Integrated System Health Management for Environmental Control and Life Support System in Manned-Spacecraft

Abstract

This paper addresses the integrated system health management (ISHM) diagnostics and prognostics for a manned spacecraft’s environmental control and life-support system (ECLSS), which ensures the safety of astronauts and guarantees space mission success. For the complex system structure, an integrated diagnostics and prognostics method is presented, which allows for the consideration of continuously monitored signals. In this method, the condition monitoring data are first classified by exploiting the comprehensive evaluation technique, and then the feature data are used to train the corresponding diagnosis models, which themselves represent the different stages of system degradation. Due to the variant behavior of the ECLSS in the space environment, variational approximation-based learning is designed in the diagnostics procedure to estimate the parameter distribution of the trained models rather than the parameters themselves. By exploiting the constructed trained models, the current ECLSS health stage and remaining useful life (RUL) can be identified. A numerical stimulation is provided to demonstrate the performance of the proposed integrated algorithm.