The prediction of indices at infall of confluent flow network of wastewater with Multivariate Linear Regression

Author

Gao Xiang ; Bai Lina

Author_Institution

Autom. Sch., Yantai Nanshan Univ., Yantai, China

fYear

2010

fDate

29-31 July 2010

Firstpage

5125

Lastpage

5129

Abstract

On the pre-estimation of indices of wastewater at the infall of Waste Water Treatment Plant (WWTP) in the confluent flow dynamic system,Multivariate Linear Regression (MLR) is presented and applied effectively in general case. However, the Partial Regression Coefficients (PRC) identified from the data model is not suitable for the continuous time varying process occasionally. Therefore, an improved Recursive Forgetting Factor Least Square (RFF-LS) algorithm for the regression of MLR is developed and implemented on an example of certain confluence flow system with a WWTP and relevant sources of wastewater to refresh the PRC adaptively now, which demonstrates better effect on predictive precision on indices of both Chemical Oxygen Demand (COD) concentration and water quantity.

Keywords

continuous time systems; data models; least squares approximations; regression analysis; time-varying systems; wastewater treatment; chemical oxygen demand concentration; confluent flow dynamic system; continuous time varying process; data model; multivariate linear regression; partial regression coefficients; recursive forgetting factor least square algorithm; waste water treatment plant; water quantity; Estimation; Fitting; Heuristic algorithms; Indexes; Prediction algorithms; Wastewater; Wastewater treatment; COD; Confluent Flow; Multivariate Linear Regression; Partial Regression Coefficients; Recursive Forgetting Factor Least Square; Wastewater Treatment;

fLanguage

English

Publisher

ieee

Conference_Titel

Control Conference (CCC), 2010 29th Chinese

Conference_Location

Beijing

Print_ISBN

978-1-4244-6263-6

Type

conf

Filename

5573161