Based on the predictions of fossil fuels depletion in the following years, as well as their negative impact due to generated exhaust fumes, eco-friendly generators, and more specifically wind generators, have arisen as a so-lution for the electric demand challenge. Wind energy consists in extracting energy from wind speed, and because of the uncertain and intermittent behaviour of this meteorological parameter, wind turbines output power cannot be optimally exploited. Although the vast majority of the research in wind speed forecasting field has consisted in the purpose of novel algorithms, these studies have not made a pre-processing step of the data in order to try to extract the maximum information from databases. Therefore, the goal of this paper consists in analysing whether the combination of time-frequency decomposition of wind speed data with different machine learning algorithms can increase the accuracy of current wind speed predictions for 10 min ahead. Obtained error metrics demonstrated that the deviation of developed wind speed forecaster was lower than 0.1% in 62% of the validation database. In addition, the root mean square error of the final forecaster was 0.34 m/s. This means an accuracy increase of 51.5% if the result is compared with benchmark model's results.

Renewable energy sources are becoming more relevant in recent decades in power generation, leading to investment in developing efficient systems. Specifically, in photovoltaic energy, modular converters are attracting interest since their characteristics enable them to work at high voltage and optimize the generated energy. However, the control strategies found the literature limit the scalability potential of modular converters. The main aim of this paper is to propose a scalable control strategy for a grid-tied CHB (Cascaded H-Bridge) converter for large-scale photovoltaic power plants. The control proposed is able to take full advantage of converter scalability and modularity, being based on the parameters needed for bipolar sinusoidal PWM (Pulse Width Modulation), and thus reducing the calculus required and simplifying its implementation. Power imbalances are overcome including the zero-sequence vector injection to allow power exchange between phases. Furthermore, the parameter used for power factor control has been discretized and discretization time analysis shows that the control strategy is stable and does not require a high-speed communication channel. For validation purposes, simulations are conducted on a downsized 12 H-bridge model.

Due to the expected lack of fossil fuels in near future as well as climate change produced by greenhouse effect as consequence of environmental emissions, renewable energy generation, and specifically solar photovoltaic generation, has become relevant in present energy generation challenge. Photovoltaic generators have strong relationship with solar irradiation and outdoor temperature in energy generation process. These meteorological parameters are volatile and uncertain in nature so, unexpected changes on these parameters produce variations on solar photovoltaic generators' output power. While many researchers have been focused in recent years on the development of novel models for forecasting involved meteorological parameters in photovoltaic generation, they commonly do not consider an analysis step of the data before using it in the developed models. Hence, the aim of this study consists in assembling a wavelet based time-frequency analysis of the used data with deep learning neural networks to forecast solar irradiation, in next 10 min, to compute solar photovoltaic generation. Results of the validation step showed that the deviation of the proposed forecaster was lower than 4% in 90.60% of studied sample days. Final forecaster's root mean square error was 35.77 W/m(2), which was an accuracy improvement of 37.52% compared against persistence benchmark model.

In recent years, the photovoltaic generation installed capacity has been steadily growing thanks to its inexhaustible and non-polluting characteristics. However, solar generators are strongly dependent on intermittent weather parameters, increasing power systems' uncertainty level. Forecasting models have arisen as a feasible solution to decreasing photovoltaic generators' uncertainty level, as they can produce accurate predictions. Traditionally, the vast majority of research studies have focused on the development of accurate prediction point forecasters. However, in recent years some researchers have suggested the concept of prediction interval forecasting, where not only an accurate prediction point but also the confidence level of a given prediction are computed to provide further information. This paper develops a new model for predicting photovoltaic generators' output power confidence interval 10 min ahead, based on deep learning, mathematical probability density functions and meteorological parameters. The model's accuracy has been validated with a real data series collected from Spanish meteorological stations. In addition, two error metrics, prediction interval coverage percentage and Skill score, are computed at a 95% confidence level to examine the model's accuracy. The prediction interval coverage percentage values are greater than the chosen confidence level, which means, as stated in the literature, the proposed model is well-founded. (c) 2021 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

In recent years, various forecasters have been developed to decrease the uncertainty related to the intermittent nature of photovoltaic generation. While the vast majority of these forecasters are usually just focused on deterministic or probabilistic prediction points, few studies have been carried out in relation to prediction intervals. In increasing the reliability of photovoltaic generators, being able to set a confidence level is as important as the forecaster's accuracy. For instance, changes in ambient temperature or solar irradiation produce variations in photovoltaic generators' output power as well as in control parameters such as cell temperature and open voltage circuit. Therefore, the aim of this paper is to develop a new mathematical model to quantify the confidence interval of ambient temperature in the next 10 min. Several error metrics, such as the prediction interval coverage percentage, the Winkler score and the Skill score, are calculated for 95%, 90% and 85% confidence levels to analyse the reliability of the developed model. In all cases, the prediction interval coverage percentage is higher than the selected confidence interval, which means that the estimation model is valid for practical photovoltaic applications.

