State of Charge Estimation of Supercapacitor Using Artificial Neural Network for Onboard Railway Applications

This paper proposes a novel method based on artificial neural network (ANN) for Estimation of state-of-charge (SOC) of supercapacitors (SC) energy storage device used in railway system for braking energy recovery. Indeed, the SOC of SCs is related to the knowledge of SCs internal voltage which is not accessible for measurement. By Using the available measurement signals, SC current and voltage, as inputs, the developed model is able to estimate the internal voltage of SC energy storage installed on board of the train. The numerical simulation processed in MATLAB/SIMULINK demonstrates the effectiveness of the ANN method to provide a good performance.