Using Estimation Techniques in Multidisciplinary Design

Abstract

Viewing the multidisciplinary design problem as a dynamical system a number of tools from the established field of dynamical system theory became available to the multidisciplinary design community. This work demonstrates the applicability of applying the Kalman filter in a manner similar to linear covariance analysis to the multidisciplinary design problem to obtain robustness characteristics. In addition to robustness characteristics, the estimation theory is shown to be applicable to design decomposition. Following theoretical development, two example problems demonstrate the applicability of applying dynamical system theory. For a linear, two contributing analysis problem showed the mean was able to be estimated with an error less than 0.08% and a matrix norm bounded the variance to less than 37.8% relative to analytic propagation. This error is shown to be a function of the geometry of the matrix two-norm and reduces as the problem dimensionality increases. The use of estimation theory is also shown to be applicable for nonlinear designs through a two-bar truss problem through successive linearization.