A new feature selection algorithm and composite neural network for electricity price forecasting

Accurate price forecasting is crucial for all market participants in electricity markets. This study presents a hybrid price forecasting framework based on a new data fusion algorithm. Owing to the complexity and distinct nature of the electricity price, a single forecast engine cannot capture all different patterns of the price signals. Hence, this study focuses on a hybrid forecasting method to extract the advantages of several forecasting engines. In the proposed method, artificial neural network, adaptive neuro‐fuzzy inference system and autoregressive moving average are employed as primary forecast engines (agents) which provide three independent price forecasts. Then, a new data fusion algorithm, the modified ordered weighted average (modified OWA), is proposed to combine the three forecasts to generate a single unified price forecast. Hopefully, the fusion's output outperforms all the agents’ forecasts. The author's proposed fusion algorithm, unlike conventional OWA, uses the feedback from the agents’ error. The proposed framework is evaluated on the Spanish electricity market. The results confirm the ability of the proposed fusion framework to provide more accurate forecasts compared with the input agents forecasts. Results are also compared with some of the recent electricity price forecasting methods.

In this paper, a new feature selection and forecast engine is presented for day ahead prediction of electricity prices, which are so valuable for both producers and consumers in the new competitive electric power markets. In a competitive electricity market, forecast of energy prices is a key information for the market participants. However, price signal usually has a complex behavior due to its nonlinearity, non-stationary, and time variance. Also, an appropriate feature selection is crucial for accurate forecasting. In this paper, a two-step approach that identifies a set of candidate features based on the data characteristics proposed and then selects a subset of them using correlation and instance-based feature selection methods, applied in a systematic way. Then, a combination of wavelet transform (WT) and a hybrid forecast method is presented based on neural network (NN) and an optimization algorithms. The proposed method is examined on PJM electricity market and compared with some of the most recent price forecast methods. These comparisons illustrate effectiveness of the proposed strategy.

Price forecast in the competitive electricity market by support vector machine

Electricity price is an essential signal in competitive electricity market. Every participant has to make a prediction of the clearing price in order to optimize the bids thus price forecast acts as the decision making basis. This study presents a comparative analysis of diverse price forecasting methods to give a comprehensive analysis of these forecasting methods. The presented methods comprise of a conventional method i.e. statistical method- auto regressive integrated moving average, adaptive method i.e. machine learning method- random forest regression, and novel deep learning method i.e. long short term memory artificial neural network. The price forecasting is formulated as a time series forecasting problem. These techniques are used for predicting electricity price of New South Wales area of Australian electricity market and Singapore electricity market over a span of 5 years ranging from 2017–2021.

This paper analyzes electricity markets in Slovenia during the specific period of market deregulation and price liberalization. The drivers of electricity prices and electricity consumption are investigated. The Slovenian electricity markets are analyzed in relation with the European Energy Exchange (EEX) market. Associations between electricity prices on the one hand, and primary energy prices, variation in air temperature, daily maximum electricity power, and cross-border grid prices on the other hand, are analyzed separately for industrial and household consumers. Monthly data are used in a regression analysis during the period of Slovenia’s electricity market deregulation and price liberalization. Empirical results show that electricity prices achieved in the EEX market were significantly associated with primary energy prices. In Slovenia, the prices for daily maximum electricity power were significantly associated with electricity prices achieved on the EEX market. The increases in electricity prices for households, however, cannot be explained with developments in electricity prices on the EEX market. As the period analyzed is the stage of market deregulation and price liberalization, this can have important policy implications for the countries that still have regulated and monopolized electricity markets. Opening the electricity markets is expected to increase competition and reduce pressures for electricity price increases. However, the experiences and lessons learned among the countries following market deregulation and price liberalization are mixed. For industry, electricity prices affect cost competitiveness, while for households, electricity prices, through expenses, affect their welfare. A competitive and efficient electricity market should balance between suppliers’ and consumers’ market interests. With greening the energy markets and the development of the CO2 emission trading market, it is also important to encourage use of renewable energy sources.

In the framework of the competitive electricity markets, electricity price forecasting is important for market participants in a deregulated electricity market. Rather than forecasting the value, market participants are sometimes more interested interval of the peak electricity price forecasting. Forecasting the peak price is essential for estimating the uncertainty involved in the price and thus is highly useful for making generation bidding strategies and investment decisions. The choice of the forecasting model becomes the important influence factor how to improve price forecasting accuracy. This paper proposes new approach to reduce the prediction error at occurrence time of the peak electricity price, and aims to enhance the accuracy of the next day electricity price forecasting. In the proposed method, the weekly variation data is used for input factors of the ANN at occurrence time of the peak electricity price in order to catch the price variation. Moreover, learning data for the ANN is selected by rough sets theory at occurrence time of the peak electricity price. This method is examined by using the data of the PJM electricity market. From the simulation results, it is observed that the proposed method provides a more accurate and effective forecasting, which helpful for suitable bidding strategy and risk management tool for market participants in a deregulated electricity market.

