Spectroscopic Quantitative Analysis For Biological Wastewater Treatment Reactors

The biological reactor for wastewater treatment is a complex system because of the presence of various types of pollutants and microorganisms.Conventional chemical analytical methods can not to fulfil the requirement of real-time monitoring of bioreactors.However,spectral analysis technique,as a rapid,simple,non-invasive and reagent-free method,is recognized as one of the promising methods for on-line monitoring bioreactors.In this thesis the utilization of three spectral analysis methods for the determination of the components in bioreactor effluents,extracellular polymeric substances（EPS） and the poly（β-hydroxybutyrate）（PHB） inside activated sludge have been proposed.The three methods included the resonance light scattering（RLS）, the near-infrared spectroscopy and the mid-infrared spectroscopy.Chemometric methods were also used for the data analysis.The main contents and results of this work are as follows:1.The RLS method is based on the interactions between biomacromolecules and dyes,which cause a substantial increase in the resonance scattering signal of dyes. Two dyes,Congo red and Neutral red,were respectively used as the RLS probes for determining proteins and carbohydrates in the bioreactor effluents.Compared with the conventional colorimetric methods,the proposed methods were satisfactory applied to the measurement of proteins and carbohydrates in the effluents from ten bioreactors for wastewater treatments,and high sensitivitiese were also achieved.Results imply that the RLS method is a sensitive,selective method for the monitoring the quality of effluents from bioreactors.2.Fourier transform near-infrared（FT-NIR） technique was used for monitoring anaerobic fermentative reactors.Orthogonal signal correction（OSC） method was employed as NIR spectral preprocessing options.Calibration models were established and the validation of the method was performed with the sucrose,ethanol and volatile fatty acids（VFAs） contents determined by the anthrone and gas chromatography methods,respectively.Spectral range and the number of internal latent variables were optimized for the best correlation coefficient.The results show that the NIR spectroscopy technique was able to rapidly quantify the contents of both VFAs and sucrose in the anaerobic reactors.3.A quantifying method for measuring the concentrations of VFAs and ethanol in the effluent of an anaerobic H₂-producing bioreactor was proposed and validated using the FT-NIR spectroscopy.The first-derivative spectra calculated by a simple numerical difference,combined with the orthogonal signal correction method,were used as spectral preprocessing options.A calibration model was established and validated using gas chromatography measurement results.The number of internal latent variables was optimized based on the lowest root-mean-square error of calibration.The calibration model established shows the satisfactory results for the lowest root-mean-square errors of prediction compared to other preprocess methods. The method developed in this work was demonstrated to be more flexible,compared with other approaches to determine the effluents VFA and ethanol concentrations.4.The FT-NIR spectroscopy was also used for the determination of the substrate concentration and chemical oxygen demand（COD） in an aerobic bioreactor. The continuous wavelet transform method was used for reducing the influence of background and noise in NIR spectra.The internal partial least square method was then used for optimizing the spectra range and establishing the calibration model of substrate concentration and COD.The spectra ware smoothed and the latent variables of the calibration model were reduced for better prediction results with the integrated method.The NIR method could be used as a new analytical means in the determing substrate concentration and COD of aerobic bioreactors.5.Fourier transform infrared（FTIR） spectroscopy in combination with two data processing methods was used for quantitative analysis of the PHB contents of activated sludge.Because of the slight variation of protein content in cells compared with the variation of PHB content,the ratio of the PHB special absorbances in 1726 cm^-1 to protein special absorbances 1654 cm^-1 was calculated to establish the calibration curve of the PHB content.In another try the partial least squares（PLS） multivariative statistical technique was used for a quantitative analysis of sludge IR spectra.The results of both methods matched those of the gas chromatography method.For the FTIR technique,there are no solvent requirements,sample preparation is minimal and simple,and analysis time is greatly reduced.The results demonstrate the potential of FTIR spectroscopy as an alternative to the conventional methods used for analysis of PHB in activated sludge.6.An integrated method of the consensus least squares support vector regression combined net analysis signal was used for analyzing the NIR spectra data of the soluble microbial products（SMP） and EPS of sludge samples.The calibration models of measing the protein,carbohydrate and humic acid contents in the sludge samples were established.The prediction results of the calibration models were satisfied.The method developed in this work was demonstrated to be more flexible than other approaches.