Author(s): Flemming Albers; Birgitta Hoernschemeyer; Malte Henrichs

Linked Author(s):

Keywords: No keywords

Abstract: This study explores the potential of machine learning (ML) for predicting flow in sewer systems, using data generated by the Storm Water Management Model (SWMM). A Long Short-Term Memory (LSTM) network, chosen for its effective handling of time series data, was trained using both hypothetical and real rainfall data. The final model achieved a mean error of 4.6 % in predicting peak flows and demonstrated prediction times that were at least 3 times faster for individual events and up to 600 times faster in mass simulations, compared to an equivalent hydrodynamic model. Tested with 5-fold cross-validation, the model exhibited significant improvements in accuracy and speed compared to its initial version, largely due to enhanced complexity in the model architecture. However, when applied to a more complex sewer system, a decrease in accuracy was observed, underscoring the need for further validation with real-world data. These findings illustrate the promising potential of ML models to boost real-time prediction efficiency but also highlight the need for model adaptation in more complex scenarios.

DOI: https://doi.org/10.71573/sdhc7935

Year: 2025