(644e) An Integrated Targeting and Design Method for Saving Water in 2nd Generation Biorefineries

According to IEA (www.ieabioenergy.com, 2009) future biorefineries are en route to participate effectively in the effort for disengagement from fossil resources used in the production of fuels and products that are fast depleting (BP, Statistical Review of World Energy, 2012) and also raise environmental concerns because of their considerable GHG emissions (EIA, 1999). Biorefineries have themselves to be sustainable with respect to material and energy consumption and waste production. At present, biorefineries exhibit large water consumption (NRC, Water Implications of Biofuels Production in the United States, 2007) with future projections for it to decrease, but not substantially (Technical Report ANL/ESD/09-1). Moreover,  it is shown that especially for biorefineries that adopt biochemical routes the water consumption is very large and by consequence the water treatment and distillation facilities may cost up to 29% of the total capital expenditure (EuropaBio, Biorefinery feasibility study, 2001). All these imply clear scope for the development of advanced water saving technologies that will optimally exploit the options for reuse, regeneration and recycle.
Biorefineries make excessive use of water and their feasibility largely depends upon the economical use, re-use and regeneration of the water resources.
Targeting technologies assume additional significance as early-stage analysis is common and the chemical flowsheet is not yet developed. An early assessment of the capabilities to save water is certainly invaluable. Aside targeting, the network development is also important when pilots need be scaled up through detailed configurations. The combinations are numerous and superstructure schemes are useful to consolidate the different scenarios. However, water mixtures are far from ideal and conventional technologies can prove very inaccurate.
The pioneering work of Water Pinch technology (Chemical Engineering Science, 49 (7), 981–1006, 1994) can be a natural starting point but the underlying assumptions usually fail to apply reliable to the study of the biorefinery water streams. Other than the apparent reason of multiple components, components can be hardly characterized as contaminants (they constitute instead useful material to recycle) and participate in concentrations and quantities largely deviating from the dilute mixtures typically considered in the literature (Computer Aided Chemical Engineering 2012, 30, 7-10). Due to interactions between components, mass transfer cannot be simplified as in usual practices whereas the superstructure optimization models require nonlinearities that are impossible to classify or handle reliably.
The work is motivated by the need to retain the potential of the targeting methodology, further extending its use to a decomposition scheme that simplifies and reduces the superstructure models. To this purpose, the paper capitalizes on transshipment models now extended with additional constraints that reflect on non-ideal aspects of the water streams. The extension makes use of step is based on Relative Residual analysis (Computer Aided Chemical Engineering 2012, 30, 7-10); the decomposition approach bears similarities to the work by the Argaez et al. (Comp. & Chemical Engineering, 1999, 23 (10), 1439-1453). The extended model assesses targets further reducing the solution space that is defined by the water streams and inlet concentrations. The approach involves a preliminary stage to define the set of constraints as required by the augmented transshipment model whose optimization determines the size and the stream network of the superstructure.
The work contains several illustrations but is also tested using data from the real-life industrial biorefinery presented by Mountraki et al. (Computer Aided Chemical Engineering 2011, 29, 1381-1385). An important advantage of the proposed solution is that it does not require the procurement of additional equipment, but exploits the distillation process in place, used to recover the large quantities of acids used for the pretreatment of biomass.
The biorefinery is using more than 44 t/h of water for removing components. Components that are removed by water are: acetic acid and formic acid. The target for minimum fresh water flowrate calculated by water pinch analysis (Computer Aided Chemical Engineering 2011, 29, 1381-1385) is 24.8 t/h, suggesting 43.6% savings. However, the quantities of acids are too large to permit a conventional water integration study (R. Smith, 2005. Chemical Process Design and Integration. John Wiley and sons). Considering only reuse, the application of the targeting model proposed in this work results in a requirement of 35.9 t/h of fresh water, setting a new target of 18.41 % savings. The reduced superstructure model results then in a network that approaches the target set by the residual model at 39.36 t/h of fresh water, achieving 10.5% savings. When the option of using water produced by the distillation processes (regenerated water) is considered, the residual model results in 57.6 % savings and 18.64 t/h fresh water consumption. Then the reduced superstructure results in a network that uses only 18.64 t/h of fresh water and reaches the exact target of the residual model: 57.6 % savings.