Using Expert Judgement Methodology to Address Uncertainty in Launch Vehicle Weight Estimates

Abstract

The conceptual design phase of launch vehicle design is subject to significant uncertainty. This uncertainty needs to be quantified and reflected in multidisciplinary design and optimization. However, a primary problem in conducting risk analysis in conceptual launch vehicle design is the lack of historical data and the lack of a technique to quantify input parameter uncertainty. Under those circumstances, asking disciplinary experts for their best judgment is sometimes the only option. This paper describes an expert judgement methodology developed at NASA, Langley Research Center to quantify input distributions for executing launch vehicle weight-risk analysis. The methodology borrowed features from the fields of psychology, knowledge engineering and computer science in determining the uncertainty associated with weight estimating relationships for a launch vehicle design study. A set of questionnaires was developed to qualify and quantify uncertainty associated with design parameters as a set of probability distributions. This approach developed has proved to be a practical way of collecting and documenting expert judgement data. Once parameter uncertainty is quantified in terms of a probability distribution, risk can then be propagated throughout the multidisciplinary system design using simulation or analytical methods to determine overall design risk and for probabilistic multidisciplinary optimization.

Aeromechanics and Vehicle Configuration Demonstrations. Volume 3: A Hybrid Probabilistic Method for Estimate Design Margin

“…Because it is a common problem, this is an active area of research with several documented methods of eliciting subjective probability estimates. [52,86,88] An organization employing EMA would be able to choose a suitable method based on their experience and organizational needs.…”

“…This value would then be multiplied by the total multiplicative effect for the matrix which in this example is equal to unity. 88 Approved for public release; distribution unlimited.…”