Chemical Synthesis Data Modeling Based on Mathematical Optimization

Abstract

As one of the most important high value-added raw materials in the chemical industry, the synthesis of C4 Olefin by ethanol coupling was of great significance in the field of the chemical industry. Different catalysts and various conditions have different effects on the chemical reaction. This paper is based on the relevant data set. Firstly, Pearson and Spearman correlation coefficient method and corresponding hypothesis test are used to get the influence of different catalysts on the chemical reaction. Ethanol conversion and C4 Olefin selectivity are positively correlated with temperature. Secondly, a multivariate linear regression model with significant core variables is constructed to investigate the effects of catalyst combination and temperature on ethanol conversion and C4 Olefin selectivity. It can be concluded that the ethanol concentration is greatly affected by temperature and CO loading, and there is a positive and negative correlation between ethanol concentration and CO loading. The selectivity of C4 Olefin is affected by temperature and is positively correlated with the charge ratio of CO/SiO2 and HAP. Finally, by using a multiple regression equation and simulated annealing model, it can be obtained that when the loading of CO is 4.75 wt%, the loading ratio of CO/SiO2 and HAP is 1 : 1.4242, the concentration of ethanol is 0.3658 ml/min, and the temperature is 448.21°C, the loading ratio of CO/SiO2 and HAP is 1 : 1.4242, the concentration of ethanol is 0.3658 ml/min, and the temperature is 448.21°C, the yield of C4 Olefin can reach a higher value.

Research on optimal design of reaction conditions based on mixed statistical model

“…Naturally, measuring the difference between two random variables inevitably introduces their respective probability distribution functions, so it is necessary to estimate the probability distribution function of the designed 𝐷 . The empirical distribution function 𝐹 of 𝐷 is constructed, as can be seen by Clivenko's theorem [1], and as the sample size increases, the empirical distribution function gradually tends to converge with the distribution function of the population in a probability 1. Therefore, in calculations, the distribution function of the population can be approximated using the empirical distribution function.…”

“…C4 olefins are involved in various fields of chemistry and are widely used, so it is of great significance to study the relationship between the reaction conditions and products of C4 olefins. At present, the research methods are mainly divided into the following categories: First, through statistical regression methods [1,2], multiple linear regression, partial least squares regression, Lasso regression, neural network and other methods are used to establish regression models, and the variable regression coefficients are obtained through significance test and calculation, so as to analyze the influence of each variable and reaction conditions on the preparation of C4 olefins. Second, the interaction of different conditions was quantitatively analyzed through cross-experiments [6], and the influence of catalytic conditions on the selectivity of C4 olefins was explored by principal component analysis and Pearson correlation coefficient.…”