A Computational Model of Networked Small-Scale Fuel Synthesis Demonstrating Greater Production Flexibility and Specificity

The rapid pace of industrial change over the past hundred years has led to any number of paradigm shifts in the way business is conducted and technologies are applied, but economies of large scale have persisted in the energy sector. In an age of automation and mass-production of small units, however, complex networking of many small energy systems can permit novel application of established technologies. This dissertation explores how established fuel synthesis technologies might behave in an automated network in which familiar units are arranged in unfamiliar ways. The flexibility afforded by automation and small scale operation allows for potentially complementary means of exploiting the fungible nature of hydrocarbon resources. Beyond any benefits of small-scale incurred from mass production and learning, fuel synthesis is a process with sensitivities to input streams that a network could exploit in a nuanced way. The completed work demonstrates that a network of small-scale fuel synthesis reactors and thermal crackers, based on current industrial practices at large monolithic scale, can be networked to dramatically sharpen the chemical spectrum they produce. In order to study the behavior of such a network in ways that are unavailable in current software, a hierarchical numerical modeling code was developed to offer greater flexibility to nest and optimize network configurations within network configurations, reflecting the modularity of the networks it is meant to simulate. This new code is capable of simulating aggressively numerically constrained networks, dynamically substituting various configurations while optimizing them across user-specified variables. Various weighting schemes were developed to facilitate more rapid convergence to a numerical solution so that highly constrained recycling schemes could be reconciled to a steady state that would produce the specified output spectrum. Modular units were coded to simulate the essential properties of real processes and technologies, with close attention paid to the sensitivity of these processes to input conditions, so that these units could be assembled in various configurations and subjected to user-specified constraints. Coded modules were designed under the principle that these individual units need not be custom-made or technologically ahead of their time; the benefits explored by network simulations are incurred not by dramatically upgrading the processes being simulated, but rather by directing and redirecting the chemical streams which are subject to those processes to tailor the outcome to the desired product. This principle was applied to chemical separation in an analytical framework in order to derive how unremarkable separators might be networked to produce remarkable precision of separation. Such precision is important because the direction and redirection of chemical streams is predicated on the ability to select the destination of a particular chemical. The effect of networking fuel synthesis reactors and thermal crackers was studied for unidirectional flows in order to understand how repeated applications of these units at smaller scale sharpen the spectrum relative to single large scale application. These fuel synthesis reactors and thermal crackers were also configured in aggressively recycled networks, imposing more severe constraints on the output spectrum. This work demonstrated that fuel synthesis at industrial output scales need not operate in monolithic units and can benefit dramatically from judicious networking, to the point that a network of units that would otherwise have produced a broad spectrum of chemical flavors can be configured to produce only a single user-specified output chemical.