Optimization-based design for lignocellulosic ethanol production: a case study of the state of Maharashtra, India

A large-scale mixed-integer linear programming model is developed for the supply chain and processing of agricultural residue to produce ethanol. The novelty in the presented model is the consideration of multiple biorefinery location options, unrestricted number of biorefineries, and the distribution of ethanol. They are implemented as extensions to an existing model which considers multiple biomass source locations, biomass types with varying seasonality, and conversion pathways. The objective is to minimize the total annualized cost while meeting a monthly demand for ethanol. A case study of ethanol production in the state of Maharashtra, India, is considered with 33 potential biorefinery locations and 35 blending sites. The impacts of varying ethanol demand and availability of key feedstock types on system design and operation are rigorously quantified. For a base case considering an ethanol blending of 5%, the model recommended four locations for setting up biorefineries with an average monthly production of 4.2 Gg ethanol/month. The cost of ethanol was estimated to be ₹54/l ($0.73/l). Varying the ethanol blending target from 2.5 to 10% increased the optimal number of biorefineries from 3 to 6, respectively. This also increased the ethanol cost by 17% due to increased feedstock and transport cost. Cotton-based biomass was the most valuable feedstock due to its low cost, and its unavailability increased the ethanol cost by 24.1% over the base case. The results provide region-specific recommendations for setting up lignocellulosic biorefineries that can be used by government and businesses.