Direct Comparison of Optimized Lunar Campaigns

Much effort has been funneled into missions that will lead humans back into space and beyond Earth's orbit (BEO). Since the fall of 2021, NASA is tasked with returning humans to the Moon with routine cadence and has selected five U.S. organizations to demonstrate further development of Human Landing System (HLS) concepts for the Artemis crewed missions, with a clear aim of minimizing risk. These missions are designed to return a human presence on the moon and serve as a foundation for further interplanetary exploration. To determine the most effective campaign, especially by means of risk reduction, Northrop Grumman developed a machine learning algorithm to quantify the efficacy of individual missions in terms of cost and risk. A domain of 15 discrete variables is iterated through to create >10
⁵
samples from which an optimal cost/risk front is determined within the design space. Further testing isolates rendezvous- and propellant-specific campaigns that are compared to the optimal front and each other. It is shown that campaigns driven by SLS launch vehicles (LVs) and fueled with hypergolic propellants demonstrate the greatest risk aversion, while a mix with smaller LVs and liquid methane conversely demonstrate greatest cost aversion. Additional comparisons are made to evaluate the sensitivity that incremental changes to independent variables, such as LV cost and Δ V requirement, affect an individual campaign's cost and risk.