Diagnosis and prediction of engine oil replacement time with the help of odor, color data and integration of color, odor and Brix data

Lubricating parts to reduce friction and wear is one of the most important functions of engine oil. When engine oil is used, the color and viscosity of the engine oil changes during the operation of the car, and due to the increase in friction and the energy required to pump the oil, it causes an increase in fuel consumption. The purpose of this study is to investigate the detection of engine oil life based on the distance traveled with the help of smell, color and integrated data of color, smell and brix using two standard and zero-one methods. In this research, electronic nose devices, refractometer and colorimeter were used. Principal component analysis (PCA), linear discriminant analysis (LDA) and artificial neural network (ANN) were used to classify the data to detect kilometers, and partial least squares (PLS) and principal component regression (PCR) were used to predict the parameters. Brix and engine oil color changes were used. The main component analysis method of the results showed that in the score chart, engine oil life detection was done better based on color, and all the oil samples were well separated based on the distance traveled. Also, the LDA method for detecting the life of engine oil with different traveled kilometers for color data separated different classes with 96.36% accuracy. Based on the disturbance matrix obtained from the artificial neural network for the color data, the classification accuracy of engine oil with different distances traveled was 93.6%. LDA method showed better classification than PCA and ANN methods. The results showed that the PLS and PCR methods performed well in predicting Brix parameters and engine oil color change, but they did not perform well in predicting the mileage parameter. Using PCR and PLS models are more suitable for Brix and color change detection.