Conceptual Design of Biorefineries Using Agrowastes, Its Decision Support and Design Rationale

This work explores the role of OUTDO, a decision support system, in the conceptual design of multipurpose-multiproduct plants producing lignocellulose-based biofuels and bioproducts.  These plants (mainly based on lignocellulose chemistry and biochemistry) are gaining importance as part of a paradigm change in the energy and process industries.  In particular, the work addresses the process of designing a class of plants known as biorefineries that employ agro/industrial wastes and enzyme-based technologies to produce ethanol and bioproducts.

The conceptual design of this type of biorefineries is part of an inter-institutional research project currently under execution in Mexico whose main objective is to develop the concept of a biorefinery using agro/industrial wastes as feedstock within the context of a mid-size economy.  Ten geographically distributed research groups participate in the project executing work packages related to each processing stage of a biorefinery and its design.  They include pretreatment, saccharification, fermentation, separation, biogas, hydrogen and electricity co-production, enzyme enhancement, synthesis of lignine-based products, plant design and Life Cycle Assessment (LCA) of the biorefinery concept.  As part of the conceptual plant design work package, seven designs have been completed in the past four years using standard techniques and tools.  Most of the process technology employed can be found in the open literature and many of the process conditions for the feedstock conditioning, pre-treatment, hydrolysis and fermentation steps have been corroborated at laboratory scale by the corresponding groups participating in the project.  Each one of the conceptual designs has been modeled and analyzed: steady–state models were solved using commercial simulators, the economic assessment was carried out according to standard profitability analysis procedures, and the environmental impact assessment is being currently carried out with LCA employing conventional assessment methods that involve the use of heterogeneous criteria (qualitative and quantitative).

During the execution of the project new data has been produced that either modifies the existing process functionality or introduces new aspects to be considered.  New stakeholders have been also incorporated along the duration of the project.  Thus, conducting the conceptual design of the biorefinery has become a highly dynamic and very sui generis activity.  It is for these reasons that the need of a flexible environment to keep track of the technical issues associated to the design and of the design process itself was identified.

OUTDO (Oxford University Tool for Decision Organisation) provides several functionalities related to decision making and information and knowledge management:

a) access to a library of multi-criteria decision analysis (MCDA) methods to improve decisions

b) maintenance of the design history for future analysis and re-use

c) representation of the decision rationale; making it explicit improves communication amongst the participants in the design and allows the exploration of the impact of changes in the technological, environmental, economic and social conditions on previous decisions (or even the impact of the change of a given decision in other decisions!) through the propagation of Global Variables throughout the network of decisions

d) integration of probabilistic forecasting methods with the MCDA methods to evaluate alternatives at different points in time (present and future)

e) operates as a repository of design models, data and documents, i.e. acting as a "glue" binding disparate pieces of information.

OUTDO provides the basis to structure the information used in and generated during the design process.  This includes the design strategy and the documentation of the design process and of decision making, e.g. the selection of a model.

To illustrate OUTDO's functionalities we will concentrate on a single design step that evolves one design stage (referred to as an artefact in the rest of the document) into a more detailed design; we will discuss some of the key decision issues and how they were resolved.

Conceptual Design of Biorefineries using Agrowastes

The first artefact is the block diagram drawn in continuous lines in Figure 1.  The first stage of the block diagram is a three-step continuous thermo-chemical Pretreatment (5:1 H₂SO₄ 0.75% v/v aqueous solution:feedstock) that soaks and cooks the feedstock.  Inhibitors in the liquid hidrolisates are eliminated by overliming.  A 92% percent conversion of hemicelluloses is achieved (80% to xylose, 12% to arabinose) and only 3% from cellulose to glucose.  The output stream is fed to the Saccharification & Fermentation stage in which hemicelluloses and celluloses are enzimatically transformed to sugars and then fermented.  A 2%-alcohol stream is sent to the Separation stage producing 99.5% ethanol.  Biogas is produced as a by-product in the Waste Water Treatment stage.  Cogeneration is considered, by burning the lignin residues.  Details of this design can be found in Sanchez et al. (2012).

