Research on the Prediction Method for Finished Gasoline Octane Number Based on TOL-WOA-XGBoost

Octane number is a key indicator for evaluating the quality of gasoline blending formulas. Due to the inability of existing near-infrared analyzers to achieve precise measurements during online blending of finished gasoline, this study proposes a prediction model for the octane number based on Extreme Gradient Boosting (XGBoost) to ensure accurate and scientific measurements. Given the complexity and difficulty in determining the optimal parameters of XGBoost, the Whale Optimization Algorithm (WOA) is improved to optimize the model's hyperparameters. Specifically, the Tent Chaotic Opposition-based Learning and Levy flight strategy (TOL) are integrated into WOA to enhance its global search capability, improve local search precision, and accelerate convergence. Finally, experiments based on real industrial data were conducted to validate the model. The simulation results demonstrate that the proposed TOL-WOA-XGBoost model is effective and feasible for predicting the research octane number of gasoline.