基于改进偏最小二乘法的近红外快速分析强化生物除磷污泥胞内PHA

聚β羟基烷酸酯(PHA)是强化生物除磷系统中颗粒污泥胞内重要碳源和能源,其快速测定对强化生物除磷机理研究有重要意义. 采用Savitzky-Golay平滑法(SG)-多元散射校正法(MSC)对污泥的近红外光谱进行预处理,通过改进的偏最小二乘法(iPLS、siPLS、biPLS)建立污泥样品近红外光谱与PHA含量的定量分析模型.结果表明,SG-MSC预处理减弱噪声和背景等表面因素对光谱的影响,采用联合区间偏最小二乘法(siPLS)将全光谱等分为30个子区间, 联合子区间13 21 24 29]建立的模型预测效果最优,其交互验证均方根误差(RMSECV)和预测均方根误差(RMSEP)分别为0.2018和0.3120,校正集和预测集相关系数分别达到0.9925和0.9391,该光谱区段与PHA分子结构中C-H的伸缩变形、弯曲振动和C=O的伸缩振动密切相关. 改进偏最小二乘有效地优化光谱建模区域,提高模型预测能力,实现污泥胞内PHA含量的快速定量分析.

Rapid determination of the intracellular PHA content during the enhanced biological phosphorus removal using NIR spectroscopy based on the modified PLS methods

Poly-β-hydroxyalkanoates (PHA) are vital intracellular storage generated by granular sludge in enhanced biological phosphorus removal (EBPR) reactor, served as carbon and energy source. The rapid determination of PHA is of great significance to the further researches on mechanism of EBPR. Savitzky-Golay smoothing (SG) and Multiplicative scatter correction (MSC) method were jointly used to preprocess the near infrared (NIR) spectra of the sludge samples, the quantitative analysis models between the NIR spectra and the PHA contents were established with modified PLS methods such as interval partial least squares (iPLS), backward interval partial least squares (biPLS) and synergy interval partial least squares (siPLS). SG-MSC pretreatment could effectively diminish the effect of spectrum interference factors. The optimal model was established by four sub-intervals 13,21,24,29] with siPLS method, divided the preprocessed spectrum into 30sub-intervals. The root mean square errors of cross validation (RMSECV) and the root mean square errors of prediction (RMSEP) were 0.2018 and 0.3120 respectively. The correlation coefficients of the calibration and prediction sets were 0.9925 and 0.9391 respectively. The best analysis spectrum band are closely related to the C—H bending vibration and C=O stretching vibration of PHA molecular. The modified PLS could not only optimize the spectra region, enhance the predictive ability of the models, but also realize the rapid and quantitative determination of intracellular PHA content in the granular sludge.