基于神经网络的新疆棉花秸秆高效酶解糖化研究

项目名称： 基于神经网络的新疆棉花秸秆高效酶解糖化研究

项目编号： No.21464011

项目类型： 地区科学基金项目

立项/批准年度： 2015

项目学科： 数理科学和化学

项目作者： 周娜

作者单位： 石河子大学

项目金额： 50万元

中文摘要： 生物质糖化是生物质高值化利用的重要途径。本项目针对新疆棉秆利用过程中影响棉秆酶解糖化的关键因素不明，棉秆三组分等影响酶解的作用机理不清楚，棉秆高效酶解糖化的预处理与酶解适配机制不明，从而致使棉秆转化率和利用率低等问题，拟选用稀酸处理、碱处理、蒸汽爆破联合碱处理3种预处理技术，研究半纤维素、木质素、以及二者的同时去除对棉秆酶解糖化的影响及作用机理，阐明影响棉秆高效酶解糖化的关键因素及机制。采用神经网络等方法分别对预处理、酶解过程进行模拟预测，并在对两步转化过程进行统计分析的基础上，建立涵盖预处理和酶解过程的综合模型，探寻棉秆高效糖化的最佳预处理与酶解适配机制，为深入理解人类干预棉秆等木质纤维素类生物质的降解过程，提高其降解转化效率奠定应用基础。

中文关键词： 纤维素；棉花秸秆；酶解糖化；机理；神经网络

英文摘要： It is very important to improve the saccharfication of biomass for high-valued biomass utilization. To investigate and improve the efficiency of enzymatically conversing cotton stalk to fermentable sugars, three kinds of methods, dilute sulfuric acid, alkaline sodium hydroxide and alkaline steam explosion were selected to focus on analyzing how the removal of hemicellulose, lignin, or both of hemicellulose and lignin, respectively, affect the saccharfication of the pretreated cotton stalk materials.The effect of different pretreatment parameter (i.e., acid/alkali concentration, temperature, and reaction time, solid/liquid ratio,etc.)on the composition and surface structure of cotton stalk were investigated, followed by enzymatically hydrolyzing these materials into fermentable sugars, the mechanism of saccharfication was also explored. To search further potential optimum pretreatment and hydrolysis conditions, an artifical neural network (ANN) was created to model both the pretreatment and enzymatic hydrolysis steps using the prior experimental data to train it. We wish this work could provide some help in understanding the degradation process of lignocellulosic biomass and eventually help improving its conversion efficiency.

英文关键词： cellulose;cotton stalk;enzymatic sacchrification;mechanism;artificial neural network