Detection of Viscosity of Engine Oil using IoT Technology

Abstract

Engine lubricants gradually deteriorate when exposed to extreme thermal conditions, oxidation, and contamination, resulting in viscosity shifts and reduced lubrication effectiveness. To address this challenge, this work presents an IoT-based framework that continuously evaluates oil condition and predicts the distance that can be safely traveled before replacement is required. The system integrates three key modules: (i) a low-cost ESP32 sensing unit that monitors oil density and temperature, (ii) a lightweight algorithm that transforms these parameters into a normalized Thickness Index (TI) to reflect viscosity behavior, and (iii) a cloud-hosted predictive model that estimates remaining useful life in kilometers (RUL-km) using TI trends combined with historical driving information. Experimental calibration links density variations to viscosity characteristics, while machine learning models are trained for prognostics. Initial validation indicates that the approach can classify viscosity grades and forecast RUL-km with a mean absolute error of less than 10%, enabling practical condition-based maintenance and enhancing vehicle reliability.