice Storage Conditioner (ISC), Thermal Energy Storage (TES) systems, and Electric Vehicles (EVs). Furthermore, the study investigates the impact of a novel energy storage system called Solar Powered Compressed Air Energy Storage (SPCAES) on the operational and environmental performance of the energy hub. To address uncertainties stemming from predicted and actual variables, a two-level stochastic optimization problem is formulated based on the Information Gap Decision Theory (IGDT) with a Risk Aversion (RA) strategy. The overarching aim is to mitigate the risks associated with the information gap confronted by microgrid operators. Subsequently, the two-level stochastic optimization problem is transformed into a single-level problem using the Karush-Kuhn-Tucker method. The proposed framework employs the RA-IGDT method and utilizes the Elephant Herd Optimization (EHO) algorithm to effectively solve the optimization problem. By applying this proposed framework to a typical energy hub, this research demonstrates the effectiveness of energy storage systems in reducing operating costs and greenhouse gas emissions during daily energy management. The study highlights the potential of energy hubs and storage systems for cost-effective and eco-friendly microgrid energy management. Scenario 4, featuring the proposed energy hub with SPCAES and ISC storage, demonstrates a 15 % reduction in operating costs and a notable 20 % increase in peak-hour electrical power for sale, underscoring its economic and operational benefits.

AI-Assisted Forecasting of Energy Prices

This study explores the application of artificial intelligence (AI) in forecasting energy prices by developing advanced models that integrate diverse data sources, including market data, weather patterns, and geopolitical events. Traditional forecasting methods, such as ARIMA and GARCH, have shown limitations in handling energy markets’ complex, non-linear relationships. In contrast, AI models demonstrate superior accuracy and adaptability, particularly deep learning and ensemble methods. The research findings indicate that AI-assisted models outperform traditional approaches, providing more reliable predictions and valuable insights for energy companies and policymakers. These models enhance risk management, optimize production schedules, and support strategic decision-making, ultimately contributing to more stable and efficient energy markets. The study also highlights the importance of integrating multiple data sources and suggests future research directions, including the development of real-time forecasting systems and explainable AI techniques. The implications of this research are significant, offering new tools for navigating the complexities of global energy markets and improving decision-making processes in both industry and policy.

Modeling market trading strategies of the intermediary entity for microgrids: A reinforcement learning-based approach

Participation of a large number of microgrids (MGs) in the energy markets faces several challenges. To address these challenges, MGs can be aggregated by an Intermediary entity (IE). In this paper, a new decision-making framework is proposed to solve the energy management problem of multiple MGs and their participation in the energy market through the IE. This framework has three main functions. First, energy price forecasting is conducted using the long short-term memory (LSTM) recurrent neural networks method. Then, the power exchange of the MGs with the main grid is optimized by solving their energy management problem. In this stage, the integrated power exchanges of the MGs with the grid are determined based on the predicted prices. In the final function, the Monte Carlo reinforcement learning technique is employed to optimize real-time pricing decisions and identify potential obstacles that may affect proposed bid prices. This approach addresses energy management challenges and enables profitable energy trading in multiple MGs within energy markets, with confirmation supported by simulation results. The results show a maximum increment of 4.55% in the profit of the IE when purchasing energy and a 3.79% maximum increment when selling energy in the real-time market compared to day-ahead decisions, respectively.

Time‐varying partial‐directed coherence approach to forecast global energy prices with stochastic volatility model

AbstractFor investors and policymakers, forecasting energy prices with accuracy is essential and plays a major role in the global bulk commodity markets. The current study proposes a novel hybrid forecasting model to predict global energy prices, namely, time‐varying partial‐directed coherence with stochastic volatility. The proposed method combines partial‐directed coherence analysis and stochastic volatility models. Accordingly, this study attempts to provide an in‐depth understanding of the relationship between energy markets and global economic conditions as well as the causality pathway between the underlined markets. Monthly data from January 1982 to July 2022 is used in this study. The results show a strong causality between global economic conditions, European oil, and natural gas prices and have profound implications for policymakers. For completeness, we extend the analysis to the forecasting ability of global economic conditions for oil and natural gas prices. The out‐of‐sample results show that the autoregressive model incorporating the global economic conditions index can significantly improve the accuracy of oil and gas price forecasts. In addition, our results are strongly robust over a variety of time horizons for forecasting, and they provide valuable insights into the forecasting choices to guide investment strategies in the energy and financial market.

