Search

The fast increase in electric vehicle (EV) usage in the last 10 years has raised the need to properly forecast their energy consumption during charge. Lithium-ion batteries have become the major storage component for electric vehicles, avoiding their overcharge can preserve their health and prolong their lifetime. This paper proposes a Machine Learning model based on the K-Nearest Neighbors classification algorithm for EV charging session duration forecast. The model forecasts the duration of the charge by assigning the event to its correct class. Each class contains the charging events whose duration is comprised of a certain interval. The only information used by the algorithm is the one available at the beginning of the charging event (arrival time, starting SOC, calendar data). The model is validated on a real-world dataset containing records of charging sessions from more than 100 users, a sensitivity analysis is performed to assess the impact of different information given as input. The effectiveness of the model with respect to the benchmark models is demonstrated with an increase in performance.

The increasing adoption of electric vehicles poses new problems for the electrical distribution network. For this reason, proper electric vehicle forecasting will be of fundamental importance for a predictive energy management system, which could greatly help the operation of the grid. This paper proposes a comprehensive novel methodology to forecast single charging sessions of electric vehicle and the resulting cumulative energy forecast of the charging infrastructure. Historical charging sessions are first clustered on the basis of similar user characteristics and their respective probability density functions are defined. From this, every charging session is predicted with a triplet of parameters, namely the arrival time, the charging duration and the average power expected during the process. The proposed method has been evaluated by considering a real case study. The results showed the ability to greatly improve the accuracy with respect to the chosen benchmark, both in terms of energy required by the station and the predicted number of charging sessions. The overall performance measured by Skill Score is 0.37 for the year 2019.

Day-ahead power forecasting is an effective way to deal with the challenges of increased penetration of photovoltaic power into the electric grid, due to its non-programmable nature. This is significantly beneficial for smart grid and micro-grids application. Machine learning and hybrid approaches are well assessed techniques, able to provide effective forecasting with a data-driven approach based on previous measurements from existing power plants. Ensemble methods can be employed to increase solar power forecasting accuracy, by running several independent forecasting models in parallel. In this paper, a novel selective approach is proposed and assessed, where independently trained neural networks are evaluated in terms of accuracy, in order to properly select a suitable forecasting. Moreover, in order to reduce the associated computational burden, suitably developed new normalization approaches are proposed and evaluated. The considered experimental case study shows that the combination of the proposed procedures is able to increase accuracy and to mitigate the overall computational load, resulting in a simple and lightweight algorithm. Additionally, a comparison with other commonly used techniques has shown that the proposed approach is robust with respect to dataset limited size and discontinuities.

Nowadays, Information and Communication Technology (ICT) infrastructures play a crucial role for human beings, providing essential services at astonishing speed. Nevertheless, such a centrality of those infrastructures attracts the interest of ill-motivated actors that target such infrastructures with cyberattacks that are every day more sophisticated and more disruptive. In this alarming context, selecting the optimal set of countermeasures represents a primary need to react against the appearance of potentially dangerous threats effectively. With the motivation to contribute to develop ing faster and more effective response capabilities against them, the paper at hand introduces a novel cybersecurity reaction methodology based on Artificial Immune Systems (AIS), for which the evolutionary computing paradigm has been adopted. By leveraging the outstanding properties of these bio-inspired techniques, the selected countermeasures to defeat cyberthreats through cloning and mutation phases in an effort to improve the quality of the solution from a quantitative perspective, being able to adjust the risk to which the assets of the protected system are exposed. Exhaustive experiments demonstrate the feasibility of the proposed approach, reducing the risk in a more than acceptable time lapse.

The amount of data generated by the current interconnected world is immeasurable, and a large part of such data is publicly available, which means that it is accessible by any user, at any time, from anywhere in the Internet. In this respect, Open Source Intelligence (OSINT) is a type of intelligence that actually benefits from that open nature by collecting, processing and correlating points of the whole cyberspace to generate knowledge. In fact, recent advances in technology are causing OSINT to currently evolve at a dizzying rate, providing innovative data-driven and AI-powered applications for politics, economy or society, but also offering new lines of action against cyberthreats and cybercrime. The paper at hand describes the current state of OSINT and makes a comprehensive review of the paradigm, focusing on the services and techniques enhancing the cybersecurity field. On the one hand, we analyze the strong points of this methodology and propose numerous ways to apply it to cybersecurity. On the other hand, we cover the limitations when adopting it. Considering there is a lot left to explore in this ample field, we also enumerate some open challenges to be addressed in the future. Additionally, we study the role of OSINT in the public sphere of governments, which constitute an ideal landscape to exploit open data.

Energy systems around the world are undergoing substantial changes, with an increasing penetration of Renewable Energy Sources. For this reason, the availability of a pool of suitable forecasting models specific for the needed time horizon and task is becoming crucial in the grid operation. In addition, nowcasting techniques aiming at provideing the power forecast for the immediate future, are more often investigated due to the spread of micro-grids and the need of facing changing electrical market environments. In this paper a novel comprehensive methodology aiming at computing the PV power forecast on different time horizons and resolutions is introduced. Moving from the 24-hours ahead prediction provided by the Physical Hybrid Artificial Neural Network (PHANN), a technique to refine the power forecast for the following 3 hours with an hourly granularity is analyzed, leveraging on newer information available during the operations. Moreover, in order to provide the power forecast for the following 30 minutes on a minutely basis, an innovative modification of a statistical technique is proposed, the robust persistence. The proposed comprehensive approach allowed to greatly reduce the overall error committed when compared with the benchmark models. Finally, the proposed methodology is validated and tested on a freely available database consisting on different parameters recorded at both the meteorological and photovoltaic test facility at SolarTechLAB, Politecnico di Milano, Milan.