A Study on the Power Prediction of Substation in DC Urban Railroad Using Train Load Prediction

To achieve carbon emission reductions, minimizing electric energy consumption in railways is essential. Recent studies emphasize active substation control and vehicle load forecasting technologies, though research remains limited. Existing models for energy consumption often struggle with accuracy due to high load variability. This paper introduces a load forecasting method using the random forest model, demonstrating its superiority over Long Short-Term Memory (LSTM) methods, and proposes a technique for predicting substation power consumption based on these forecasts.