Near infrared spectrometry of humic acid content in fertilizers at different levels of granularity

3rd International Conference on Energy Engineering and Environmental Protection

能源学;生态环境科学

Gong, Xue^1 ; Li, Yuhuan^1 ; Wang, Yue^1 ; Wang, Ruiyan^1 ; Yuan, Xiujie^1

Shandong Agricultural University, China^1

关键词: Chemical enterprise;    Classical least squares;    Correlation coefficient;    Fertilizer industry;    Hyperspectral models;    Near infrared spectrometry;    Near-infrared absorption;    Prediction precision;   

Others  :  https://iopscience.iop.org/article/10.1088/1755-1315/227/5/052055/pdf
DOI  :  10.1088/1755-1315/227/5/052055

学科分类：环境科学（综合）

来源: IOP

【 摘 要 】

ObjectivesThe adaptability of generally sampling tests and traditional method of chemometrics has trended to be more and more fragile in modern chemical enterprises which products chemical components required to be quickly measured with higher precision. Therefore it is a challenge to quickly and exactly calculate the humic acid contents in fertilizer industry while giving up traditional methodology. But it is possible currently that the combination of modern near-infrared spectroscopy and stoichiometry will become an effective means. MethodsThere were 74 samples randomly selected from humic acid raw materials, and strictly grading in four classes for each sample as 0.55mm, 0.25mm, 0.18mm, 0.15mm and exactly determining the humic acid content In light of the national criterion (GB/T11957-2001), and precisely surveying the near-infrared absorption spectra (NIRAS) data for 296 samples with the NicoletTMiNTM10 ft-ir spectrometer. So that the mathematical model was established using the 64 by Classical Least Square method (CLS), and the other 10 were verified. ResultsThe result of the model verification shows that hyperspctral model has a good predictive effect on the fertilizer humic acid content. The model's coefficients of determination are all above 0.75. The relative errors are all below 0.35. The lowest RMSEP is 1.15. And the correlation coefficient can reach 0.812. ConclusionsFinally the results show that the use of hyperspectral model can predict the humic acid content in humic acid fertilizers effectively. The powdery humic acid determination has overall good prediction effect based on CLS and NIRAS data, and the prediction precision is influenced by particle size; the finer grain diameter is, the higher the precision of the model. This study will provide the technical support for quick sampling of humic acid raw materials.