PubAg

Main content area

Artificial Intelligence and fourier-transform infrared spectroscopy for evaluating water-mediated degradation of lubricant oils

Author:
Chimeno-Trinchet, Christian, Murru, Clarissa, Díaz-García, Marta Elena, Fernández-González, Alfonso, Badía-Laíño, Rosana
Source:
Talanta 2020 v.219 pp. 121312
ISSN:
0039-9140
Subject:
Fourier transform infrared spectroscopy, artificial intelligence, discriminant analysis, lubricants, neural networks, oils, reflectance, reflectance spectroscopy, viscosity
Abstract:
The presence of water in lubricant oils is a parameter related to the lubricant deterioration, which can be indicative of a serious loss of tribological efficiency and, therefore, an increase in maintenance costs. Likewise, controlling the aging of the lubricant oil is a keynote issue to prevent damage on the lubricated surfaces (e.g. engine pieces). The combination of Attenuated Total Reflectance (ATR) techniques with Fourier-Transform Infrared Spectrometry (FTIR) result in an easy, simple, fast and non-destructive way for obtaining accurate information about the actual situation of a lubricant oil. The analysis of this ATR-FTIR information using Artificial Neural Networks (ANN) as well as Linear Discriminant Analysis (LDA) results in the proper classification of lubricant oils regarding the presence/absence of water, age and viscosity. The methodology proposed in this work describes procedures for identifying the deterioration degree of oils with as high as 100% success (aging week) or 97.7% (for viscosity and water presence).
Agid:
7009142