Investigation of Octane Number Loss Based on Particle Swarm Optimization

Abstract

The Particle swarm optimization (PSO) is a swarm intelligence algorithm that simulates the predatory behavior of birds. It is inspired by the social behavior of bird flocking. It is widely used in many fields because of its easy implementation, high accuracy, and fast convergence. In this article, we propose a method to improve the performance of the PSO algorithm by combining it with a gradient boosting regression (GBR) model. We apply our algorithm for the optimization of octane number (express in RON) loss in the gasoline industry. RON is the most significant indicator that reflects the combustion petrol performance and it is the commercial brand name of petrol (e.g., 89#, 92#, 95#). Our simulation results demonstrate that RON average loss rate was greater than 30%, under the product's sulfur content was no greater than 5μg/g (Euro VI standard is no greater than 10μg/g).