|
| Construction of detection models for p-anisidine value of soybean oil and sunflower seed oil based on low-field nuclear magnetic resonance |
| |
| DOI：10.19902/j.cnki.zgyz.1003-7969.240266 |
| KeyWord:low-field nuclear magnetic resonance vegetable oil p-anisidine value model |
| FundProject:河南省科技攻关项目（242102320278）；河南省A类专业创建建设专项（HN-HautFood-100）；河南工业大学“双一流”本科生科技创新能力提升专项项目（HN-HautFood IAEG-012） |
|
| Hits: 8 |
| Download times: 2 |
| Abstract: |
| In order to achieve the rapid detection of oxidative quality of vegetable oils, taking soybean oil and sunflower seed oil as the research objects, the changing rules of low-field nuclear magnetic resonance(LF-NMR) relaxation signals during the oxidation process were systematically analyzed, and the correlations between the p-anisidine value (p-AV) and relaxation characteristic indicators were studied. The calibration models for p-AV of soybean oil, sunflower seed oil and total samples (soybean oil + sunflower seed oil) were constructed based on LF-NMR, and the effects of modeling data, preprocessing methods and modeling method on p-AV model performance were examined. The results showed that the p-AV of soybean oil and sunflower seed oil gradually increased, while the relaxation time gradually decreased during the heating process. The order of peak area variation was S23>S22>S21, and there was a significant correlation between the p-AV and the relaxation characteristic indicators. The optimal modeling conditions for the p-AV models of soybean oil, sunflower seed oil and total samples were as follows: modeling data 1-1 000, preprocessing method orthogonal signal corretion(OSC), none and OSC, and modeling method partial least square (PLS). Under the optimal modeling conditions, the p-AV model prediction precision for soybean oil and sunflower seed oil were 3.448 and 2.572 respectively, while those were 4.523 and 4.437 using the total sample model. In conclusion, it is feasible to detect the p-AV of vegetable oils based on LF-NMR, and the prediction precision of the single vegetable oil model is higher, while the total sample model of two vegetable oils has better universality. |
| View full text View/Add Comment Download reader |
| Close |
|
|
|
