Hydrogen sulfide production assessment based on sewage physicochemical properties using artificial neural network

Abstract

In sewers, the emission of hydrogen sulfide is of great danger to the environment. The aim of this work is two folds, firstly to explore sewage water characteristics as key element of sulfide estimation in water phase. Secondly, to quantify the influence of each parameter on sulfide concentration, applying statistic method of artificial neural network. The obtained data is collected from Morocco, Casablanca city. The sulfide concentration is correlated to physicochemical parameters using artificial neural network (ANN). The reliability of the ANN model has been successfully tested with a correlation coefficient equal to 95% and a mean squared error of 5%. Hence, the model is used to calculate the importance of the parameters that affects sulfide production using Garson’s algorithm. Results show that the water quality meets multiple required conditions for sulfide buildup. Also, temperature has the most significant impact on sulfide production, followed by BOD5.