Very short-term load demand forecasters are essential for power systems' decision makers in real-time dispatching. These tools allow traditional network operators to maintain power systems' safety and stability and provide customers energy with high reliability. Although research has traditionally focused on developing point forecasters, these tools do not provide complete information because they do not estimate the deviation between actual and predicted values. Therefore, the aim of this paper is to develop a very short-term probabilistic prediction interval forecaster to reduce decision makers' uncertainty by computing the predicted value's upper and lower bounds. The proposed forecaster combines an artificial intelligence-based point forecaster with a probabilistic prediction interval algorithm. First, the point forecaster predicts energy demand in the next 15 min and then the prediction interval algorithm calculates the upper and lower bounds with the user's chosen confidence level. To examine the reliability of proposed forecaster model and resulting interval sharpness, different error metrics, such as prediction interval coverage percentage and a skill score, are computed for 95, 90, and 85% confidence intervals. Results show that the prediction interval coverage percentage is higher than the confidence level in each analysis, which means that the proposed model is valid for practical applications.

Photovoltaic generation has arisen as a solution for the present energy challenge. However, power obtained through solar technologies has a strong correlation with certain meteorological variables such as solar irradiation, wind speed or ambient temperature. As a consequence, small changes in these variables can produce unexpected deviations in energy production. Although many research articles have been published in the last few years proposing different models for predicting these parameters, the vast majority of them do not consider spatiotemporal parameters. Hence, this paper presents a new solar irradiation forecaster which combines the advantages of machine learning and the optimisation of both spatial and temporal parameters in order to predict solar irradiation 10 min ahead. A validation step demonstrated that the deviation between the actual and forecasted solar irradiation was lower than 4% in 82.95% of the examined days. With regard to the error metrics, the root mean square error was 50.80 W/m(2), an improvement of 11.27% compared with the persistence model, which was used as a benchmark. The results indicate that the developed forecaster can be integrated into photovoltaic generators' to predict their output power, thus promoting their inclusion in the main power network. (C) 2021 Elsevier Ltd. All rights reserved.

Electrical load forecasting plays a crucial role in the proper scheduling and operation of power systems. To ensure the stability of the electrical network, it is necessary to balance energy generation and demand. Hence, different very short-term load forecast technologies are being designed to improve the efficiency of current control strategies. This paper proposes a new forecaster based on artificial intelligence, specifically on a recurrent neural network topology, trained with a Levenberg-Marquardt learning algorithm. Moreover, a sensitivity analysis was performed for determining the optimal input vector, structure and the optimal database length. In this case, the developed tool provides information about the energy demand for the next 15 min. The accuracy of the forecaster was validated by analysing the typical error metrics of sample days from the training and validation databases. The deviation between actual and predicted demand was lower than 0.5% in 97% of the days analysed during the validation phase. Moreover, while the root mean square error was 0.07 MW, the mean absolute error was 0.05 MW. The results suggest that the forecaster's accuracy is considered sufficient for installation in smart grids or other power systems and for predicting future energy demand at the chosen sites.

The aim of this study was to develop an artificial intelligence-based tool that is able to predict wind power density. Wind power density is volatile in nature, and this creates certain challenges, such as grid controlling problems or obstacles to guaranteeing power generation capacity. In order to ensure the proper control of the traditional network, energy generation and demand must be balanced, yet the variability of wind power density poses difficulties for fulfilling this requirement. This study addresses the complex control in systems based on wind energies by proposing a tool that is able to predict future wind power density in the near future, specifically, the next 10 min, allowing microgrid's control to be optimized. The tool is validated by examining the root mean square error value of the prediction. The deviation between the actual and forecasted wind power density was less than 6% for 81% of the examined days in the validation step, from January 2017 to August 2017. In addition, the obtained average deviation for the same period was 3.75%. After analysing the results, it was determined that the forecaster is accurate enough to be installed in systems that have wind turbines in order to improve their control strategy.

The development of data-driven models for the identification of the degradation state of industrial components is challenged by several issues, such as the unavailability of large datasets containing historical data, the presence of measurement noise, and the intrinsic stochasticity of the degradation process. The proposed method merges the degradation state assessments provided by self-organizing maps (SOMs) trained using data collected from the component under test (component-based) with that provided by SOMs trained using data collected from a fleet of similar components (population-based). Within an ensemble approach, the outcomes of the SOM models are then aggregated using a dynamic weights proportional method based on the individual model local performances in situations similar to the one under test. The proposed SOM-based ensemble approach has been verified with respect to an experimental case study concerning the identification of the degradation state of insulated-gate bipolar transistors, which are known as one of the most critical components in power systems.

The HEMIS project is concerned with health monitoring of electrical powertrain in order to enhance the maintainability and safety of electric vehicles. Electromagnetic emissions from electrical powertrain will change the electromagnetic environment in vehicles, which may have implications for electromagnetic compatibility and, because of the high electrical power levels, human exposure to electromagnetic fields. This paper outlines some of the electromagnetic issues under investigation in the project, including limitations of existing test methods for electromagnetic compatibility, and a lack of test methods for the assessment of in-vehicle human exposure, as well as for both electromagnetic compatibility and human field exposure issues relating to wireless charging of traction batteries.