This thesis develops quantitative methods for evaluation and design of large-scale biofuel production systems with a particular focus on bioreactor-based fuel systems. In Chapter 2, a lifecycle assessment (LCA) method is integrated with chemical process modeling to select from different process designs the one that maximizes the energy efficiency and minimizes the environmental impact of a production system. An algae-based ethanol production technology, which is in the process of commercialization, is used as a case study. Motivated by this case study, Chapter 3 studies the selection of process designs and production capacity of highly distributed bioreactor-based fuel system from an economic perspective. Nonlinear optimization models based on net present value maximization are developed that aim at selecting the optimal capacities of production equipment for both integrated and distributed-centralized process designs on symmetric production layouts. Global sensitivity analysis based on Monte Carlo estimates is performed to show the impact of different parameters on the optimal capacity decision and the corresponding net present value. Conditional Value at Risk optimization is used to compare the optimal capacity for a risk-neutral planner versus a risk-averse decision maker. Chapter 4 studies mobile distributed processing in biofuel industry as vehicle routing problem and production equipment location with an underlying pipeline network as facility location problem with a focus on general production costs. Formulations and algorithms are developed to explore how fixed cost and concavity in the production cost increases the theoretical complexity of these problems.

A mixed integer programming model is developed to determine a logistical design for maximizing rates of return to harvest, storage, transportation, and bioreflning of herbaceous crop residue for production of biofuels and feed for ruminant animals. The primary objective of this research is to identify the optimal location, scale, and number of pretreatment and biorefinery plants in northeastern North Dakota. The pretreatment and biorefinery plants are modeled under the assumption that they utilize recent technological advancement in AFEX and Simultaneous Saccharification and Fermentation, respectively. Potential feedstocks include wheat straw, barley straw, Durum straw, and com stover. Results indicate that the minimum ethanol rack price that will effectively trigger the production of cellulosic ethanol is $1.75 per gallon.

The objective of this study is to develop a model to evaluate the economic feasibility of biofuels production, and in particular to isolate the variables crucial to feasibility. The model constructed to define these variables is unique in its ability to accommodate a variety of plants and to integrate all portions of the production process; it was tested on a case study of a Euphorbia lathyris industry. The model minimizes costs of production to determine the best configuration for the industry. Total cost equals the sum of costs incurred in each segment of the process: growth, harvest, transport, and extraction. The solution is determined through a non-linear transportation- transshipment algorithm which describes production as a series of nodes and links. Specific application of the model was analysis of E. lathyris biofuel production in Arizona. Simulations were run examining the sensitivity of biocrude cost to changes in input parameters. Conclusions are summarized as follows. * No change in any single element can reduce final cost sufficiently to enable competitive production in the near future. * The major factor necessary to bring cost into range is improvement in biological yield. Two components of equal importance are tonnage produced per acre and percentage extractables recovered in processing. * Lowering cropping costs provided the most effective improvements of economic inputs. Perennial crops significantly reduced farm costs. * Transportation costs outweighed economies of scale in extraction; extractor location close to crops is more efficient than centralized. The cost minimization model was successfully used to isolate the critical factors for an E. lathyris industry in an arid region. Results determine that this industry would not be competitive in Arizona without dramatic improvements in yields and moderate changes in a combination of input costs. Viability is critically dependent on improvements in tonnage yield produced per acre and percent extractables recovered.