期刊
INDUSTRIAL CROPS AND PRODUCTS
卷 66, 期 -, 页码 52-61出版社
ELSEVIER SCIENCE BV
DOI: 10.1016/j.indcrop.2014.12.022
关键词
Artificial neural network; Fermentable sugars; Lignocellulosic materials; Optimization; Stepwise regression
资金
- National Council of Technological and Scientific Development (CNPq) [402102/2012-6]
- Coordination for the Improvement of Personnel in Higher Education (CAPES)
- Sao Paulo Research Foundation (FAPESP) [2013/17497-5]
Various pretreatment techniques can change the physical and chemical structure of lignocellulosic biomass and improve the hydrolysis rates. High-intensity ultrasound could be a promising technique in the biomass pretreatment process. The objective of this work was to study the effect of biomass concentration, pH, ultrasonic power level and sonication time on the production yield in total sugars (S-T) and reducing sugars (S-R) during the pretreatment of banana flower-stalk biomass. A qualitative evaluation was carried out by scanning electron microscopy, showing a disruptive effect on the biomass structure at high ultrasonic power levels and low biomass concentrations. An experimental design with three-levels for the four-variables was used in order to set the conditions for the pretreatments. Stepwise regression (SRG) and an artificial neural network (ANN) were applied in order to establish mathematical models that could represent and be used to study the dependence of the factors on both the S-T and S-R yields. The statistical results indicated that the ANN approach provided a more accurate estimation than SRG. (C) 2014 Elsevier B.V. All rights reserved.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据