Simulating bioregenerative life support for deep space missions without resupply: development of a fully closed stoichiometric model

For long-duration space missions it is imperative to use (bio)regenerative life support systems (BLSS) to reduce mass and volume. In most BLSS studies, only a fraction of the supplies is provided by the system (e.g., 40% of the food supply). The rest is considered to be taken on board at departure, or provided through resupply missions. In the E|A|S (Evolving Asteroid Starships) research project, autonomous long-duration space missions are envisioned without the possibility of resupply. In such a scenario, a fully closed BLSS, with minimal or no material loss is vital. A stoichiometric model is presented, based upon previous work (Vermeulen et al., 2019), updated with a detailed review of publicly available MELiSSA literature from 1989-2022. A total of 23 studies provide detailed descriptions of the stoichiometry of the MELiSSA BLSS. Eleven of these studies describe the entire loop, while the others describe one or two of the five MELiSSA compartments. None of these studies provided a model for a fully closed stoichiometry, either because the goal was to only provide a part of the food or oxygen supply, or because some part of the stoichiometry was open, and outside resupply was deliberately integrated. The resulting stoichiometry of the presented study describes the cycling of the CHON elements through all five MELiSSA compartments and one auxiliary compartment. This is done through a compact set of chemical equations with fixed coefficients. No kinetic limitations were integrated because this exceeded the level of detail deemed necessary for further agent-based modeling studies. A spreadsheet model was developed that simulated the flow of all relevant compounds for a crew of six. The results showed that at steady state a very high degree of closure was attained, with 12 out of 14 compounds exhibiting zero loss in between iterations, and oxygen and CO2 displaying a minor loss that amounted to respectively 0.85 g and  2.31 g per person per year. At steady state, the VFA spectrum that was produced by the liquefying Compartment 1 consisted of 48.88% acetate, 13.25% propionate, 37.53% butyrate and 0.34% valerate, a proportion that is in line with observations of anaerobic digestion in literature. This stoichiometric model is used as the foundation of an ABM of the MELiSSA loop. In this ABM, humans, plant plots and bioreactors are all modeled as virtual agents that interact with each other. The ABM approach enables the introduction of heterogeneity at the level of agent attributes, and allows for the study of complex interactions that cannot be observed in more common equation-based modeling techniques. The most recent version of the ABM will also presented with a range of results from different simulation experiments.