In this paper, we investigate the use of Physics-Informed Neural Networks (PINNs) for the solution of an optimal control problem related with the optimal control of a stationary electric storage system (ESS) installed in a service station for plug-in electric vehicles (PEVs). The ESS is used to balance the PEVs charging power, in order to mitigate the impact on the grid, and keep low the power flow at the point of connection with the grid. The proposed PINN is trained in order to learn the optimality conditions of the optimal control problem so that, after training, it can provide the solution with no significant computation effort. This one represents a promising alternative to the analytical computation of the optimal control (which is possible only in very simple settings), and to the solution of the optimal control problem with numerical methods, which requires significant time in the more complex and realistic settings. Numerical simulations are presented to evaluate the effectiveness of the trained PINN in solving the optimal control problem.

In digital transformation, Industrial Data Space (IDS) is a key enabler for industry 4.0 to improve the industrial process, efficiency, and energy consumption by exploiting learning-based techniques. The present paper focuses on improving the decision-making process in complex industrial environments by developing a Deep Reinforcement Learning (DRL) based real-time assistant. Mainly, we address a use case from the space industry to improve the launcher throughput and efficiency and reduce cost by optimally managing the industrial resources. A mathematical formulation of the Industrial Production System (IPS) and a simulated environment are developed to train the DRL-based Proximal Policy Optimization (PPO) agent. The proposed method is scalable and in line with the dynamic nature of the industrial production systems to overcome the domain-dependent heuristics extensively used in the manufacturing industry. Furthermore, simulation results show that the proposed method can provide industrial operators and managers with a real-time decision support system to increase the Return on Assets.

This paper presents a decentralised control strategy for the management of simultaneous charging sessions of electric vehicles. The proposed approach is based on the model predictive control methodology and the Lagrangian decomposition of the constrained optimization problem which is solved at each sampling time. This strategy allows the computation of the charging profiles in a decentralised way, with limited information exchange between the electric vehicles. The simulation results show the potential of the proposed approach in relation to the problem of shaving the aggregated power withdrawal from the electricity distribution grid, while still satisfying drivers’ preferences for charging.

This paper proposes a control strategy for an Energy Storage System (ESS) installed on a secondary substation of an electricity distribution line in order to mitigate power variations with respect to the day-ahead planning caused by renewable energy sources on the distribution line.In particular, the aim of the controller is to keep the power profile of at primary substations close to a reference profile foreseen on a day-ahead basis while guaranteeing the stable operation of its ESS, in terms of their state-of-charge dynamics. The inclusion of the ESS contribution to the network operation is attained by the integration of properly defined power flow constraints in a Model Predictive Control Framework. The proposed approach has been validated through numerical simulations, representative of real operative scenarios.