A novel frequency domain identification (FDI) strategy for the identification of radiation force models from frequency domain hydrodynamic data is proposed. First, a subspace identification method is augmented with a convex constraint that guarantees a stable solution. Then, in a second convex optimization problem, constraints on low- and high frequency asymptotic behavior and passivity are enforced. This novel method, constrained frequency domain subspace identification (CFDSI), is validated by comparing both SISO and MIMO CFDSI results with the state-of-the-art FDI toolbox, which is part of the Marine Systems Simulator MATLAB toolbox. In two test cases, it is shown that the novel algorithm can successfully identify a model with either a SISO or MIMO structure, where stability, passivity and the desired low- and high-frequency asymptotic behavior are guaranteed. For the two test cases presented, the quality of the CFDSI models matches the quality of the state-of-the-art FDI models.

This content is only available via PDF.
You do not currently have access to this content.