Temperature Compensation of Lubricating Oil Impedance Based on Random Forest Regression

Electrochemical impedance spectroscopy (EIS) is an important method for equipment health monitoring to analyze the deterioration of lubricating oil quality. However, the characteristic value of the AC impedance of the oil is affected by the operating temperature and there is a lack of theoretical correlation model, so a method is proposed to compensate the effect to AC impedance by temperature in the process of deterioration of the lubricating oil quality of gas engine based on random forest regression. Firstly, the oil temperature and the corresponding impedance changes during the actual operation of the engine was analyzed. Based on the mode of oil operating temperature and its corresponding impedance everyday, the deviation of oil temperature (Δ<tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$T$</tex>) and and impedance mode (ΔIm) from the corresponding benchmark value at each sampling point was calculated separately. Then, the correlation model between ΔT and ΔIm was established based on random forest regression algorithm. After temperature compensation was carried out on the all original impedance data with the reference temperature, the corresponding equivalent impedance values of the original test data at all times are obtained. Finally, the equivalent impedance data at the reference temperature is used to analyze and predict the oil deterioration trend.