Towards supervisory Model Predictive Control for circular life support systems in long-term space missions

Regenerative Life Support Systems (LSS) fulfill the essential functions for human survival in space, such as atmosphere revitalization, water recovery, food production, and waste management, and are crucial for long-term space missions where the resupply of resources from Earth is not feasible or reliable. Operating a regenerative LSS poses several challenges, mainly related to its complexity, efficiency, and reliability. A set of heterogeneous subsystems involving mechanical, chemical, biological, and energetic processes has to be optimally coordinated in order to meet the requirements on mass, power, crew time, safety, reliability, sustainability and efficiency. In this paper, we address these challenges by proposing a supervisory control layer based on a nonlinear and time-varying Model Predictive Control (MPC) approach. The mathematical framework for deriving the prediction model addresses generic regenerative LSS. The MELiSSA (Micro-Ecological Life Support System Alternative) project developed by the European Space Agency is used here as the test case. For the first time, a complete dynamical model including all the MELiSSA compartments connected on all the phases (solid, liquid, gas) is derived, simulated, and controlled by a supervisory MPC. The design of such a controller follows a large set of requirements pre-defined by the MELiSSA project. Results on a mission lasting 14 weeks, which also includes a system failure scenario, are reported and evaluated for a specific MELiSSA network architecture.