This paper analyses the recent evolution of the electricity price of one of the EastEuropean countries’ Balancing Markets (BM) – Romania, aiming to understand the prices trend and predict them in the current economic and geopolitical context. This is especially important as the electricity producers have to allocate their output between wholesale electricity market, ancillary services markets and BM targeting to maximize value and achieve a sustainable economic development. Therefore, in this paper, we propose an AI-powered electricity price forecast using several types of standout Machine Learning (ML) algorithms such as classifiers and regressors to predict the electricity price on BM. This approach, consisting of two steps, identifies the imbalance sign and significantly enhances the performance of the price forecast. The proposed method offers valuable insights into the market participants’ trading opportunities using two prediction solutions. The first prediction solution consists of averaging the results of five ensemble ML algorithms. The second one consists in weighting the results of the five forecasting ML algorithms using either a linear regression or a decision tree algorithm. Thus, we propose to combine supervised and unsupervised ML algorithms and find the fundamentals for creating optimal bidding strategies for electricity market players.

Day ahead price forecasting of electricity markets by a mixed data model and hybrid forecast method

A comparative analysis of neural networks and statistical time-series methods in the electricity price forecasting of the Iran power market is conducted. An error back propagation algorithm is applied to train neural networks. The time series methods discussed in previous research were considered as benchmarks in the comparison. Two case studies are investigated based on the proposed methods to examine the impact of load ramp in the electric energy price. Only the timeseries of energy price is considered as inputs in the first case and both of the price and load are considered as inputs in the second case. The computational results validate the capability of neural networks in energy price forecasting and indicate that the load ramp is one major factor for determining the energy price.

IntroductionThe monitoring of the electricity market and preventing incorrect usage behavior are well-known issues in the competitive electricity market (see for example Marulanda et al., 2006; Nagypal et al., 2015; Strielkowski and Lisin, 2016; Nagaj, 2016) that are well-embedded into the strategy of sustainable energy development (Streimikiene et al., 2016). The monitoring of the electricity market plays a key role in the free electric energy market (Kasperowicz, 2011; Lisin et al., 2016). Supervisory actions are important for the System Operator and the Regulatory Commissions in order to ensure the efficient market performance. For Federal Energy Regulatory Commission in USA (Haymes, 2000), the market monitoring has been identified as a standard market design component and a basic function in the deregulated electricity market. According to Rahimi et al. (2003) and London Economics et al. (2007), the electricity market monitoring could be organized as an internal system or an institution in the Independent System Operators (ISO) structures, as well as could function also outside of the ISO structures and can provide service for ISO or other recipients (Goldman and Lesieutre, 2004).Furthermore, the availability of a well-developed monitoring system would also help agents to optimize their bidding strategies (David and Wen, 2000; Dun-nan et al., 2006; Pires Manso et al., 2015). Currently agents are concentrating on increasing their efficiency in order to maximize profits through a complicated system of product and sales management. Due to the fact that the process of decision making is very complex, whereas complete information is often unavailable, the electricity market monitoring appears to be difficult (London Economics, 2007). Therefore, if the market monitoring system functions properly, the transparency and reliability of the information concerning the electricity market dynamics may be ensured.The main activities of market monitoring as described in Rahimi et al. (2003), London Economics et al. (2007), (Dehdashti, 2005) and (Guler et al., 2005) include: 1) Detection of attempts to exercise market power and fraudulent behavioral, 2) Monitoring of market performance, 3) Identification of market design imperfection, 4) Transmission and generation blackout controlling, 5) Market participants' behavioral monitoring (activities and transactions).After an in-depth analysis of the existing literature devoted to exploring the issue of electricity market monitoring, it is possible to state that the information regarding the general methods of monitoring are available. However, there is a deficit of information related to the monitoring techniques focused on economical parameters of Spanish electricity market. Furthermore, the conducted analyses of the market structure and of the available source data related to Spanish power system have proven that that there is a need for a modification or a creation of new methods and indices, which shall be applied in order to correctly monitor the economical parameters of Spanish electricity market.This paper firstly introduces the concept and functions of the electricity market monitoring system (section I). Subsequently, the main methods and indices applied to the monitoring (market power, market performance and behavioral monitoring) are outlined (section II). Section III outlines the particularities of Spanish market. Subsequently, section IV presents conclusions.1. The global overview of the monitoring systemThe result of the extensive analysis of the existing literature on market monitoring is represented graphically in Figure 1.The Figure 1 presents the market monitoring framework: the elements and the correlation between them. This extensive analysis is based inter alia on the following sources: London Economics et al. (2007), Goldman and Lesieutre (2004), Dun-nan et al. (2006), Chakrabarti et al. (2008), Newbery (2004), Asgari et al. …

