Identification and classification of adulteration in some fossil fuel products using an electronic nose

Abstract

The objective of this study was to identify and classify adulteration in certain fossil fuel products (including gasoline, diesel, and kerosene) using an electronic nose (e-nose). In this research, blends of gasoline-diesel, gasoline-kerosene, and diesel-kerosene were prepared at volumetric ratios of 5%, 10%, 15%, 20%, 25%, and 30%. Data acquisition was performed using an e-nose system equipped with 10 metal oxide semiconductor (MOS) sensors. The collected data were analyzed using various methods, including Linear Discriminant Analysis (LDA), Quadratic Discriminant Analysis (QDA), Support Vector Machine (SVM), and Artificial Neural Network (ANN). Among the sensors, MQ135, TGS2611, and TGS2610 demonstrated the highest performance. The results showed that the identification and classification of pure fuels using QDA, SVM, and ANN achieved 100% accuracy, while the LDA method achieved 98.8% accuracy in distinguishing pure fuel types. Based on the results, the e-nose system demonstrated over 90% accuracy in detecting and classifying fuel adulteration, outperforming traditional methods based on conductivity and pH measurements.