Rapid Determination of the Total Phosphorus and the Nitrate Nitrogen in Denitrifying Phosphorus Removal with iPLS and Near Infrared Spectroscopy

Abstract

Denitrifying phosphorus removal was realized in sequencing batch reactor with alternate anaerobicanoxic procedures. Quantitative analysis models of the total phosphorus and nitrate nitrogen were established with near infrared spectroscopy combined with interval partial least-squares (iPLS). Second derivative, multiple scatter correction and wavelet denoising were respectively used to pretreat the raw near infrared spectral data. The results show that the iPLS model established by wavelet denoising method was the optimal one. As to the interval partial least-squares calibration models (iPLS models) of the total phosphorus and nitrate nitrogen, the results showed that the coefficients (r c ) of the correlation between the actual values and corrected values were respectively 0.9449 and 0.8681, with the root mean square errors of cross validation (RMSECV) being 0.0254 and 0.0315. In addition, the testing results of the interval partial least-squares calibration models of the total phosphorus and nitrate nitrogen indicated that the coefficients (r p ) of correlation between the actual values and predicted values were 0.9294 and 0.9014, with the root mean square errors of prediction (RMSEP) being 0.0235 and 0.0298. This study suggested that the iPLS models of the total phosphorus and nitrate nitrogen may provide a method for synchronous detection of the total phosphorus and nitrate nitrogen during denitrifying phosphorus removal.