The model of chaotic sequences based on adaptive particle swarm optimization arithmetic combined with seasonal term

Electricity market liberalization and the absence of cost-efficient energy storage technologies have led to the transformation of state-owned electricity companies into complex electricity market entities, each having a different time horizon. Deregulation has intensified competition, giving rise to increased uncertainty caused by a multitude of interrelated exogenous factors, resulting in unexpected fluctuations in electricity prices. As a consequence, market participants encounter elevated risks and seek effective mitigation strategies. In this paper, the challenges described in the literature are addressed by studying price distribution histograms in the German and Finnish electricity markets. The objective is to identify normal price intervals that can serve as a foundation for an integrated Day-Ahead price forecasting methodology. A novel approach utilizing the Extreme Learning Machine in combination with Bootstrap intervals is proposed and applied to both markets. The findings demonstrate that Bootstrap intervals effectively capture normal prices, whereas extremely high prices typically align with the upper limits of Bootstrap intervals. Conversely, negative prices tend to fall outside the lower boundaries of the intervals. In order to assess the performance of the proposed methodology, a comparative analysis of its forecasting accuracy against the well-established Generalized AutoRegressive Conditional Heteroskedasticity and AutoRegressive Fractionally Integrated Moving Average models is conducted. In addition, both the computational efficiency and forecasting accuracy of the Extreme Learning Machine in comparison to the Artificial Neural Network are assessed. The results reveal the superior efficiency of the Extreme Learning Machine. The developed forecasting model could potentially assist market participants in making well-informed decisions and executing optimal bidding strategies in response to various scenarios before the Day-Ahead market closes. Notably, the proposed methodology transcends the limitations of fixed price thresholds and effectively addresses market nuances, including the occurrence of negative prices, thus offering a more comprehensive approach for electricity price forecasting.

With the continuous promotion of the unified electricity spot market in the southern region, the formation mechanism of spot market price and its forecast will become one of the core elements for the healthy development of the market. Effective spot market price prediction, on one hand, can respond to the spot power market supply and demand relationship; on the other hand, market players can develop reasonable trading strategies based on the results of the power market price prediction. The methods adopted in this paper include: Analyzing the principle and mechanism of spot market price formation. Identifying relevant factors for electricity price prediction in the spot market. Utilizing a clustering model and Spearman’s correlation to classify diverse information on electricity prices and extracting data that aligns with the demand for electricity price prediction. Leveraging complementary ensemble empirical mode decomposition with adaptive noise (CEEMDAN) to disassemble the electricity price curve, forming a multilevel electricity price sequence. Using an XGT model to match information across different levels of the electricity price sequence. Employing the ocean trapping algorithm-optimized Bidirectional Long Short-Term Memory (MPA-CNN-BiLSTM) to forecast spot market electricity prices. Through a comparative analysis of different models, this study validates the effectiveness of the proposed MPA-CNN-BiLSTM model. The model provides valuable insights for market players, aiding in the formulation of reasonable strategies based on the market's supply and demand dynamics. The findings underscore the importance of accurate spot market price prediction in navigating the complexities of the electricity market. This research contributes to the discourse on intelligent forecasting models in electricity markets, supporting the sustainable development of the unified spot market in the southern region.

In a competitive electricity market, forecast of energy prices is a key information for the market participants. However, price signal usually has a complex behavior due to its nonlinearity, nonstationarity, and time variancy. In spite of all performed researches on this area in the recent years, there is still an essential need for more accurate and robust price forecast methods. In this paper, a combination of a feature selection technique and neural network (NN) is proposed for this purpose. The feature selection method is a modified version of the relief algorithm, proposed for the feature selection of price forecasting. Then, by means of the most relevant, explanatory and irredundant features, a neural network (NN) predicts the next values of the price signal. The adjustable parameters of the whole method are fine-tuned by a cross-validation technique. The proposed method is examined on PJM electricity market, forecasting day-ahead locational marginal prices (LMPs), and compared with some of the most recent price forecast methods especially some other popular and validated feature selection techniques. These comparisons indicate the validity and robustness of the proposed forecasting method.

Day-ahead electricity price prediction applying hybrid models of LSTM-based deep learning methods and feature selection algorithms under consideration of market coupling