Analytical and chemometric applications in the study of automotive and related lubricant degradation

Chemometric techniques have been utilised for the study of automotive lubricant oil degradation. The initial investigations were performed by analysis of data from top ring zone engine test datasets. Principal component analysis (peA) was used to explore the ring zone data. The difference in the performance of various lubricant formulations sampled from the ring zone region of operating Petter AA-I diesel and Petter W-1 petrol engines was established by the partial least squares discriminant model (PLS-DA). The results from the study of the test engine data provided optimised insight into the break down of the chemical/physical parameters of the lubricants during operative conditions. This work proceeded onto condition monitoring techniques. Over a hundred used oil samples were obtained from the sump of various petrol and diesel engine vehicles, in addition fresh oil samples were also collected. Series sets of used oil samples were acquired by periodic sampling from a Honda 1.8 L petrol engine, a Peugeot 1.9 L diesel engine and a diesel engine sump test. Following sample acquisition, each oil sample was analysed using FTIRIATR and conductimetric titrations were performed. These analytical equipments are used to monitor and assess the extent of degradation. A novel model was developed to enhance the IP 400 conductimetric titration method of measuring base number of new and used lubricants. This nonlinear least-squares model was integrated into the titration programme along with two linear least-squares curve fitting methods. The models were effectively used to estimate the titration endpoint which was subsequently used in base number calculation. The results demonstrated the robustness of the three endpoint estimation methods and indicate reliability of the titration equipment and programme. peA was used to analyse the FTIR spectra data of the oil samples. peA performed on different sets of pre-processed data uncovered objective information on the condition of the lubricants. peA models of the series set oil samples highlighted difference between samples as a result of progressive degradation. The effect of adding fresh oil (top-up) to the engine was apparent. This work has demonstrated the importance and efficiency of mathematical/statistical models in lubricant oil engine testing and condition monitoring.