System Analysis of Source Water Indicators at Drinking Water Production in the Central Water Supply System

Abstract

The work is devoted to the system analysis of the given indicators of source water at the production of drinking water in the central water supply system of a large settlement. Today, within increasing negative anthropogenic impact on the environment, there is a deterioration in the state of many sources of drinking water supply in a wide range of indicators, in particular, such as organoleptic properties of water. As a consequence, there is a problem for drinking water. The paper presents the process of preparing data on the parameters of the source water taken from the reservoir, which were taken into account monthly at the enterprise (from 2002 to 2014) at the water deodorization plant. The above parameters have a significant effect on the organoleptic properties of the final water. Preparation of data for analysis is carried out by K. Pearson's principal components analysis. The data obtained in the space R9 is transferred to a space of lower dimension R3. Dimension reduction allows to reduce autocorrelation between components. Selection of components in the space R3 is carried out according to the Pareto rule. Clustering is carried out in the space R3 by the method of spherical clustering of the "Trout" data with a constant clustering radius. A step-by-step visual representation of the clustering algorithm in the space R3 is presented. It is shown that there are clusters in these indicators of the quality of the initial water. Correlation-regression analysis of the data presented in the main components is carried out. Regression dependences of the indicators of organoleptic properties on the main components from the space R3 are shown.