Rapid and nondestructive quality analysis of thermally oxidized rapeseed oil based on ultrasonic diagnostic technology: A study on temperature compensation mechanisms.

This study applies ultrasonic detection technology combined with temperature compensation models to evaluate the quality of thermally oxidized rapeseed oil, which was continuously heated at 180 °C for 72 h. The results indicate that the effects of heating treatment and increased detection temperature on ultrasonic indicators exhibit opposite trends, with the impact of detection temperature on ultrasonic signals being significantly higher than that of thermal treatment. The total polar component prediction model, which integrated the random forest algorithm with the global temperature compensation model, exhibited the best predictive performance, with external validation results of R = 0.906, RMSE = 1.402, and MAE = 1.007. These findings confirm that the temperature compensation mechanism effectively mitigates the temperature effects during detection. This proposed temperature compensation mechanism exhibits significant potential for application in the edible oil production industry, aiming to achieve online, rapid, and non-destructive monitoring of oil quality.