The stability of highly consumed vegetable refined oils after discontinuous frying of potatoes was compared. Both those vegetable oils containing additives and those that did not were considered. Vegetable oil samples were evaluated using refractive index, anisidine and peroxide values, UV absorbance and dielectric constant-based determination of the content of total polar compounds. Chemical changes caused over the frying time were monitored and multivariate modelling of the data was carried out. A new gas chromatographic-mass spectroscopy method was intended to record a fingerprint of both polar and non-polar compound fractions. Multivariate models of chromatographic fingerprints were also developed, and the results obtained from both approaches were verified to be statistically similar. In addition, multivariate modelling also allows to differentiate among vegetable oils according to oxidation performance. Indeed, it was initially observed that olive oils presented the highest natural thermo-oxidative stability compared to other seed oils, although it should be noted that these differences were not significant when regarding olive pomace oils and seed oils containing synthetic additives.

Most physicochemical and sensory properties of edible vegetable oils are not stable over time. One of the main causes of quality depletion of vegetable oils is oxidation, which influences sensory acceptability and nutritional value, and could even lead to toxic compounds. That negative influence affects international refined oil prices and the variety of its culinary applications. Modelling quality depletion of vegetable oils and establishing the shelf life, generally accepted as the time until rancidity becomes evident, already remains a challenge for the industry. Hence, this paper will show a promising chemofoodmetric methodology, as an easy and straightforward tool to estimate the current shelf-life of refined vegetable oils, based on a comprehensive characterisation of quality depletion-related changes over storage time under real market conditions. The methodology for building a multivariate kinetic ageing-based model is described, taking into account all time-related physicochemical parameters and chemometric processing tools. From a particular ageing state, multiparametric models are able to reliably infer the expected storage time for each vegetable oil so that it remains consistent with acceptability requirements. The results of the study pointed out the accuracy of multivariate shelf-life modelling with regard to univariate modelling. Discrepancies were found in the oxidation rates of oils extracted from different plant seeds.