Planning of Economically Optimal Biorefineries under Consideration of Location, Capacity and Setup using Evolutionary Algorithms

– What kind of biorefinery should be built where in order to operate it most economically? –

The conservation and substitution of fossil resources by using renewable energy is one of the major challenges of the present. A very comprehensive and versatile utilization of different biomasses can be achieved using biorefineries. Depending on their setup, biorefineries have an output product portfolio which is similar to conventional fossil refineries. Although the technology is generally available, such a plant has not been realized yet, due to lacking profitability. This work develops approach to simultaneously optimize a biorefinery in regards to location, capacity and setup while considering the geographic availability of biomass sources and product sinks.

Main database of the analysis presented in this paper is the Corine Land Cover Survey of 2006. This data set divides Europe?s area into 44 land cover categories, of which 37 exist in Germany. About 15 of those are relevant for supplying residual biomass (biomass without any other high value use). By assigning biomass yields to each category, the geographic availability of biomass can be modeled as a basis for optimization.

The optimization problem has a large number of decision variables which occur in different representations (mainly real and binary). Evolutionary algorithms and evolutionary strategies in particular, are suitable heuristics for finding the best ? or at least a very good ? solution in such cases. Potentially problematic conversions into linear, continuous or concave forms are not necessary.

After initializing the optimization process by randomly generating a number of feasible biorefinery locations with associated capacity and setup. Those solutions are altered by means of recombination and mutation, followed by a selection of those solutions with the best objective function values, namely profit or net present value. Over a large number of iterations, the biorefineries? setups and locations converge towards very good solutions of the problem.