Technical Report
Engineered Resilient Systems: Tradespace Tools Research
-
Systems Engineering and Systems Management Transformation
Report Number: SERC-2015-TR-102-5
Publication Date: 2015-06-01
Project:
Engineered Resilient Systems
Principal Investigators:
Dr. Tommer Ender
Co-Principal Investigators:
The Department of Defense’s (DoD) Science & Technology priority for Engineered Resilient Systems (ERS) calls for adaptable designs with diverse systems models that can easily be modified and re-used, the ability to iterate designs quickly and a clear linkage to mission needs. Towards this end, tradespace analysis is of great importance. The Georgia Tech Research Institute (GTRI) has been developing a web-based, collaborative modeling and simulation (M&S) tool that uses a model-based systems engineering approach to address the analysis of alternatives for the acquisition programs to assess cost, schedule and performance risk.
The following document describes the development of an integrated toolset to support the evaluation of earlystage design alternatives relevant to ERS. This includes research and development of methodologies to conduct Analysis of Alternatives (AoA) relevant to evaluating different dimensions of resiliency for these systems. This document also begins to describe the ways in which various toolsets, as part of the ERS Architecture effort, may interact to develop and build tradespaces for materiel solution analysis. This includes conducting research and development of methodologies to conduct AoA’s as relevant to evaluating different dimensions of resiliency for these systems. The primary toolset developed is ERS TRADESPACE, which provides a user interface for analysts to quickly and accurately assess and compare alternatives to execute materiel solutions analysis. ERS TRADESPACE consists of a core, containing the data structure, a layer containing the analytical process blocks, and a layer through which analyses are ordered, managed, and executed. OpenMDAO, and open source tool developed by NASA, is a key element of this latter layer, and it is here that a tradespace is generated and interaction with externally hosted models is controlled. These models are reachable via tools external to ERS TRADESPACE and exist as some synthesis of a server, a library of various analytical models and simulations, and an interface through which external software may submit inquiries. EASE and the Thrift interface used by ERDC are two such examples.
A tradespace may be developed and built through various means. Any one or more of the models in various use cases may reside locally and thus be accessed directly via ERS TRADESPACE. Additionally, any one or more of these models or simulations may reside externally, accessed through EASE or Thrift. In this latter scenario, ERS TRADESPACE uses specifically designed interfaces to EASE and Thrift to identify potential analytical models available and relevant to the given problem as well as specifying input and output parameters. Within ERS TRADESPACE, OpenMDAO then governs the interaction with these models by way of managing the input and output metadata, especially in terms or units and allowable ranges, and orders the execution of the various analytical blocks to generate a complete tradespace.
Resiliency in the context of a living, maturing framework does not preclude the development and maturation of stable analytical constructs and methods. As we operationalize ontological formalisms into analytical methods that capture and evaluate various resiliency dimensions, we begin by ensuring consistency with specifications of those dimensions. This work strives to lay the initial foundations for systems engineering the analysis of resiliency in the context of linking early-stage design to modeled operational characteristics and capabilities. The analytical methods are derived from existing ontological bases, and seek to promote consistency and comparability from one analysis to the next. Further, this work strives to develop modular and composable analytical constructs and processes that will also scale to tradespace dimensions common across ERS evaluations. “Flexibility of a Designed System to Future Engineering Change” contains its own analytical building blocks and processes as well to evaluate different dimensions of resiliency for Engineered Resilient Systems. Currently analytical methods focus on evaluating evaluates “Robustness of Fielded System Capabilities and Capacity with respect to Operational Requirements” and “Flexibility of a Designed System to Future Engineering Change” for complex engineered systems relevant to the DoD.
The approaches described here are currently being used in collaboration with government programs, and future publications will include results of application sharable with the broader community. Both of the analytical building blocks and their associated methods and workflows presented here may be readily applied to a raw, 8 freshly generated tradespace or one already filtered by some other means. The methods are robust to either scenario and are designed for intuitive application as stand-alone modules or through a rational synthesis with other methods.
Systems Engineering resiliency involves engineering the analytical constructs, methods, combinations thereof, and
the supporting processes, and tools together in a consistent, traceable, comparable, and executable manner. Each
development set must be consistent with the overall specification and what has come before. Further, we must
understand and develop methods whereby combinations of these metrics lead to more complete analysis of
resiliency dimensions for a system type. The development philosophy strives to enable design and development
of resiliency analyses that are transparent, intuitive, rational, and quantifiably traceable. In this way, we begin to
lay modular and composable foundations for resiliency analysis in the ERS context, and provide the software tools
to practically realize this vision. These foundations can form building blocks for future work and help promote
comparability and reuse of resiliency analyses across systems and designs in the future.