Structure Optimization of Ensemble Learning Methods and Seasonal Decomposition Approaches to Energy Price Forecasting in Latin America: A Case Study about Mexico

The energy price influences the interest in investment, which leads to economic development. An estimate of the future energy price can support the planning of industrial expansions and provide information to avoid times of recession. This paper evaluates adaptive boosting (AdaBoost), bootstrap aggregation (bagging), gradient boosting, histogram-based gradient boosting, and random forest ensemble learning models for forecasting energy prices in Latin America, especially in a case study about Mexico. Seasonal decomposition of the time series is used to reduce unrepresentative variations. The Optuna using tree-structured Parzen estimator, optimizes the structure of the ensembles through a voter by combining several ensemble frameworks; thus an optimized hybrid ensemble learning method is proposed. The results show that the proposed method has a higher performance than the state-of-the-art ensemble learning methods, with a mean squared error of 3.37 × 10−9 in the testing phase.

Short-Term Forecasting of Global Energy and Metal Prices: VAR and VECM Approaches

This study introduces a set of multivariate models with the aim of forecasting global prices of 1) crude oil, 2) natural gas, 3) iron ore, and 4) steel. Various versions of vector autoregression and error-correction models are applied to monthly data for the short-term prediction of nominal commodity prices six months ahead, and to examine forecast accuracy. The fundamentals for metal and energy price predictions include inter alia, stock changes, changes in commodity production volumes, export volumes by the largest players, changes in the manufacturing sector of the largest consumers, the state of global real economic activity, freight rates, recession, and so on. Kilian's (2009) index of global real economic activity is found to be a useful proxy for global demand and a reliable input in forecasting both energy and metal prices. The findings suggest that models with smaller lag orders tend to outperform those with a higher number of lags. At the same time, selected individual models, while showing a standalone high performance, have varying forecast precision during different periods, and no individual model outperforms others consistently throughout the forecast horizon.

Hourly electricity price forecast for short-and long-term, using deep neural networks

Abstract Despite the practical importance of accurate long-term electricity price forecast with high resolution - and the significant need for that - only small percentage of the tremendous papers on energy price forecast attempted to target this topic. Its reason can be the high volatility of electricity prices and the hidden – and often unpredictable – relations with its influencing factors. In our research, we performed different experiments to predicate hourly Hungarian electricity prices using deep neural networks, for short-term and long-term, too. During this work, investigations were made to compare the results of different network structures and to determine the effect of some environmental factors (meteorologic data and date/time - beside the historical electricity prices). Our results were promising, mostly for short-term forecasts - especially by using a deep neural network with one ConvLSTM encoder.

Energy Price Prediction Integrated with Singular Spectrum Analysis and Long Short-Term Memory Network against the Background of Carbon Neutrality

In the context of international carbon neutrality, energy prices are affected by several nonlinear and nonstationary factors, making it challenging for traditional forecasting models to predict energy prices effectively. The existing literature mainly uses linear models or a combination of multiple models to forecast energy prices. For the nonlinear relationship between variables and the mining of historical data information, the prediction strategy and accuracy of the existing literature need to be improved. Thus, this paper improves the prediction accuracy of energy prices by developing a “decomposition-reconstruction-integration” thinking strategy that affords medium- and short-term energy price prediction based on carbon constraint, eigenvalue transformation and deep learning neural networks. Considering 2011–2020 as the research period, the prices for traditional energy resources and polysilicon in clean photovoltaic energy raw materials are selected as representatives. Based on energy price decomposition using the Singular Spectrum Analysis (SSA) method, and combining it with Learning Vector Quantization (LVQ) cluster technology, the decomposed quantities are aggregated into price sequences with different characteristics. Additionally, the carbon intensity is considered the leading market’s overall constraint, which is input with the processed price data into a Long Short-Term Memory network (LSTM) model for training. Thus, the SSA-LSTM combined forecasting model is developed to predict the energy price under carbon neutrality. Four indices are employed to evaluate the prediction accuracy: Root Mean Squared Error (RMSE), Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE) and R-squared. The results highlight the following observations. (1) Using a sequence decomposition clustering strategy significantly improves the model’s prediction accuracy. This strategy enhances predicting the overall trend of the price series and the changes in different periods. For coal price, the RMSE value decreased from 0.135 to 0.098, the MAE value decreased from 0.087 to 0.054, the MAPE value decreased from 0.072 to 0.064, and the R-squared value increased from 0.643 to 0.725. Regarding the polysilicon price, the RMSE value decreased from 0.121 to 0.096, the MAE value decreased from 0.068 to 0.064, the MAPE value decreased from 0.069 to 0.048, and the R-squared value increased from 0.718 to 0.764. (2) The prediction effect is better in the case of carbon constraint. Considering “carbon emission intensity” as the overall constraint of the leading market, it can effectively explore the typical characteristics of energy price information. Four evaluation indicators show that the accuracy of the model prediction can be improved by more than 3%. (3) When the proposed SSA-LSTM model is used to predict both prices, the results show that the evaluation index of the prediction error remained at about 1%, while the model’s accuracy was high. This also proves that the proposed model can predict traditional energy prices and new energy sources such as solar energy.

