Xiuquan Li, Mingwan Zhuang, Weirong Yang, Xiaohong Zhu, and Qiyuexin Wang
Neural network, energy storage power plant, urban rail transit, regenerative energy, dynamic thresholds
The rapid expansion of China’s urban railway transport sector has heightened the need to minimise the substantial train energy consumption. Ground-mounted supercapacitor (SC) energy storage plants have proven to be effective in reclaiming regenerative braking energy and are hence, valuable in reducing train energy usage. To enhance efficiency, the present study refines the energy management approach utilising the conventional voltage–current double-loop control strategy while considering both the operational state of the train and the energy storage power station demand. To conduct simulation experiments, an exclusively designed platform for urban rail transportation power supply systems is employed, with the Batong Line of the local metro used as the sample. The key novelty of this investigation is a regenerative energy (RE) forecasting model that employs a neural network. This model takes into account both the current location and power of the train to make precise predictions about the RE needs of the energy storage facility. The study demonstrated that the improved strategy resulted in a decrease in output energy usage by 0.2 kWh within the substation and a reduction in energy expenditure on the braking resistor by 0.178 kWh, making it highly beneficial in terms of energy conservation when compared to the fixed-threshold strategy. Hence, the study provides a useful technical reference for energy-saving initiatives in urban rail transportation and has practical applicability.
Important Links:
Go Back
