The Optimization Model for Reducing RON Loss in Gasoline Refining Process

As gasoline is the main fuel of small vehicles, exhaust emissions from its combustion will affect air quality. The focus cleaning to reduce sulfur and olefin content in while maintaining RON as much possible. reduction bring great economic losses enterprises. Therefore, it very important for petrochemical enterprises construct a loss model refining process. construction, which reduces during refining, question this paper. By Python SPSS software, we got two variable filtering methods: random forest importance PCA filtering, combined with SVR models, product were predicted. order original data by Excel maximum minimum removal, 3σ criterion deletion too many sites incomplete data, filling empty values mean within hours. Several prediction models established help variables selected compared one based on Gaussian, linear, polynomial, Sigmoid kernel functions; other model. was constructed, use evaluation functions such MSE, R 2 , RMSE evaluate subject condition. We convert problem into linear nonlinear optimization problems: function forest; combination forest. Optimizing each sample, method find optimal solution local sample. are evaluated perspectives degree, rate, running speed, etc.