Mislav Hruškar

Model predictive control of battery energy storage for application in railway transport systems

 master thesis, 2016.

 

In order to increase energy and economic efficiency of railway transport systems, use of different types of energy storage is becoming more frequent. Energy storage are used to make better use of energy produced by regenerative braking of trains, reduce peak power demand, minimize system costs and ultimately apply advanced control concepts that would enable achieving a better working performance and integration of railway system in advanced power systems.

For the purposes of control problem formulation, it is necessary to identify a good mathematical model of the process which is managed. An algorithm for the identification of PWA model is developed that involves clustering-based procedure. By using the developed algorithm, piecewise affine (PWA) state space model is identified. After identification, by using PWA battery model, a concept of model predictive control is applied with battery energy storage in the railway system in order to increase energy efficiency and minimize costs of the system. Finally, the control algorithm, that is the objective function is extended by a quadratic term which while achieving better economic energy exchange balance also extends battery lifetime.

 

Keywords: battery storage, energy efficiency, piecewise affine model estimation, battery state of health preservation, model predictive control, mixed-integer quadratic program