Quantitative probabilistic modeling of environmental control and life support System resilience for long-duration human spaceflight

Abstract

The future of human space exploration will see crews travel farther and remain in space for longer durations than ever before. For the first time in the history of human spaceflight, the Environmental Control and Life Support Systems (ECLSS) that sustain the crew in their habitat will have to function without rapid resupply or abort-to-Earth capability in the event of an emergency. In this environment, reliability and resilience will become more dominant design drivers, and will need to be considered alongside traditional system metrics such as mass and cost early in the design process in order to select the optimal ECLSS design for a given mission. This thesis presents the use of semi-Markov process (SMP) models to quantify the resilience of long-duration ECLSS. An algorithm is defined to translate ECLSS design data - including system architecture, buffer sizes, and component reliability information - into an SMP and then use that SMP to calculate resilience metrics such as the probability of system failure before the end of mission and the number of spares for each component that are required to achieve a certain probability of success. This methodology is demonstrated on a notional ECLSS, and then used to determine logistics requirements for a Mars One surface habitat Life Support Unit and examine the trade between resupply mass and the probability that sucient spares are supplied. This case study found that, if sparing is performed at the processor level, 10,410 kg of spares would have to be provided in each resupply mission in order to provide a probability greater than 0.999 that sucient spares are available to complete all required repairs. This is equivalent to over 75% of the mass of consumables that would be required to sustain an open loop system for the same duration. When coupled with the increased uncertainty associated with regenerative systems, the low mass savings associated with the selection of regenerative rather than open loop indicate that, at current reliability levels and with spares implemented at the processor level, regenerative ECLSS may not be the optimal design choice for a given mission. The SMP methodology described in this thesis provides an analytical means to quantify system resilience based on system design data, thereby facilitating the use of formal multiobjective optimization methods and trade studies to create ECLSS with the appropriate balance between mass and resilience for a given mission.