NASA Logo, National Aeronautics and Space Administration

Resilience Analysis and Design

The Resilience Analysis and Design (RAD) project pursues the development of computational analysis tools and design methods to enable the consideration of resilience in the design process. Resilience is the property of a system to actively and inherently prevent, mitigate, and recover from hazardous scenarios as they unfold over time. A system can be resilient because of inherent flexibility placed in the design and operational profile, or because of active contingency management measures. Considering the resilience of a system in the design process thus requires analyzing how these different approaches affect its dynamic response to hazards over time.

Towards that goal, the RAD project:

1. Develops the fmdtools simulation toolkit, a Python toolkit for simulating dynamic hazardous scenarios over time

2. Develops design processes for considering resilience in design, including optimization methods for automatically exploring the space of possible solutions to a given resilience optimization problem.

3. Develops resilient design case studies which show how these methods can play out in practice.

fmdtools

Fmdtools is a Python toolkit for simulating the dynamic effects of hazardous scenarios in complex engineered systems and analyzing the associated resilience of the system. To support early design work, fmdtools uses a function-based representation which can evolve as the concept increases in fidelity, enabling network-based, static, and dynamic behavioral representations of the system. This function-based representation additionally enables conduction early functional hazard assessment (FHA) processes in simulation, thus enabling an iterative design process that can explore and systematically compare the resilience of a wide range of designs. There are three main components of the fmdtools package:

1. Model definition constructs which enable systematic early specification of the high level structure and behaviors of a system with concise syntax.

2. Simulation methods which enable the quantification of system performance and propagation of hazards over a wide range of operational scenarios over a wide range of model types.

3. Analysis methods for quantifying resilience and summarizing and visualizing behaviors and properties of interest.

The main impetus for the development of the fmdtools project was a lack existing tools to enable early function-based fault simulation for early functional hazard assessment. Researchers thus had to re-implement modelling, simulation, and analysis approaches for each new case study or methodological improvement. The fmdtools resolves this problem by separating resilience modelling, simulation, and analysis constructs from the model under study, enabling reuse of methodology between case studies.

The RAD project continues to develop this package to:

  • Expand the set of possible use-cases from mechanical systems to complex Systems of Systems (SoS) with human/software/component interactions,
  • Advance the simulation approach by enabling more complex behaviors and scenarios to be run, including probabilistic and statistical methods for simulating and analyzing a given system’s performance and resilience, and
  • Encourage the usage of this tool through tool integrations, interfaces, computational performance improvements, and tool documentation which make it easier to learn and more practical to use.

  • Resilience Optimization

    Incorporating resilience in the design process can be difficult because the design space is large and heterogenous (involving many different variables and considerations) and because the evaluation of solutions generally requires running complex simulations. Resilience Optimization is the systematic process of exploring this design space to improve the resilience of the design using mathematical optimization techniques.

    The RAD project develops frameworks for optimizing resilience when the problem has complexities with methods to interacting design, operational, and resilience objectives, constraints, and variables and optimizing a system under partial information through a design process.

    Case Studies/Applications

    SMARt-STEReO: RAD is developing a model of wildfire emergency response operations that showcases advanced System of Systems (SoS) resilience simulation and analysis capabilities.

    Rover: RAD is developing a model of an autonomous lane-following rover to better understand AI faults and human interactions.

    Publications and Presentations

    Daniel Hulse, Hongyang Zhang, Christopher Hoyle, "Understanding Resilience Optimization Architectures With an Optimization Problem Repository ," International Design Engineering Technical Conferences and Computers and Information in Engineering Conference IDETC/CIE-2021; Virtual Conference; 17 August 2021
    [presentation of conference paper (PDF, 2.2MB)]

    Sequoia Andrade, Eleni Spirakis, Daniel Hulse, Hannah S. Walsh, Misty D. Davies, " The System Modeling and Analysis of Resiliency in STEReO (SMARt-STEReO)," AIAA Aviation 2021; ZOOM/Virtual; 06 August 2021
    [presentation (PDF, 594KB)]

    Daniel Hulse, Sequoia R. Andrade, Eleni Spirakis, Hannah S. Walsh, Misty D Davies"SMARt-STEReO: Preliminary Model Description," NASA Technical Memorandum; 30 September 2020
    [technical memorandum (PDF, 2.1MB)]

    Updated 1 December 2021


    Active Members

    Daniel Hulse
    Sequoia Andrade
    Lukman Irshad
    Hannah Walsh

    University Collaborators

    Christopher Hoyle, Oregon State University

    Interns

    Inga Girshfeld


    First Gov logo
    NASA Logo - nasa.gov