• Spectroscopy and Spectral Analysis
  • Vol. 37, Issue 8, 2486 (2017)
XU Hong1, MING Dong1, LI Ying-xin2, QI Hong-zhi1, SUN Mei-xiu2, and YANG Ji-chun2
Author Affiliations
  • 1[in Chinese]
  • 2[in Chinese]
  • show less
    DOI: 10.3964/j.issn.1000-0593(2017)08-2486-07 Cite this Article
    XU Hong, MING Dong, LI Ying-xin, QI Hong-zhi, SUN Mei-xiu, YANG Ji-chun. Research and Application of Linear Mixing Model of Two-Component Mixed Edible Oil[J]. Spectroscopy and Spectral Analysis, 2017, 37(8): 2486 Copy Citation Text show less

    Abstract

    A linear mixing model of the two-component mixed edible oil was proposed based on the assumption that the absorption coefficient of the two-component mixed edible oil can be obtained by the linear combination of absorption coefficients of two raw oils according to the mixing proportion. Based on the Lambert-Beer’s Law, a series of absorption equations of the incident light was constructed to describe variations of light intensity after the same source light respectively passing through the two raw oils and their proportional mixture with the same thickness. As a result, a formula was derived to calculate the mixing proportion of the two-component mixed edible oil, which was a new method to quantitatively analyze the proportion of adulterated edible oil. According to the theory of errors, the calculation difference of mixing proportion to the actual value was analyzed by the total differential methods, and the result demonstrated that there was an optimal wavelength where the products of the absorption light intensity of two raw oils and the difference value between their absorbance were larger, and less calculation difference could be obtained. A VIS/NIR spectral detection system was assembled and used to validate the mixing proportions of corn oil adulterated in peanut oil, soybean oil adulterated in peanut oil and soybean oil adulterated in corn oil. The results showed that the principle of choosing the optimal testing wavelength was correct, and the relative errors between calculation values and actual values were less than 5% for adulterating proportions above 10%(v/v), as well as the correlation coefficients were 0.999 4, 0.999 7 and 0.999 3 with RMS errors of 0.006 9, 0.005 1 and 0.007 6. The results proved that the proposed linear mixing model was practicable. The comparative experiments, which were operated on the stacked series of two separate oil samples with same mixing proportions, also verified the correctness of the linear mixing model and the method for choosing the testing wavelength. In addition, these comparative experiments revealed that the parallelism of light source and verticality of oil vessels were important to the detection accuracy of mixing proportion. This paper introduces a new method of detecting the mixing proportion of two-component mixed edible oil which differs from the traditional method of spectroscopy combined with chemometrics. In this method, the mixing proportion of two-component mixed edible oil can be precisely computed only after three times data collection of absorption spectra of the two raw oils and their mixture.
    XU Hong, MING Dong, LI Ying-xin, QI Hong-zhi, SUN Mei-xiu, YANG Ji-chun. Research and Application of Linear Mixing Model of Two-Component Mixed Edible Oil[J]. Spectroscopy and Spectral Analysis, 2017, 37(8): 2486
    Download Citation