An on-line fast model predictive control of highpower ultracapacitors charging current for renewable energy urban rail vehicle

Ultracapacitors have been supplied as main power in urban rail vehicle due to their characteristic of high efficiency. Given the model complexity and hard constraints of the ultracapacitors, the charging current control of ultracapacitors under this condition is challenging as fast charging is required. In this paper, an on-line fast model predictive control is proposed to address the problem based on extended interior-point method. By utilizing fixed barrier parameter and warm start mechanisms, the on-line fast MPC substantially simplifies the solving process and regulates the charging current of ultracapacitors rapidly and precisely. Meanwhile, a Kalman filter is integrated to estimate the system states and to adapt the load uncertainty. Experiment results are presented to demonstrate the effectiveness of the proposed control method.