Adjoint-Based Shape Optimization of Marine Energy Saving Devices Using Partially Moving Reference Frame Propeller Resolution Method
Please login to view abstract download link
The environmental footprint and operating cost of a ship are primarily governed by fuel consumption, which is strongly influenced by the hydrodynamic performance of the hull, the propeller, and their mutua interaction [1]. By modifying the hull wake upstream of the propeller, energy-saving devices (ESDs) play a key role in improving propulsive efficiency and overall hull–propulsor interaction. Consequently, simulation-based optimization requires methods that accurately resolve the complex flow features in the vicinity of both the propeller and the ESD, while remaining computationally efficient for resolving the flow around the entire ship. High-fidelity methods, in which the transient rotational motion is resolved by a corresponding rotation of the propeller grid, are afflicted by substantial computational costs due to the unsteady nature of the application. This particularly applies to adjoint-based optimization frameworks. To alleviate this burden, the Multiple Frames of Reference (MRF) method [2] and variations of it [3], have been proposed as a computationally efficient propeller-resolving alternatives in both primal and adjoint [4]. These methods typically trade computational efficiency for a reduced level of flow-field resolution in the propeller region. To this end, we present a modified Partially Moving Frame of Reference method for resolving rotational motion as well as its corresponding adjoint. We show that the proposed method achieves a better balance between omputational efficiency and high-fidelity flow resolution. The example included refers to the established Japan Bulk Carrier case and its propeller. An energysaving device is optimized to improve the flow field in the vicinity of the propeller and thereby enhancing the propulsive efficiency of the vessel using the proposed rotational-motion resolution method and its corresponding adjoint framework.
