engleski

Multivariate statistical methods for characterization of pollution indicators of municipal wastewater in split

In this paper a set of quantitative descriptive and analytical data collected after the primary treatment of municipal wastewater wider area of Split has been processed by multivariate statistical methods in other to characterization of pollution indicators. Monitoring daily measurement of 13 parameters was performed on three locations in the region of Split during 2006-2009. The obtained results show and explain the value movement of pollution indicators in relation to the maximum permissible concentration prescribed in the Regulations as well as some significant correlation of individual variables for a specified period. For handling the results different chemometrics methods were employed, such as basic statistical methods, correlation coefficients, the principal component analysis (PCA) and cluster analysis (CA). Based on the obtained experimental data statistical parameters were determined and multivariate linear regression models: multivariate linear model (MLR) and local linear regression model (PLR) with 9 variables were developed. The study allows drawing new information from the data sets such as patterns of similarity between sampling locations, sources of pollutions in the environment, seasonal behavior of pollution indicators and time trends. The implementation of these analysis required the proceedings in the preliminary study for the selection of treatment technologies and making the project of device and a precondition for setting up the complex mathematical models that will enable an effective control and optimize the process of disposal and wastewater treatment

multivariate linear model; principal component analysis; pollution indicators; municipal wastewater

nije evidentirano