Model Development of Nitrification in Premise Plumbing using Artificial Neural Network

Nitrification is the process by which reduced forms of nitrogen are oxidized to produce nitrite and nitrate. When chloramine is used in potable water systems for secondary disinfection, there is a serious potential issue. The usage of monochloramine as a secondary disinfectant is growing in place of free chlorine. Water is treated with ammonia to promote the creation or breakdown of monochloramines. Regulations may be broken as a result of nitrification's detrimental effects on water quality. A study was conducted to look into the quality of the water, the impact of pipe material on the beginning of nitrification, and the effects of nitrification on In-Premise Plumbing. Here, the impact of pipe material on nitrification in premise plumbing was investigated along with examining its effects on water quality. Empirical modeling approach using Artificial Neural Network (ANN) was taken to observe and predict these effects. Input variables of ANN modeling are copper dose concentration, pH level and number of days while output variables are nitrite and ammonia utilization. The best-fitted models are, for ammonia utilization, ANN model with Levenberg–Marquardt algorithm and 50 hidden neurons, which had a coefficient of determination of 0.738 and a mean squared error of 65.8, and, for nitrite utilization, ANN model also using Levenberg–Marquardt algorithm and 50 hidden neurons, which had a coefficient of determination of 0.601 and a mean squared error of 0.0063.