Expected Productivity-Based Risk Analysis in Conceptual Design
- Paper number
IAC-05-D1.3.01
- Author
Ms. Julie Wertz, Massachussets Institute of Technology (MIT), United States
- Coauthor
Dr. David Miller, Massachussets Institute of Technology (MIT), United States
- Year
2005
- Abstract
During the design process, risk is mentioned often, but very infrequently actually impacts major design or architecture decisions. One of the main reasons for this is that risk is considered a qualitative parameter. To be more effectively used as a parameter to search a trade-space and make design decisions, it is important to bring risk into the design process as a quantitative parameter that engineers can both understand and trade. One parameter that is almost always used when making design decisions is the productivity of a design or architecture. If instead of using simple productivity it were possible to examine the expected productivity, taking into account failures and risk items, the concept of risk could be brought into the design process in a very quantitative and tradable way.
Expected productivity modeling has been used sparingly in the aerospace industry in the past. However, in previous expected productivity modeling work, the productivity was not time nor path dependent. While assuming the productivity is path-independent greatly simplifies calculations, it is an invalid assumption for many real life systems. The productivity in each state often depends on what action the system is executing at that time. This, in turn, depends on the amount of time the previous actions took, making the productivity path-dependant. An approach has been developed to model the systems described above in a more efficient and effort saving manner than the current method of a Monte Carlo simulation. While the approach developed includes many mathematical details, the basic principle behind the approach is to find the expected path and then find the expected productivity given the expected path. This approach has been tested against Monte Carlo simulations in terms of both the accuracy of the calculations and the computational speed of completing the simulations, with excellent results in both categories. The new simulation approach can achieve results that consistently fall within the 95% confidence interval of the Monte Carlo results, while completing the simulation anywhere from 10 to almost 275 times faster than the equivalent Monte Carlo. Additionally, the approach of using expected productivity to represent risk in a conceptual design trade has been demonstrated using the JPL Terrestrial Planet Finder Interferometer (TPF-I) mission. Due to the work described in this paper, early design decisions for TPF-I have been effected by risk modeling and degraded state analysis. This paper will discuss the need for the new modeling methodology, the details of the methodology that has been developed, comparisons between this new methodology and a Monte Carlo simulation, and finally an example of how to use expected productivity in a trade study for a real mission, TPF-I.
- Abstract document
- Manuscript document
IAC-05-D1.3.01.pdf (🔒 authorized access only).
To get the manuscript, please contact IAF Secretariat.