Publication Detail

ESRDC Annual Report FY08

Chryssostomos Chryssostomidis, Franz Hover, George Karniadakis
2008
17 pp.

Executive Summary
Probabilistic analysis and simulation of electric networks in transient states has not been very advanced due primarily to the lack of stochastic tools for time-dependent systems. The work at MIT has focused on developing such tools in the context of the All-Electric-Ship (AES). We have formulated efficient new methods in modeling uncertainty for electro-mechanical systems based on different extensions of polynomial chaos, and have applied them to various models extracted from the ONR Challenge problem. We have also developed new physical models to couple the ship hydrodynamics - and, hence, the randomness of sea states - to the integrated power system (IPS), thereby coupling the electrical, mechanical and hydrodynamic domains.

In particular, in this report we review progress on several tasks: In the first one, results from stochastic simulations of the ONR integrated power system (IPS) testbed have been analyzed for sensitivities, and similar tests have been run on the real time digital simulator at the Center for Advanced Power Systems at Florida State University. We also explored applications of grid-based stochastic simulation techniques to nonlinear filtering. Further work on automating ship maneuvering has been performed, focusing on a hybrid approach that uses high-authority, open-loop maneuvers connected with simple “level” trajectories and feedback elements to control errors. The planning and control results have been tested extensively with scale model experiments.

In other tasks, we employ a sensitivity method we developed for the AES to identify the most important parameters of the IPS under certain operating conditions. This information is then used to construct low-dimensional models for the IPS, which are very valuable for the efficient design of IPS. Specifically, we show that our low-dimensional models agree with the full (expensive) models even in the presence of 50% perturbations in the stochastic inputs. Finally, in the last task we develop new models that couple the ship motion and the propeller to the IPS, hence creating a new framework of analysis of IPS, the ship, and the sea states, for first time. Both simplified models but also full 3D solvers are involved in our analysis in order to obtain both accurate and efficient methodologies for the electric ship designer of future AES.

The above results form a comprehensive framework based on which we plan to pursue future research activities of integrated power systems in mixed domains and design of AES under uncertainty.

The impact of our work has been both on fundamental advances in stochastic modeling as well as on establishing a new modeling approach to AES. Many research groups within the ESRDC consortium and beyond have adopted the generalized polynomial chaos methods that we have developed.

type: Technical reports

This publication is not currently available from MIT Sea Grant. Please try again later or contact MIT Sea Grant for more information.

Parent Project

Project No.: 2008-ESRDC-01-LEV
Title: Electric Ship Research and Development Consortium (ESRDC)

Other publications by this author (as Lead)