(249d) Superstructure Optimization for Catalytic Conversion of Lignocellulosic Biomass to Fermentable Sugars

The major challenge in utilizing the lignocellulosic biomass for production of chemicals and fuels is the extraction of sugars trapped inside the biomass effectively. Recently, a novel nonenzymatic technology¹ has been reported for production of sugars via single-stage conversion of lignocellulosic biomass using biomass-derived gamma-valerolactone (GVL) as a solvent. Although this single-stage conversion technology is very promising since it leads to a high sugar yield while eliminating costly pretreatment steps, obtaining fermentable sugars with high purity and developing a separation system for effective GVL recovery and recycling are critical for economic feasibility of the process. Therefore, a series of effective separation technologies should be developed and integrated with the upstream conversion technology for a cost-effective production of fermentable sugars.

There are various alternatives for separating biomass-derived sugars from the GVL solvent, lignin and biomass residues. The biomass-derived sugars can be separated from GVL, lignin and residues by supercritical CO₂ extraction, phase separation upon salt addition, or multi-stage extraction using an alkylphenol solvent. Following the separation of sugars, GVL solvent can be recovered from lignin and residues using precipitation of solids, evaporation of GVL, or adsorption of lignin. Later, GVL can be recycled back to the sugar production step, while lignin and biomass residues can be utilized for steam and power generation in a utility plant. However, it is not clear that what combination of separation technologies will make the production of fermentable sugars more cost-effective, which requires a systematic evaluation of all the alternatives. To this end, we generated a superstructure that integrates the upstream sugar production technology with the separation alternatives to obtain fermentable sugars with high purity, while achieving high recovery and recycling of the GVL solvent, and utilizing the biomass residues in a utility plant for steam and power generation.

In the utility plant, the energy generated from combustion of biomass residues is used to produce high-pressure (HP) steam in a boiler. In addition, HP steam can be purchased to satisfy the utility demand. HP steam can be used for heating, electricity generation, and/or converted to medium-pressure (MP) or low-pressure (LP) steam to be used for electricity generation and heating. Electricity can be generated in steam turbines and/or purchased to meet the electricity demand. In case of electricity surplus, excess electricity can be sold to the grid to generate extra revenue. The utility demands are determined by heat integration of process streams. To minimize the requirement for utilities, the maximum heat recovery should be obtained in a heat exchange operation, which is referred as a heat cascade. The heat cascade can be modeled as a transshipment problem², but with variable heat duties (to be determined by the superstructure model). After the heat integration, remaining heating requirement can be satisfied using steam produced in the utility plant, which is also included in the superstructure model. Selection of how much electricity to produce and/or purchase, selection of the type and amount of steam (HP, MP, LP) to produce and its conditions (temperature and pressure; superheated, saturated or wet steam) are the decision variables for the utility plant.

In summary, our superstructure model includes: (1) sugar production alternatives, (2) the heat recovery network, and (3) the utility system. The modeling of the major unit operations is based on experimental parameters (e.g. conversions, partition coefficients). The proposed optimization formulation is a mixed-integer nonlinear programming (MINLP) model with detailed equipment models, constraints imposing experimental inlet/outlet conditions and purity constraints. The objective function is the maximization of the net revenue obtained from the sugar product and excess electricity sales minus the costs of raw materials and utilities as well as operating and annualized capital costs of selected process equipment units. Using this model, we identified various process alternatives which we discuss in detail.  

References

1. J. S. Luterbacher, J. M. Rand, D. M. Alonso, J. Han, J. T. Youngquist, C. T. Maravelias, B. F. Pfleger, and J. A. Dumesic, Science, 2014, 343, 277-280.
2. S. A. Papoulias and I. E. Grossmann, Computers and Chemical Engineering, 1983, 7, 707-721.