Forecasting energy prices: Quantile‐based risk models

AbstractThe goal of this paper is to use a new modelling approach to extract quantile‐based oil and natural gas risk measures using quantile autoregressive distributed lag mixed‐frequency data sampling (QADL‐MIDAS) regression models. The analysis compares this model to a standard quantile auto‐regression (QAR) model and shows that it delivers better quantile forecasts at the majority of forecasting horizons. The analysis also uses the QADL‐MIDAS model to construct oil and natural gas prices risk measures proxying for uncertainty, third‐moment dynamics, and the risk of extreme energy realizations. The results document that these risk measures are linked to the future evolution of energy prices, while they are linked to the future evolution of US economic growth.

A New ANN Approach for Time Series Analysis

Time series analysis and accurate forecasting of energy prices are critical for both policymakers and market participants. In the practical analysis of price time series, the coefficients play vital roles; however, their accurate estimation is a challenging issue as they are affected by external factors. This study proposes a new modelling approach for Artificial Neural Networks (ANNs) models based on fuzzy logic. For this purpose, we reformulated an ANN model as a fuzzy nonlinear regression model to capture the advantages of both fuzzy regression and ANN methodologies. This clear-box model can be applied to not only uncertain, ambiguous, and complex environments, but it is also capable of modelling nonlinear patterns. To illustrate the capability of the proposed approach, we report a case study of liquified natural gas (LNG) prices in Japan’s market (the world’s largest natural gas importer). The results support that the performance of the proposed approach is acceptable; moreover, it can deal with uncertain and complex environments as a clear-box model.

A novel dynamic time-delay grey model of energy prices and its application in crude oil price forecasting

The correct forecasting of energy prices can better regulate energy market allocation, reduce the unit energy allocation cost, and maximize the economic benefits of the energy allocation cost. First, this paper aims to study the time-delay in the energy supply and demand relationship, which directly affects changes and fluctuations in the energy price. Starting from the market economic model, this paper analyses the dynamic time-delay process of energy prices by using the relationship between energy supply, demand and prices, and it establishes the time-delay differential equation of energy prices. Second, a novel dynamic grey time-delay forecasting model of energy prices is selected based on the differential information of the differential equation and difference equation and the principle of data reduction. The parameters and optimization methods of the model are studied to obtain the modeling process of the model. Finally, the novel model is applied to predict the Novel York Mercantile Exchange (NYMEX) West Texas Intermediate (WTI) crude oil futures price, the Intercontinental Exchange (ICE) Brent crude oil futures price, and the Organization of the Petroleum Exporting Countries (OPEC) basket crude oil spot price. The forecasting results are better than those of the other four grey prediction models, which indicates that the novel model can effectively predict the oil price and that the novel model can describe the dynamic change law of the energy price system with a time-delay.

Forecasting energy prices using a novel hybrid model with variational mode decomposition

Forecasting energy prices accurately has always played the important role in the country's energy security and environmental impacts of policies. This paper proposes a novel decomposition-ensemble model to predict the energy prices which fluctuate wildly. The several steps process as follows: (1) The original energy prices are decomposed into sublayers with different frequencies by variational mode decomposition (VMD). (2) The autoregression model (AR) predicts the first low frequency component and Elman neural network (ELMAN) forecasts the last high frequency component. Besides, the improved bidirectional long short-term memory (IBiLSTM, the attention-based convolutional neural network and bidirectional long short-term memory) predicts other sublayers. (3) The prediction of the sublayers with different models is reconstructed as final predicted results with the non-linear integration approach. Combining econometric and artificial intelligence methods with the asymmetric feature makes the forecasting performance more accurate. The novel model outperforms other related comparative models under different training sets lengths. In general, experiments on two cases of energy prices: natural gas and carbon futures prices demonstrate the validity and reliability of the proposed model AR-IBiLSTM-ELMAN with VMD. The advanced model could simultaneously exploit the unique advantages of each model which provides an effective forecasting tool for governments and enterprises.

