Control of electric vehicles to help stabilise electric grids
The popularity of electric vehicles (EVs) is growing rapidly and is poised to help foster much greener transport in Europe's near future. There may be yet another important added benefit emerging from widespread use of electric cars, namely their potential to help stabilise and balance demands on national electric grids using their electric storage capabilities. To illustrate, if only 5.5 % of Germany's car fleet becomes electric, the country will have enough storage capacities to provide all required stabilising power to the grid from EVs. In this context, the EU-funded COGENT (Control for green energy-efficient technologies) project looked at how EVs can help stabilise the grid, taking into account their varying connectivity. Keeping in mind that EV sales will triple in the next decade, the project studied the role of drivers and their habits in utilising the batteries for grid-balancing while satisfying charging demands. It then developed optimal control techniques that improve the vehicles' energy efficiency and power-balancing capabilities. To achieve its aims, the project built on recent developments in sensor and communication technologies providing data about the power grid. It designed a control system to exploit the data in order to respond to the needs of the grid. This was developed using model predictive control (MPC) and powerful modelling techniques from machine learning. COGENT also researched learning-based identification methods and how they can be integrated in an MPC environment. It developed safety guarantees for learning-based control methods with focus on combining control theory with learning techniques. Project achievements include new techniques for implementing predictive controllers for energy management of EVs addressing real-time requirements and the distributed structure of the problem. The new learning-based control strategies developed will support efficient energy management of EVs and support their role in optimising power grids. The next step beyond the project is to undertake an extended EV study to demonstrate the benefits of learning-based and distributed predictive control for their optimal energy management. Once this happens, the dream of a smart grid that provides green and efficient energy generation will be well on the way to becoming reality.
Keywords
Electric vehicles, COGENT, electric grids, model predictive control, machine learning, energy management
