摘要
采用傅里叶中红外光谱对强化生物除磷过程中污泥胞内糖原物质进行了表征,并将污泥样品与糖原标样的红外光谱进行对比,确认糖原物质在1020 cm-1峰附近有较明显的特征吸收贡献.因此,采用红外光谱945~1150 cm-1区域内的吸收光谱数据,结合测得胞内糖原的含量,应用偏最小二乘法和BP神经网络算法分别建立了污泥红外光谱与糖原含量的定量分析模型.结果显示,偏最小二乘法得到的测量值与预测值具有较小的平均绝对误差(MAE),其相关系数达到0.921.
The intracellular glycogen in the sludge from the enhanced biological phosphorus removal( EBPR) process was characterized by Fourier Transform Infrared( FTIR) spectroscopy. Compared with the FTIR spectrum of glycogen standard,it was confirmed that the peak near 1020 cm- 1was attributed to the intracellular glycogen. Accordingly,a simple and effective quantitative method for glycogen content measurement was established using the FTIR spectroscopy with partial least squares( PLS) method and BP neural network algorithm and the infrared spectra ranging from 945 ~ 1150 cm- 1.Results showed that a lower mean absolute error( MAE) could be achieved with PLS method between the measured and predicted glycogen values,and the correlation coefficient was 0. 921,implying the validity of the proposed method for the analysis of the intracellular glycogen content in sludge.
出处
《环境科学学报》
CAS
CSCD
北大核心
2015年第3期705-712,共8页
Acta Scientiae Circumstantiae
基金
国家自然科学基金(No.51378017
21007001)
安徽省高校省级优秀青年人才基金重点项目(No.2013SQRL100ZD)~~
关键词
中红外光谱
强化生物除磷
糖原
偏最小二乘
BP神经网络
Fourier Transform Infrared spectroscopy(FTIR)
enhanced biological phosphorus removal(EBPR)
glycogen
partial least squares
BP neural network algorithm
