Print Email Facebook Twitter Optimization of Ice-Class Propellers Title Optimization of Ice-Class Propellers Author Huisman, T.J. Contributor Van Terwisga, T.J.C. (mentor) Faculty Mechanical, Maritime and Materials Engineering Department Marine & Transport Technology Programme Ship Hydromechanics / Resistance and Propulsion Date 2015-11-11 Abstract The main objective of this Master’s thesis is to develop an optimization routine to improve ice-class propeller design methodology using the design space within the ice-class rules. Ice impacts on a ship propeller give additional design demands to ensure reliability and safety. Consequently, ice class propellers feature thicker blades, therewith compromising fuel efficiency. However, ships trading the Baltic states and Scandinavia only sail two to five percent of their time in ice-infested waters. Propulsive efficiency should hence be optimized for ice-free conditions only, while still having sufficient ice performance and strength. The Finnish Swedish Ice Class Rules prescribe loads on the propeller blade as five load cases of uniform pressure that should be applied on the propeller blade. The Non-Dominated Sorting Genetic Algorithm II (NSGAII) is coupled to MARIN’s in-house propeller geometry generator, hydrodynamic boundary element analysis method PROCAL and a finite element analysis to evaluate the propeller blade strength. Both the radial and chordwise propeller distributions are parameterized by means of Bézier curves into optimization design variables. With these expansions, the computational framework is capable to automatically satisfy the ice-class stress constraints while converging to the best possible objective values. Each propeller within the optimization is iterated on mean pitch towards a design thrust. The four optimization objectives that are considered in this Master’s thesis are propeller efficiency, thrust variation throughout the ship’s wake field, propeller mass and ice-induced loading. Efficiency is considered as main objective while thrust variation is intended to provide interaction with the wake field. Besides the practical importance of the mass objective, it also guides the optimization towards high efficiency and maximum allowable material stresses. Based on a steady simulation of ice milling by means of an idealized ice-load pressure distribution, the ice-induced loading can be estimated as quantification of ice-performance. Best practice guidelines on the usage of PROCAL within the optimization are developed based on grid refinement and numerical uncertainty studies. Four different implementations of the finite element method are compared to the solution from a dense tetrahedral solid element mesh. Linear shell elements appear to perform best, both in terms of computational time and accuracy. A case study shows that ice-induced loading can be reduced as function of particularly the pitch distribution and blade profile geometry. It is also observed that the optimization searches for the weaknesses within the computational methods. For instance, it appears that the current ice-class rules allow highly skewed propellers, despite damage cases in practise. The optimization results are encouraging for future work concerning the optimization of blade profiles, although further work is required. It appears that the thrust variation objective steers towards flat chordwise pressure distributions. Cavitation computations are not yet included in the optimization, nonetheless, the optimized propellers show only little cavitation in the tip region. In conclusion, the optimization seems to provide a well-balanced starting point towards the design of high efficiency ice-class propellers. Subject icepropelleroptimizationgeneticclass To reference this document use: http://resolver.tudelft.nl/uuid:0607236e-c6d1-41a7-873b-c487065cea34 Part of collection Student theses Document type master thesis Rights (c) 2015 Huisman, T.J. Files PDF Huisman_-_2015_4_-_Optimi ... _Study.pdf 3.26 MB PDF Huisman_-_2015_10_-_Optim ... Thesis.pdf 3.84 MB Close viewer /islandora/object/uuid:0607236e-c6d1-41a7-873b-c487065cea34/datastream/OBJ1/view