Regressional models that describe oil absolute viscosity

Equations that describe the temperature dependence (298–338°K) of absolute viscosity (μ) of 21 oils and oil‐liquid fat mixtures were obtained based on two different approaches. Fitting each particular viscosity profile to a quadratic extension of the Andrade equation provided the best predictive models (R2>0.96). However, the coefficients associated with temperature effect did not have a physical‐chemical meaning. In contrast, the multiple variable regressional approach fitted, in just one equation, the μ of all 21 oil systems (R2≈0.93). This equation included terms associated with structural parameters of acylglycerides, namely the degree of unsaturation (i.e., iodine value) and chainlength (i.e., saponification value) of the fatty acids. The models described effects of thecis double bonds and fatty acid chainlength on the acylglycerides’ interactions that determine both the μ of the system and its capability to crystallize. Therefore, multiple variable regressional analysis might be an excellent tool to better understand the quantitative structure‐functional property relationships in lipids systems.