Determination of Compatibility Parameters of Crude Oils by Near-Infrared Spectroscopy

The prediction of crude oil blend properties is an essential part of the purchasing and planning process of refineries. One of the most important properties to predict is the compatibility of crude oils, as this parameter is associated with serious upsets in refinery systems. Conventional analyses to determine compatibility are time-consuming, and therefore, they are difficult to establish as routine analytical tests on refineries and crude oil distribution and storage sites. In this work, the feasibility of using near-infrared (NIR) spectroscopic measurements for determining compatibility was evaluated. The evaluation of compatibility is based on the creation of two models to determine the Heithaus parameters of the crude oils: solvent power (Po) and solvent requirement (Ra). Based on a set of more than 100 crude oils, it was found that the optimal number of variables for Po and Ra was 5 and 6, respectively. Several statistical tests indicated that the models produced reasonably good values for both parameters. Furthermore, it was demonstrated that the application of blending rules to determine P-values of crude oil blends based on the NIR-predicted values yields lower error margin compared to conventional methods (±0.07 vs ±0.30). These results are promising as NIR methods are fast, reliable, and can be used online for plant control.