Predicting wastewater temperatures in sewer pipes using abductive network models.

Area of Science:

• Environmental Engineering
• Hydrology
• Wastewater Management

Background:

• Accurate wastewater temperature prediction is crucial for sewer system management.
• Existing models may be complex or require extensive data.
• Abductive predictive models offer a simpler alternative with reduced computational demands.

Purpose of the Study:

• To develop and validate simple predictive models for estimating wastewater temperatures in sewer pipes.
• To identify the key input parameters necessary for accurate wastewater temperature prediction.
• To assess the performance of abductive models in forecasting sewer temperatures.

Main Methods:

• Utilized a predictive modelling technique, specifically abductive models.
• Trained and evaluated models using 12 months of data from five sewer pipes (33,900 training, 39,000 evaluation entries).
• Developed distinct models for urban upstream combined sewers and large downstream collector sewers.

Main Results:

• Achieved high correlation between measured and predicted wastewater temperatures (R² up to 0.98).
• Reported root mean square error (RMSE) for temperature change between 0.15 °C and 0.33 °C.
• Identified upstream wastewater temperature and downstream in-sewer air temperature as the sole necessary input parameters for accurate predictions.

Conclusions:

• Simple abductive predictive models are effective for estimating wastewater temperatures in sewer systems.
• Minimal input parameters (upstream wastewater and in-sewer air temperature) are sufficient for accurate predictions.
• The developed models offer a computationally efficient and accurate solution for wastewater temperature monitoring.