A NOVEL QUANTITATIVE ANALYSIS METHOD OF THREE-DIMENSIONAL FLUORESCENCE SPECTRA FOR VEGETABLE OILS CONTENTS IN EDIBLE BLEND OIL

摘要

Edible blend oil is a mixture of vegetable oils. Eligible blend oil can meet the daily need of two essential fatty acids for human to achieve the balanced nutrition. Each vegetable oil has its different composition, so vege-table oils contents in edible blend oil determine nutritional components in blend oil. A high-precision quan-titative analysis method to detect the vegetable oils contents in blend oil is necessary to ensure balanced nu-trition for human being. Three-dimensional fluorescence technique is high selectivity, high sensitivity, and high-efficiency. Efficiency extraction and full use of information in three-dimensional fluorescence spectra will improve the accuracy of the measurement. A novel quantitative analysis is proposed based on Quasi-Monte-Carlo integral to improve the measurement sensitivity and reduce the random noise. Partial least squares method is used to solve nonlinear equations to avoid the effect of multicollinearity. The recovery rates of blend oil mixed by peanut oil, soybean oil and sunflower are calculated to verify the accuracy of the meth-od. It is increased, compared the linear method used commonly for component concentration measurement.