Robust Energy Management System for Multi-Source DC Energy Systems—Real-Time Setup and Validation

This brief aims at providing a proof of concept of a systematically designed linear parameter varying (LPV)/ $\mathcal {H}_\infty $ -based energy management system (EMS) for coordinated multi-variable control of multi-source electrical systems. A three-source electrical system representing the power supply system on board of an electric vehicle has been chosen as a representative example of irregular and generally not a priori known load variation. The power supply system is composed of the fuel cell, battery, and supercapacitor. Each power source is coupled to a dc–dc converter, all converters being connected in parallel to a common dc-bus in order to feed the load represented by the vehicle’s electrical motor. The system is modeled as an LPV system—as its operating point depends on the load—and the control objectives are cast into the $\mathcal {H}_\infty $ formalism as a disturbance-rejection problem. A dedicated hardware-in-the-loop system was built for the proof-of-concept purpose, with real-world battery and supercapacitor being used, while the fuel cell system is entirely emulated. A dSPACE MicroAutoBox II device embeds the designed EMS, due to its flexibility and ease of programming with MATLAB. A driving cycle from [Institut Français des Sciences et Technologies des Transports, de l’Aménagement et des Réseaux (IFSTTAR)] is chosen as a pertinent scenario of load variation due to its rich frequency content able to challenge all the three sources. Effectiveness of the EMS is assessed in relation to the imposed control objectives—dc-bus voltage regulation, dynamical separation of power sources’ current variations depending on the specialization range of each source, and imposing desired steady-state behavior for each of the three power sources—with very promising results.