Empirical modelling of total suspended solids from turbidity by polynomial neural network in North Eastern India
Ritabrata Roy, Mrinmoy Majumder · Desalination and Water Treatment · 2018
ABSTRACT Suspended solids in water cause turbidity. Suspended solids reduce dissolved oxygen,hinder the respiration of aquatic organisms and induce bacterial growth. Due to such significant impact on water quality and aquatic ecosystem, turbidity and total suspended solids are important water quality parameters. Now, turbidity of water can be estimated easily and instantly by modern devices, but total suspended solids can be measured only by conventional analytical procedure. Such analytical procedure is troublesome, time consuming and requires laboratory setup. This study aims the modelling of total suspended solids from turbidity through polynomial neural network to save the time, effort and cost of analytical method to estimate total suspended solids. Different polynomial neural network methods were used for the modelling. Field data from 176 water bodies of North Eastern India were used for the training and testing of those models. An empirical equation was developed from those models which can calculate total suspended solids from turbidity. The equation was tested to be valid for its high correlation (correlation coefficient and coefficient of determination = 0.999) and very low deviation (mean absolute percentage error = 3.477 or less and root mean square error = 1.904 or less) from the actual values.