A crucial element of biobased economies is the development of economically sustainable biorefining systems enabling a full exploitation of biomass (and its macrocomponents, i.e. cellulose, hemicellulose and lignin) for the production of chemicals, materials and energy. The huge variety of biomass processing paths in second generation biorefineries leads to complex product portfolios indicating the necessity of maximizing the value derived from biomass feedstock when treated in a biorefinery with the aim to create economically feasible and environmentally sustainable biobased production systems (Kokossis and Yang, 2010). One of the most important and challenging aspects in the quest of producing a set of more sustainable biobased products, is the design of a more profitable, better integrated and more sustainable biorefinery supply chain network (Akgul et al., 2011; Martín and Grossmann, 2010; Čuček et al., 2014). The biomass pretreatment, which fractionates biomass into its three macrocomponents, usually represents the highest costing part of the entire biorefining system (CIMV, 2015). Optimization tools could play a decisive role supporting the decision making process in new biorefinery systems, through the performance of sensitivity analysis, the assessment of technical or economic uncertainties as well as the identification of the major cost drivers (Kim et al., 2013).
Supply chain optimization has emphasized the integration of all the existing material and information flows between the supply chain nodes (i.e. biomass production, storage, transportation, pretreatment, conversion as well as product storage and transportation), while dealing successfully with uncertainties that usually exist in the biobased market (Sharma et al., 2013; De Meyer et al., 2014).
Complex bio-based production systems can also be handled in a powerful way with technology superstructure models. These consist of synthesis blocks that for given objective functions (e.g. profit or net present value maximization, cost or required energy or greenhouse gas emissions minimization, etc.) provide a holistic approach to the design of optimal configurations of the examined production network (Tsakalova et al., 2011; Yuan and Chen, 2012).
This work focuses on the integration of a technology superstructure with a lignocellulosic supply chain model, formulated as a mixed-integer linear programming (MILP) model and applied to an Organosolv-based biorefinery in the South-West of Hungary. The model represents a multi-echelon, multi-feedstock and multi-period supply chain model, aiming at determining the configuration of the biomass network to supply a biorefinery with the necessary biomass feedstock, maximizing the overall profit of the system and at the same time fulfilling targeted regional bioproduct demand over the entire planning horizon, including uncertainties on key parameters. Spatially-explicit features are also embedded in the model formulation, expressing the geographical availability of biomass crops and transportation links. This methodology enables a systematic design of biorefining systems as well as the evaluation of key performance indicators (KPIs) (e.g. costs, environmental impacts). GAMS® is the software tool used for this optimization task (Rosenthal, 2014).
Based on previous works (Hugo and Pistikopoulos, 2005; Zamboni et al., 2009; Giarola et al., 2011), the problem is formulated assuming as given: sets of candidate plants using known technologies (i.e. Organosolv, saccharification and co-fermentation), potential geographical sites for the plants location, available transportation modes (i.e. trucks), spatially-explicit availabilities of the raw material and a regional demand for a set of products. The planning and operational decision variables will be optimized to identify and locate the cultivation and storage sites, the production facilities, the feedstock mix supplied to the selected plants and the optimal logistics.
The biomass market features, considering 3 different types of biomass feedstock (i.e. winter wheat straw, winter barley straw and corn stover), alongside its seasonal and geographical availability, assuming 4 months of available harvest (i.e. July, August, September and October), the technical and economic characteristics of pretreatment and conversion processes as well as competitive issues are the main input elements of the proposed MILP model (BIOCORE Final Report, 2014; NOVA, 2014; Humbird et al., 2011). The examined Hungarian region, for better data accuracy, is discretised into 102 cells with a surface of 225 km2 each, focusing on the arable land and its factors (i.e. arable land distribution and cereal straw availabilities per cell).
A single economic objective function is considered, aiming at maximizing the total profit of the whole network, subject to production, storage, transportation and demand constraints, over the entire planning horizon, which is defined to be 1 year, divided into 12 months. Subsequently, the total profit is calculated from the annual revenues and costs, where the latter term includes both annualized capital (i.e. installation investments of selected plants) and operational costs (i.e. biomass purchase, storage, transport, processing as well as product storage and transportation).In addition, the supply chain behavior is captured through the definition of both biomass and product mass balances, alongside logical constraints that have to be satisfied at each of the supply chain nodes.
The model embedding the definition of the process technology superstructure for a portfolio of selected biobased products (e.g. cellulose, hemicellulose, lignin, ethanol), will enable the analysis of the key sources of uncertainty in the planning of biorefining systems (e.g. capital and operating costs, technological yields) and their evolution over time.
The EC (Reneseng-607415 FP7-PEOPLE-2013-ITN) is gratefully acknowledged for supporting this work.
Akgul, O., Shah, N., & Papageorgiou, L., G. (2011). ESCAPE21 Journal, 1799-1803.
BIOCORE Final Report (2014):
Čuček, L., Martín, M., Grossmann, I., E., Kravanja, Z. (2014). Computers and Chemical Engineering, 66, 57-70.
De Meyer, A., Cattrysse, D., Rasinmäki, J., & Van Orshoven, J. (2014). Renewable and Sustainable Energy Reviews, 31, 657-670.
Giarola, S., Zamboni, A., & Bezzo, F. (2011). Computers and Chemical Engineering, 35, 1782-1797.
Hugo, A., & Pistikopoulos, E. (2005). Journal of Cleaner Production, 13, 1428-1448.
Humbird, D., Davis, R., Tao, L., Kinchin, C., Hsu, D., & Aden, A. (2011). Process Design and Economics for Biochemical Conversion of Lignocellulosic Biomass to Ethanol, NREL Report, Golden Colorado, USA.
Kim, J., Sen, S., M., & Maravelias, C., T. (2013). Energy Environ. Sci., 6, 1093-1104.
Kokossis, A., C., & Yang, A. (2010). Computers and Chemical Engineering, 34, 1397-1405.
Martín, M., & Grossmann, I., E. (2010). ESCAPE20 Journal, 943-948.
NOVA (2014). NOVA paper #5 on bio-based economy 2014-11 (available at: www.bio-based.eu/markets).
Rosenthal (2014). GAMS – A User’s Guide. GAMS Development Corporation, Washington, DC, USA.
Sharma, B., Ingalls, R., G., Jones, C., L., & Khanchi, A. (2013). Renewable and Sustainable Energy Reviews, 24, 608-627.