Predicting unregulated disinfection by-products in water distribution networks using generalized regression neural networks

Disinfection by-products (DBPs) formation in water distribution networks (WDNs) is a common type of quality failure. A reliable DBPs modeling can be way to prevent In this study, generalized regression neural network (GRNN)-based models were developed predict the occurrence three unregulated i.e. dichloroacetonitrile (DCAN), trichloropropanone (TCP), and trichloronitromethane (TCNM). Water sampling data several WDNs used develop models. parameters regulated as predictors The results validated verified. Besides, key identified followed by sensitivity analysis. coefficient determination (R2) GRNN-based was >75% for DCAN TCP; whereas TCNM, R2 < 45% observed. exhibited better prediction accuracy compared with recently multiple linear proposed framework other contaminants.