Non-deposition self-cleansing models for large sewer pipes

Area of Science:

• Environmental Engineering
• Hydraulic Engineering
• Water Resource Management

Background:

• Existing sediment transport models for sewers are limited to smaller pipe diameters (<500 mm).
• There is a need for accurate models applicable to larger sewer systems.

Purpose of the Study:

• To develop and validate new self-cleansing models for sewer sediment transport.
• To evaluate the performance of existing and newly developed models on diverse datasets.
• To provide reliable tools for designing both small and large sewer pipes.

Main Methods:

• Collected new experimental data on a 595 mm diameter sewer pipe.
• Developed two new self-cleansing velocity models based on the new dataset.
• Evaluated and compared model accuracy using root mean squared error (RMSE) and mean absolute percentage error (MAPE).

Main Results:

• Existing literature models show overfitting, with high accuracy on author-specific data but poor generalization.
• Newly developed models demonstrate improved prediction accuracy across various datasets.
• The new models are effective for predicting self-cleansing velocity in pipes with and without deposited beds.

Conclusions:

• Newly developed self-cleansing models offer enhanced accuracy for sewer design across different pipe sizes.
• Overfitting is a significant issue in existing sewer sediment transport models.
• The research provides practical tools for engineers designing urban drainage systems.