A Neuro Fuzzy Method for Hydrochemical Data Processing in River Flow Analysis

Abstract

The production, social, and ecological requirements for maintaining the quality of inland waters necessitated establishing a network of hydrochemical observation posts. The variability in the monitored indicators required implementing regular chemical and analytical studies. Conventional rigid statistical methods in analytical chemistry often fail to address the specifics of studying fuzzy experimental data, such as series of impurity concentration values in a river flow over space and time. In this context, alternative soft computing tools, particularly those based on neuro fuzzy hybrid algorithmic structures related to the ANFIS architecture, are more suitable. An analysis of chemicoanalytical data arrays for copper and zinc in the Volga River, considering water flow at various distances from the shore and depths, revealed a complex oscillatory behavior in the concentrations of both substances. This analysis concluded that the neuro-fuzzy processing scheme of the monitoring results enables a more in-depth study of the poorly understood processes of hydrochemical dynamics in systems far from thermodynamic equilibria, such as natural watercourses.