(390ag) A Novel Superstructure Framework for Tailor-Made Middle Distillates for a Net-Zero Carbon Economy.

As the world advances, the demand for middle distillates, particularly jet fuel and diesel grows rapidly, driven by factors such as rising global demand for transportation and energy fuels. This demand translates to an increase in CO₂ emissions as the current fuels’ sources are fossils. Notably, hard-to-decarbonize sectors such as aviation, shipping, and heavy hauling transport are challenging to transition away from fossil fuels since they cannot be easily electrified. Therefore, an alternative solution such as biofuels can be a sustainable substitute for fossil fuels to decarbonize these sectors. Renewable feedstocks like lignocellulosic biomass can be transformed into low-carbon fuels to partially meet the existing demands of fuels for hard-to-decarbonize sectors. In the last 30 years, a plethora of conversion pathways and catalytic upgrading alternatives for lignocellulosic materials have been discovered. However, a crucial decision in chemical process design, which involves the selection of methods for process representation, simulation, and optimization in a cost-effective and sustainable manner becomes challenging. Typical approaches such as detailed simulations or lifecycle analyses struggle to efficiently explore all possible alternatives because of the huge design space.

Hence, to fill this gap, in this research, we developed a novel superstructure-based optimization framework and mathematical models to study the diverse pathways for middle distillates production. Our superstructure contains 500 catalysts, 165 biomass upgrading chemistries and 320 individual biofuels. Notably, our superstructure allows three optimal decisions: the choice of catalysts, biomass upgrading chemistries, and blend compositions of middle distillates. We use machine learning (graph neural network, artificial neural network) to parameterize the properties of biofuels that have not been tested experimentally as jet fuels or diesel. Then, the rationally designed blends of the transformed biomass molecules, identified via optimization, can be used as novel fuel products, such as diesel and jet fuels, which could match or outperform existing fossil fuel counterparts. We formulate this integrated process and product design as a multi-objective mixed integer non-linear programming problem. Our multi-objectives are aimed at identifying the minimum selling price of our desired fuels and their minimum greenhouse gas emissions (CO₂ equivalent).

Using the developed framework, we examine the tradeoffs between the economics (cost) and environmental sustainability of different biofuel designs. This research aims to inform policymakers and industries on novel product blends which are cost-effective and environmentally sustainable for decarbonizing the transportation sector, contributing to a net-zero carbon economy.