Parametric Design and Analysis to Support Model-Based Systems Engineering Using SysML
- Created: Friday, 01 October 2010
What is commonly called Model-Based Systems Engineering (MBSE) proposes to replace traditional document-centric approaches to systems engineering with model-centric approaches in which the complete definition of the system or product is captured in a single repository and model. By “complete definition,” this repository and model needs to capture information spanning the so-called “four pillars” of systems engineering: requirements (what the system must do), architecture (how the system is structured), behavior (how the system operates or is used), and parametrics (formal rules and constraints describing the system). The MBSE paradigm, it is argued, offers
- Improved quality
- Enhanced system design integrity
- Improved specification of allocated requirements to hardware/software
- Fewer errors during integration and test - More rigorous requirements traceability
- Increased productivity
- Improved impact analysis of requirements changes
- Reuse of existing models to support design/technology evolution
- Auto-generation of documentation
- Reduced risk
- Improved cost estimates
- Early/ongoing requirements validation and design verification
There are a number of software tools available to systems engineers choosing to practice MBSE. Of particular interest are those supporting the SysML graphical programming language, an open-source standard maintained by the Object Management Group.
NASA’s program and project organizations are almost always a complex mix of NASA centers, other agencies (US and foreign), contractors, and universities. Use of standards-based tools, languages, and data formats can minimize, if not eliminate, interoperability issues that often plague collaborative engineering efforts in which information from various sources must be integrated and exchanged.
While adopting the MBSE approach and use of SysML is an important first step in the right direction, what is truly required is that we then create links from the available MBSE tools to NASA’s arsenal of discipline/domain engineering tools used for requirements management, cost modeling, design (MCAD, ECAD), and simulation (CAE, FEM, CFD). Solutions that take advantage of open standards such as AP233 (Systems Engineering), AP203/ AP214 (Mechanical Engineering), AP209 (Structural Engineering), AP210
(Electrical Engineering), and STEP-TAS (Thermal Engineering), are once again clearly preferred.
Systems engineering activities such as trades, optimization, verification, and validation can be conducted within an integrated tools framework that will support more effective decision making. Systems engineers would be able to tap into the computational power of the discipline/ domain engineering tools to go far beyond simple spreadsheet-based models typically employed today. And discipline/domain engineers would have a single authoritative source, the MBSE data repository, for the design information needed to create their models and simulations, and conduct their own analyses in support of subsystem development.
NASA has only begun to adopt MBSE methodologies and tools to support its programs and projects. We have funded efforts through the Small Business Innovation Research (SBIR) program to create interfaces between one or two MBSE tools and a small subset of our discipline/domain tools, but much work remains to achieve the comprehensive goals identified herein.