The second artefact represents a coproduction scheme of ethanol, hydrogen as well as acetic and butyric acids.  The Pretreatment stage was modified to use a batch pressure cooking process (again 5:1 H₂SO₄ 0.75% v/v aqueous solution:feedstock).  The resulting liquid hidrolisates are sent to a Dark Fermentation stage (thermophilic bacteria, 60 C, 25 hours residence time) that produces hydrogen and CO₂ (2:1 molar ratio) (2.23 mol H₂/ mol pentoses).  Acetic and butyric acids are the co-products sent to the Waste Water Treatment stage for biogas production.  A six-fold increment of biogas is achieved compared against the ethanol-only production scheme (the first artefact).  The anaerobic reactor converts 93% of the organic contents to biogas (CH₄ and CO₂, 3:1 molar ratio) with mesophilic bacteria, at 25-45 ¼C and 6 hours residence time.  The residual stream goes to the aerobic reactor where almost all the organic content is removed.  Details of this flowsheet are presented in Sanchez et al. (2011).

Some of the main reasons for changing the Pretreatment technology were that the continuous technology requires the payment of patent royalties, and that there was no adequate experimental equipment available amongst the eight participant groups to test and validate the operating conditions of the continuous operation mode.

New technology emerged during this period with experimental evidence of an improved production of hydrogen and biogas.  An agreement was achieved to collaborate with another research group to incorporate the hydrogen production stage into the design of the biorefinery.

Figure 1.  Biorefinery Process Flowsheet.

Before proceeding with the LCA, an economic comparison of these two artefacts was made calculating capital investment, cost contributions per stage and total production cost of ethanol as a function of plant capacity and feedstock price.  Polysaccharide concentration of feedstock was established as a positive monotonic function of price and was adjusted to local market conditions.  Typical results are shown in Figure 2 and Table 1.  Note that the lowest production costs were obtained for coproduction schemes using cheap feedstock (i.e. 0.42 $/l EtOH @2,100 ton/day DB (dry basis)).  However, for capacities below 1,000 ton/day DB using the ethanol production scheme may be more attractive from an economic point of view under the established conditions.  Capital investment and patent royalties had a significant impact on the decision to discard the use of the continuous pretreatment.  But expected improvements in technology (not to mention the additional consideration of other criteria such as environmental, financial or political) may modify, in the short term, the economic conditions thus shifting this decision. 

Figure 2.  Ethanol Total Production Costs. Enhanced coproduction (i.e. ethanol, biogas and hydrogen) in green and ethanol-biogas in blue.         

Table 1.  Boundary Ethanol Total Production Cost for ethanol plant (columns 2 and 4) and biorefinery (columns 3 and 5).

USD/L	100 ton /day DS		2,100 ton/day DS
35 %	1.80	2.22	0.51	0.42
80 %	1.14	1.34	0.61	0.54

Conclusions

The use of OUTDO is a relatively recent event in the biorefinery design project, so part of the work has been to record an idealised decision trail of the seven designs that have been completed in the last four years.  This exercise is exposing some decisions that should be re-evaluated and is improving the organisation of the accumulated information.  More importantly, OUTDO is being used within the project to maintain its undergoing design history, improve current decision processes (through the use of its MCDA toolkit and the forecasting mechanisms) and foster discussion around the values and preferences of the design participants.  It is expected that as the project unfolds the recorded rationale will allow swift identification of those decisions affected by changes to various design variables (costs, operation conditions, targets, regulation limits, etc.) and the opportunity to backtrack and revise them.

References

A. Sanchez, V. Sevilla-Guitron, G. Magana, P. Melgoza and H. Hernandez. "Co-production of ethanol, hydrogen and biogas using agro-wastes. Conceptual plant design and NPV analysis for mid-size agriculture sectors".  Proc. 21th European Symposium on Computer Aided Process Engineering. Pto. Carras, Greece (2011). Computer-Aided Chemical Engineering Vol. 29. pp. 1884-1889.

A. Sanchez, G. Magana, C. Moreno and V. Sevilla-Guitron. "Conceptual Design and Process Economics of a 2G-Ethanol Production Plant in Medium-Scale Agriculture Sectors". Submitted to Industrial Engineering Chemistry (2012).