A Novel Engine Oil Health Classification using Physical Properties and DWKNN Algorithm

Engine oil plays a vital role in maintaining the engine's overall health. It acts as a lubricant and reduces friction between the engine components to minimize wear and tear, thus increasing its lifespan. Oil is changed during scheduled vehicle service without having a proper quantitative estimate of its life. This leads to wastage of engine oil as its health is misjudged. This paper focuses on enabling efficient usage of the engine oil by monitoring its health with the help of oil's physical properties. Conductivity and transparency are the parameters taken into consideration, and the overall health of the oil is predicted using the machine learning algorithm, DWKNN (Distance Weighted k - Nearest Neighbor) Classifier. After testing with various samples of engine oil, the model has obtained an accuracy of 95%. A data acquisition unit using sensors has been developed with the help of an Arduino Nano to facilitate data collection.