Detection oxidative degradation in lubricating oil under storage conditions using digital images and chemometrics

Abstract Early detection of the degradation process of lubricating oil caused by storage condition can prevent extensive losses for rural producers and mining companies, which operate far from urban centers and need to purchase large volumes of lubricating oil. This work concerns the development of an approach combining computer vision-based analytical methods and chemometrics tool for lubricating oil oxidation detection. This is presented here for the first time to the best of our best knowledge. Samples of the non-oxidized lubricating oil acquired in the local market were analyzed by ATR-FTIR and digital images (DI). Subsequently, the samples were subjected to heating and exposure to radiation to force the degradation of the samples and simulate the process that occurs during storage. Then, they were analyzed in the same way as non-oxidized oil. These conditions were able to promote the degradation of the samples, with information being accessed by the intensity increase on the 720 cm −1 band. In addition, PCA revealed that the DI could be used to differentiate non-oxidized samples from those that had suffered degradation at some level. A model based on PLS-DA, proved effective in the classification of these samples, with success in all cases. This new strategy will allow the fast and in situ acquisition of the results, without the need to analyze the samples in a distant lab. Besides, the proposed method is in line with green chemistry as it does not use sample handling and does not generate residues.