Study on preparation of Olefins by Coupling ethanol based on simulated annealing Model and Neural Network algorithm

Abstract

Ethanol molecules can be prepared by biomass fermentation, with a wide range of sources, green and clean, and the production of high value-added C4 alkenes with ethanol as a platform molecule has great application prospects and economic benefits has attracted wide attention at home and abroad. Meanwhile, optimizing the reaction conditions for the preparation of butene can also appropriately reduce the waste of resources and make a certain contribution to environmental protection. This paper collects relevant response data, uses fit type and Linear model for data fitting and comparative analysis, calculates the average value of each item of data selected from the five experimental groups, and then uses MATLAB two-dimensional broken line drawing for analyzing the changing law of the data more vividly and intuitively. Finally, the simulated annealing algorithm model and the genetic algorithm model are used to optimize the temperature and catalyst combination of the highest C4 olefin yield, the GRGNon-Linear method, the simple simplex method, and the Evolutionary evolution algorithm and Nesterovmethod artificial firefly algorithm are used to analyze and calculate. Thus, the optimal catalyst combination is 200mg1wt%Co/SiO2-200mgHAP ethanol concentration 0.9ml/min, and the temperature of 425℃ is the maximum yield of C4 olefins. Moreover, 350 degrees is the optimum temperature in the temperature range below 350 degrees. Finally, we use the experimental data of the BP (BackPropagation) neural network prediction algorithm to predict and analyze the feasible scheme and explain the detailed reasons for choosing the experimental conditions to determine five groups of better experimental schemes.