A two-stage predictive model to simultaneous control of trihalomethanes in water treatment plants and distribution systems: adaptability to treatment processes

Domínguez-Tello, Antonio; Arias-Borrego, Ana; García Barrera, Tamara; Gómez-Ariza, JoséLuis

A two-stage predictive model to simultaneous control of trihalomethanes in water treatment plants and distribution systems: adaptability to treatment processes

2017

Domínguez-Tello, Antonio | Arias-Borrego, Ana | García Barrera, Tamara | Gómez-Ariza, JoséLuis

The trihalomethanes (TTHMs) and others disinfection by-products (DBPs) are formed in drinking water by the reaction of chlorine with organic precursors contained in the source water, in two consecutive and linked stages, that starts at the treatment plant and continues in second stage along the distribution system (DS) by reaction of residual chlorine with organic precursors not removed. Following this approach, this study aimed at developing a two-stage empirical model for predicting the formation of TTHMs in the water treatment plant and subsequently their evolution along the water distribution system (WDS). The aim of the two-stage model was to improve the predictive capability for a wide range of scenarios of water treatments and distribution systems. The two-stage model was developed using multiple regression analysis from a database (January 2007 to July 2012) using three different treatment processes (conventional and advanced) in the water supply system of Aljaraque area (southwest of Spain). Then, the new model was validated using a recent database from the same water supply system (January 2011 to May 2015). The validation results indicated no significant difference in the predictive and observed values of TTHM (R ² 0.874, analytical variance <17%). The new model was applied to three different supply systems with different treatment processes and different characteristics. Acceptable predictions were obtained in the three distribution systems studied, proving the adaptability of the new model to the boundary conditions. Finally the predictive capability of the new model was compared with 17 other models selected from the literature, showing satisfactory results prediction and excellent adaptability to treatment processes.

Показать больше [+]