ASSESSMENT OF THE EFFECT OF OPERATING PARAMETERS ON GASOLINE QUALITY

Abstract

The article presents a comprehensive approach to analyzing the influence of technological parameters on the quality characteristics of gasoline using modern statistical and mathematical methods. Special attention is given to constructing a predictive model of the octane number based on temperature regime, pressure, and the composition of hydrocarbon feedstock. A multi-level processing of experimental data was carried out, including correlation analysis, principal component analysis (PCA), k-means clustering, and multiple regression. Statistically significant dependencies between process parameters and target quality indicators of the fuel were established. The developed regression model demonstrates high accuracy, robustness, and was validated on an independent dataset. Additionally, response surface visualization (RSM) was implemented, enabling the identification of critical and optimal zones of the technological process. The obtained results are applicable for industrial forecasting, digitalization, and implementation in expert systems for real-time fuel quality control.