Multiobjective Optimization Approach for Cellulosic Biomass Pretreatment
详细信息 查看全文
- 作者：Lorena Pedraza-Segura ; Hector Toribio-Cuaya ; Antonio Flores-Tlacuahuac
- 刊名：Industrial & Engineering Chemistry Research
- 出版年：2013
- 出版时间：April 17, 2013
- 年：2013
- 卷：52
- 期：15
- 页码：5357-5364
- 全文大小：280K
- 年卷期：v.52,no.15(April 17, 2013)
- ISSN：1520-5045
文摘
Ethanol manufactured from cellulosic residues has the potential to become one of the alternative energy sources in the portfolio of renewable energies. Before reaching that status some technological problems must be resolved. Among them are the availability of reliable mathematical models for simulation, design, optimization, and control purposes as one the major hurdles that remains to be tackled. Another challenge refers to process design and operation when several conflicting and/or contradictory goals must be met. In this work we use corn cob, a widely available cellulosic residue, for ethanol manufacturing. Among all the steps that must be carried out for ethanol manufacturing we concentrate on the pretreatment process. Hence, the corn cob thermo-chemical pretreatment step was studied in order to model plant cell wall degradation and the release of polymers and monomeric sugars. During the pretreatment step xylose, xylan and some unwanted products are attained. We are mostly interested in maximizing the concentration of xylose while simultaneously minimizing the amount of xylan and undesired side products. However, increasing the amount of xylose will also increase the amount of xylan and side products. Therefore, both goals cannot be achieved simultaneously, and a trade-off solution must be sought. In this work trade-off solutions are computed as the point on the Pareto front that features minimum distance to the Utopia region. To verify the goodness of the trade-off solutions such a theoretical trade-off solution was implemented on experimental facilities. The experimental results clearly demonstrate that maximum xylose yield, and a minimum amount of xylan and side products can be obtained by using multiobjective optimization techniques especially when compared against similar results obtained from linear design of experimental techniques.