Proceedings of the Eighth International Conference on Probabilistic Safety Assessment & Management (PSAM)
66 An Intelligent Agent-Oriented Approach to Risk Analysis of Complex Dynamic Systems with Applications in Planetary Missions (PSAM-0202)
Download citation file:
- Ris (Zotero)
- Reference Manager
NASA's future planetary missions will use electric propulsion system. These missions have typically different phases in which the operating time of the propulsion system changes as the spacecraft enters or escapes orbits of planets. These missions require operation of complex dynamic systems over a long period of time.
Reliability of propulsion systems is crucial for the success of planetary missions. Conventional fault tree, event tree and reliability block diagram prove to be cumbersome to capture dynamics and loss of redundancy of such systems. The other constraint is modeling the CCF (common cause failure) within and between independent dynamic assemblies and system blocks. Without a practical model for the system failure, modeling of CCF is not practical.
Such complexities restrict using available commercial fault tree or event tree analysis software tools in which popular reliability modeling methods have been applied.
In this work, an agent-oriented approach was adopted to make a computer-based direct simulation for the dynamic failure logic of the system. In this system simulation, every part of the system is replaced by an intelligent piece of software that represents properties and behaviors of its real counterpart of the system. This method offers a higher level of abstraction when the main logic of the failure needs to be implemented.
In this work CCFs are considered using the conditional probability values calculated based on global screening data available for each component. Again a Monte Carlo based sampling will combine the available knowledge on CCFs with the direct simulation of the system.
For a given trial, the computer program developed for this purpose, proceeds along the mission time and generates an operational scenario by checking the status of contributing agents. In the next step based on requirements (success criteria) of each phase, failure or success of the mission in the simulated operational scenario is evaluated. The computer program tracks these trials up to failure of a particular phase or success of the whole mission, and records the total number of failures in each phase. At the final step, the time dependent reliability of propulsion system including CCF of components is presented and related issues are highlighted.