基于激光诱导击穿光谱和化学计量学的水体COD实时测量方法

发布时间：2018-03-26 04:06

本文选题：COD　切入点：LIBS　出处：《光谱学与光谱分析》2017年11期

【摘要】：光谱学传感技术测量水体COD是现代环境监测发展的趋势,相比于传统的化学分析方法具有在线连续检测的突出优势,适合对环境水体COD的实时监测。采集实际水样,利用激光诱导击穿光谱技术(LIBS)获取水样的光谱数据。通过不同的光谱预处理方法结合偏最小二乘法建立水样的COD定量预测模型,对水体COD的LIBS光谱测量方法的定量预测及相关模型参数进行分析。发现用基线校正叠加S-G求导预处理后的光谱建立的PLS模型得到最佳预测效果,校正集集R2为0.9958,预测集R2为0.975 3,RMSEC为4.438 7,RMSEP为9.733 9。通过实验结果分析表明光谱传感技术可用于环境实际水体COD的定量预测分析,为开发便携式水体检测设备奠定了理论基础。
[Abstract]:Spectroscopic sensing technology is the trend of modern environmental monitoring and development. Compared with the traditional chemical analysis method, it has outstanding advantages of on-line continuous detection, and is suitable for real-time monitoring of environmental water COD. The spectral data of water samples were obtained by laser induced breakdown spectroscopy (LIBS). The COD quantitative prediction model of water samples was established by using different spectral pretreatment methods and partial least square method. The quantitative prediction of LIBS spectrum measurement method of water body COD and the analysis of related model parameters are carried out. It is found that the best prediction effect is obtained by using the PLS model established by the pre-processing of S-G derivative preprocessing with baseline correction superposition. The corrected set R2 is 0.9958, and the prediction set R2 is 0.975 3rMSEC. The experimental results show that the spectral sensing technique can be used for quantitative prediction and analysis of COD in environmental water bodies, which lays a theoretical foundation for the development of portable water detection equipment.
【作者单位】：桂林电子科技大学;北京农业智能装备技术研究中心北京市农林科学院;
【基金】：国家(863)计划项目(2013AA10230202) 国家自然科学基金重点项目(31622040)资助
【分类号】：O657.3;X832

【相似文献】