Research on Ethanol-coupled Preparation of C₄ Olefins Based on Multi-model Combination and Genetic Algorithm

Methods such as Ridge Regression analysis, Partial Least Squares regression are common techniques to analyze the correlation degree of factors in different scenarios. Ridge Regression analysis can effectively deal with multicollinearity problems, Partial Least Squares regression rule is popular for its simplicity and intuitiveness. This paper comprehensively use these three algorithms to discuss the multivariate correlation between ethanol conversion, C4 olefin selectivity, and catalyst temperature. In this paper, the regression model between multiple influencing factors was established by using Ridge Regression and Partial Least Squares regression method and the relevance in ethanol-coupled preparation of C4 olefins was revealed. Genetic algorithms are employed to find the optimal catalyst and temperature combination for maximizing C4 olefin yield. Their global optimization capability and adaptability make them well-suited for nonlinear, multi-peak optimization problems. A fitness function targeting C4 olefin yield was constructed to systematically search for the best combination. Through this series of analyses, this paper not only provides a theoretical basis and practical guidance for optimizing the ethanol conversion process, but also lays a foundation for further research and explores the possibility of improving the reaction efficiency under multivariate conditions.