• Energy Research

In self-consumption (SC) configurations, energy management systems (EMSs) are increasingly being implemented to maximise the self-consumption ratio (SCR). Recent studies have demonstrated that prediction-based EMSs significantly enhance decision-making capabilities compared to non-predictive EMSs. This paper presents the design, implementation, and testing on a real system of two machine learning (ML)-type predictive models capable of forecasting the electricity consumption of an individual building using a small dataset. A nonlinear autoregressive with exogenous input (NARX) neural network model and a support vector regression (SVR) model were designed and compared. These models predict day-ahead hourly electricity consumption using forecasted meteorological data from Meteo Galicia (MG) and building occupancy data, both automatically obtained and pre-processed. In order to compensate for the lack of recurrence of the SVR model, the effect of introducing an additional input, a time vector, was analysed. It is proved that both ML models trained with a small dataset are able to predict the next day’s average hourly power with a mean MAPE below 13.96% and a determination coefficient (R2) greater than 0.78. The model that most accurately predicts the hourly average power of a week is the SVR, which achieves a mean MAPE and R2 of 10.73% and 0.85, respectively.

The evolution towards the Industry 4.0 concept implies a huge development of the connectivity and the smart automation in all the industrial areas. For the chemical processes, the supervision systems are generally fairly basic, and a number of operations are still conducted by the operators. In order to simplify their daily tasks, expert systems able to discern each stage of the process and any possible drifts can be considered. To be implemented, the expert system needs a knowledge base which, in our case, will contain a set of images representative of the main stages of the process. The embedded system which will make the image capture must be autonomous from an energy point of view. Thus, the energy consumption optimization is necessary to ensure the maximal lifetime for the system. In this frame, this paper presents the energy management solution applied to an image acquisition system, based on a Raspberry Pi, for the chemical industry. The energy saving is achieved by the choice of the physical architecture, together with the application of a duty-cycle strategy.

RESUMEN: El objetivo de este trabajo ha sido diseñar un controlador digital robusto para controlar la tensión de un sistema de energía híbrido aislado. La única carga del sistema es resistiva y variable, y está alimentada mediante un inversor asociado a un filtro LCL. Después de plantear el problema, se explica cómo se obtiene el modelo de la planta. A continuación se analiza su comportamiento dinámico oscilante y la dependencia de ese comportamiento con respecto a la carga resistiva. Se describe después la estructura y el diseño del controlador global que contiene dos lazos de control en cascada, el primero de corriente y el segundo de tensión. Se diseñan los controladores de los dos lazos haciendo hincapié en su robustez y su capacidad de amortiguar el sistema en lazo cerrado. Para acabar, se realizan simulaciones en las que se estudia el comportamiento dinámico de los dos lazos en regulación y seguimiento en distintos puntos de funcionamiento. Los resultados demuestran la calidad del control en términos de robustez, amortiguamiento y rapidez de la respuesta. Palabras Clave: Sistema de Energía Híbrido, Control Robusto, Asignación de Polos, Control Digital, Sistema no Estacionario, Inversores

Abstract Nowadays, the importance of electrical generation based on renewable energies is increasing, due to its low emissions of greenhouse gases. At the same time, Distributed Generation and Microgrids (MG) are becoming an important research line because of their peculiar characteristics. MGs are composed of small power sources which can be renewable, placed near customer sites. Moreover, they have the inherent property of islanding: the disconnection of either the MG from the main grid or a portion of a MG from the rest of the MG. There are two kinds of islanding: intentional or planned (for maintenance purposes), and unintentional or unplanned. The latter is mainly due to disturbances and it is used to avoid damages in sources and loads. It is the most critical case because it must be detected as soon as possible to activate all the control systems which allow continuing the energy production and distribution despite the disconnection. In islanding, it is crucial to ensure the power and the electrical signal quality. In grid-connected mode, the inverters use the electrical signal of the main grid as reference. Once in islanding, the main grid reference is lost and new control techniques for the inverters are needed in order to obtain the correct values of voltage magnitude and frequency in the MG. The main objective of this paper is to make a survey on MGs focussed on two important features: unplanned islanding and control of inverters in that scenario. The idea is to present the basic architectures and regulation techniques of MGs and to study the islanding behaviour, mainly the different detection techniques and the inverters’ control once islanded.

In the context of global energy demand increase, working on energy efficiency is essential. This paper deals with energy harvesting on car suspensions. In order to have a real added value, some criteria must be considered: the need to design a system that would be easily integrated into cars, the possibility to locally use the recovered energy to add new functionalities that can improve the security or the comfort of the car, and the necessity to not degrade and, if possible, to improve (semi-active or active dampers) the performances of the suspension. From the mechanical point of view, the functional analysis is used to define and to characterize the main suspension parts, to investigate the connexions and the energy flows and to identify the key elements for energy recovery. Then, quarter car and half car models implemented with Matlab/Simulink software are presented in order to evaluate the quantity of energy that could be recovered. Three locations are presented and evaluated. Simulations results will finally give an overview on the implementation opportunities.

ABSTRACTThe traditional electrical grid must evolve to an intelligent smart grid (SG) to achieve efficiency, reliability, environmental and power quality improvements. Communication techniques and protocols will be crucial to manage the SG. This paper studies the communication requirements for SGs and describes the most suitable communication protocols, wired and wireless, with special attention to the latest proposals in this field. The security issues derived from data transmission through electrical grids are also analysed with a deep description of the current best practices employed to assure the SG protection face to intrusions. Copyright © 2013 John Wiley & Sons, Ltd.

The electrical networks are very complex systems, presently in full evolution. With the increasing penetration and apportionment over large areas of the renewable energies, the centralized nature of the power production evolves toward a more distributed form. In this context, the interest of the microgrids, as subcomponents of the main grid, rises especially by the specific services that they can provide. To ensure the stability and the dependability of a microgrid is a difficult task particularly in islanded mode. The microgrid is a mix of different types of elements and its structure may evolve. The multi-agents systems technique, with its distributed and adaptive character, can be an interesting tool for the control of each microgrid component and for the real time decision making related to the microgrid management. The paper starts with a presentation of microgrids and their specific issues, and follows with some basic concepts on multi-agents systems. Then, a review of the proposed techniques and algorithms that implies multi-agents systems in microgrids management is presented, focusing on the safety and the dependability in microgrids.