Forecasting Nodal Price Difference Between Day-Ahead and Real-Time Electricity Markets Using Long-Short Term Memory and Sequence-to-Sequence Networks

Price forecasting is at the center of decision making in electricity markets. Much research has been done in forecasting energy prices for a single market while little research has been reported on forecasting price difference between markets, which presents higher volatility and yet plays a critical role in applications such as virtual trading. To this end, this paper takes the first attempt at it and employs novel deep learning architecture with Bidirectional Long-Short Term Memory (LSTM) units and Sequence-to-Sequence (Seq2Seq) architecture to forecast nodal price difference between day-ahead and real-time markets. In addition to value prediction, these deep learning architectures are also used to develop classification models to predict the price difference bands/ranges. The proposed methods are tested using historical PJM market data, and evaluated using Root Mean Squared Error (RMSE) and other customized performance metrics. Case studies show that both deep learning methods outperform common methods including ARIMA, XGBoost and Support Vector Regression (SVR) methods. More importantly, the deep learning methods can capture the magnitude and timing of price difference spikes. Numerical results show the Seq2Seq model performs particularly well and demonstrates generalization capability to extended forecasting lead time.

Modelling the Future of the Baltic Energy Systems: A Green Scenario

Abstract The electricity sector in Europe and in the world is undergoing rapid and profound changes. There is a sharp increase in the capacity of renewable energy sources, coal and nuclear power plants are being closed and new technologies are being introduced. Especially rapid changes are taking place in the energy systems of the Baltic States. Under these conditions, there is an emerging need for new planning tools particularly for the analysis of the power system properties in a long-term perspective. The main contribution of this article lies in the formulation and solution of optimization problems that arise when planning the development of power systems in the Baltic States. To solve this problem, it is necessary to use models of various power plants and make a number of assumptions, the justification of which requires the following actions: to briefly review the current situation of the production and demand of energy in the Baltic power systems; to conduct an overview of the Baltic interconnections and their development; to make forecasts of energy prices, water inflow, energy production and demand; to set and solve the problems of optimization of power plant operation modes; to demonstrate the possibility and limitations of the developed tools on the basis of real-life and forecast data. In this paper, a case study is performed using the main components of the overall modelling framework being developed. It focuses on the Baltic power systems in 2050 under the conditions of significant expansion in the installed capacity of renewable energy sources (RESs) and diminished fossil fuel power plant activity. The resulting electricity generation mix and trade balance with neighbouring countries is assessed, showing that even with significant RES expansion, the Baltic countries remain net importers and because of the intermittency of RESs, there are hours within the year when the demand cannot be met.

Learning spatiotemporal dynamics in wholesale energy markets with dynamic mode decomposition

Energy markets balance electricity generation (supply) and demand while ensuring non-discriminatory access. Understanding energy market dynamics is essential to improving grid efficiency and optimizing the development of new energy conversion and storage technologies. Accurate energy price forecasts are important for many energy storage technologies to be profitable from price arbitrage. In this paper, we apply Dynamic Mode Decomposition (DMD), a popular spatial-temporal reduced-form modeling technique, to forecast 6587 locational marginal prices in the California Day-Ahead Market (DAM). DMD is a promising equation-free modeling technique in systems with inherent periodic tendencies in time such as financial markets and fluid dynamics. Yet we show, for the first time, that DMD cannot reliably forecast day-ahead energy prices due to the periodicity of the data. Instead, we show Augmented DMD (ADMD) overcomes these limitations and is a fast and accurate price forecaster. We benchmark DMD, ADMD, and backcasting forecasting methods for optimal price arbitrage with an energy storage system. Using ADMD, a market-connected energy storage system can capture up to 92% of allowable revenues in rolling horizon simulations. Moreover, ADMD is up to 1000-times faster than time-series forecasting methods and requires orders of magnitude less data than deep/machine learning techniques.