International Journal of Scientific Research and Engineering Development

This study was carried out to monitor the concentrations of trihalomethanes (THMs) in both raw and treated water, and to develop a model for predicting THM concentration in drinking water systems within the largest drinking water treatment plant in Bangladesh. A mathematical model was developed using the multiple linear regression (MLR) approach. THM concentration was designated as the dependent variable, while various water quality parameters—namely pH, temperature, COD, TOC, DOC, UV254, SUVA, ammonia nitrogen, bromide, chlorine contact time, chlorine dose, and residual chlorine—were defined as independent variables. Multiple linear regression analysis of the data indicated that COD, ultraviolet light absorbance at 254 nm (UV254), initial chlorine dose, and ammonia concentration were the most significant variables. COD was found to be the most influential parameter responsible for THM formation, followed by UV254, chlorine dose, and ammonia-N. The residual free-chlorine concentration, pH, reaction time, bromide ion, DOC or TOC or SUVA had little or no significance. Analysis of combinations of the TOC, DOC, SUVA, COD and the ultraviolet absorbance indicated that use of COD along with the ultraviolet absorbance alone provided the best prediction of the experimental data. The relationships between the variables were initially examined using simple correlation analysis. Multiple regression analysis was then applied to evaluate the statistically significant variables in the system. The significance level for including a variable in the model was set at 0.05. The developed model provided satisfactory estimations of THM concentrations, with a model regression coefficient of 0.65. Both the R² and Durbin–Watson statistics were found to be statistically significant. The Durbin–Watson value was 2.14, which falls within the acceptable range. Predicted THM concentrations were compared to actual concentrations measured during the sampling program. The results showed a good agreement between measured and calculated THM concentrations (R² = 0.77), indicating that the method presented in this paper can be effectively used to estimate THM concentrations throughout the system. The correlation and regression analyses used to examine the relationship between the independent variables and THM concentrations showed promising results, with strong relationships. Validation of the model revealed no significant differences between predicted and observed values, and the prediction error was low. The model developed in this study can be used to predict THM concentration levels in drinking water supplies under conditions typical of Bangladesh. It is noteworthy to mention that no previous attempts to assess, monitor, and predict THM concentrations in public drinking water have been reported for the country although a large fraction of the population consumes chlorinated public drinking water. Until recently, there was no information available that pertained to the concentration level of THMs in drinking water, lest the development of a model thus making our research almost the first one of this kind.

Md. Serajuddin, Md. Aktarul Islam Chowdhury, Md. Mehedi Hasan, Tanzir Ahmed Khan,"Developing a Multiple Regression Model for Predicting Trihalomethane Concentrations in Drinking Water Supply: A Case Study" International Journal of Scientific Research and Engineering Development, V9(2): Page(1498-1504) Mar-Apr 2026. ISSN: 2581-7175. www.ijsred.com. Published by Scientific and Academic Research Publishing.