Disclaimer: Any mention of commercial products within NIST web pages is for information only; it does not imply recommendation or endorsement by NIST.
The Object Management GroupⓇ (OMGⓇ) recently released an updated standard called SysML Extension for Physical Interaction and Signal Flow Simulation (SysPhS). The update specifies standards for translations between:
- Overall systems engineering modeling, done with Systems Modeling Language or SysML
- Simulations of physical interactions and signal flows within the system being engineered
Previously, system engineers used SysML and other modeling languages to create a model of an intended system and then described to engineers in other domains – electrical, mechanical, production, etc. – what they needed and how their products fit into the system. The other domain engineers used their own tools to specify system components and simulate their behavior. They then brought all the information together and incorporated it into a model of the overall system. However, differences between the two simulations often produced inconsistencies that were difficult for engineers to resolve.
NIST’s Conrad Bock (now in Smart Connected Systems, Communications Technology Laboratory) played a leading role in addressing this problem. Bock worked with stakeholders in the National Agency for Finite Element Methods and Standards to establish a need for a standard. He and other NIST researchers then conducted research underlying the standard, validated it through software, documented the standard, and shepherded it through the OMG standardization process. The resulting standard was published in May 2021. Since then, NIST researchers have interacted with industry to ensure that the standard can be implemented and used effectively, and they have identified additional needs to be addressed in future versions.
The update improves system engineering reliability. It shows how physical interactions and signals work together in a single system. It also includes a method for debugging physical interaction models, which are more difficult than signal flow models, due to bidirectional interactions between components.