urbazewski's Journal: resilience of life support systems

. . .

I love deadlines ---
I love the whooshing sound they make as they fly by.
Douglas Adams

I'm working on a paper that uses a hybrid (discrete and continuous) model of a generic water revitalization system to develop and test new measures of system resilience to component level faults. One key idea is that resilience is a dynamic property of the system that cannot be understood in terms of the fault probalities or MTBF (mean time before failure) of individual system components. I've been working on these ideas on and off for a couple of years, starting with a small award from the Director's Discretionary Fund when I was working as a contractor at the NASA Ames Research Center.

One novel method of measuring resilience that we developed estimated the transition probabilities of a Markov chain on the two states "nominal -- demand for clean water met" and "performance failure -- demand for clean water not met" and used the ratio of those parameters as a measure of resilience. Paper available.

The current paper will take a broader view of competing ways to measuring resilience. It's titled: "Alternative Metrics for Evaluating the Resilience of Advanced Life Support Systems" Ann Maria Bell, Orbital Sciences Corporation; Richard Dearden, Research Institute for Advanced Computer Science; Julie A. Levri, NASA Ames Research Center. Abstract:

Ensuring the safety of the crew is a key performance requirement of a life support system. However, a number of conceptual and practical difficulties arise when devising metrics to concretely measure the ability of a life support system to maintain critical functions in the presence of anticipated and unanticipated faults. Resilience is a dynamic property of a life support system that depends on the complex interactions between faults, controls and system hardware. We review some of the approaches to understanding the robustness or resilience of complex systems being developed in diverse fields such as ecology, software engineering and cell biology and discuss their applicability to regenerative life support systems. We also consider how approaches to measuring resilience vary depending on system design choices such as the definition and choice of the nominal operating regime. Finally, we explore data collection and implementation issues such as the key differences between the instantaneous or conditional and average or overall measures of resilience. Extensive simulation of a hybrid computational model of a water revitalization subsystem (WRS) with probabilistic, component-level faults provides data about off-nominal behavior of the system. The data are used to consider alternative measures of resilience as predictors of the system?s ability to recover from component-level faults.

So far so good, now all I need is a paper to go with the abstract.

The deadline for submitting the first draft of the paper was March 7.