A Design Tool for Machinery Space Arrangement

A Design Tool for Machinery Space Arrangement

van der Bles, Roel (TU Delft Mechanical, Maritime and Materials Engineering; TU Delft Ship Design, Production and Operations)

Kana, Austin (mentor)
Hopman, Hans (mentor)
Mayenburg, Nicky (mentor)
Atasoy, Bilge (graduation committee)

Delft University of Technology

Marine Technology

2019-04-10

The subject of this master thesis is the development of a design tool to help Vuyk Engineering Rotterdam B.V. arrange the machinery spaces (engine room, switchboard rooms, separator rooms) aboard their vessels early in their design process. It applies the field of automatic design generation to a step in Vuyk’s design process to save time and easily explore alternatives. First the general design process at Vuyk and the design of the engine room and other machinery spaces in particular are investigated. This is followed by a review of literature on design of machinery spaces, both in maritime applications in ships and in non-maritime process plant lay-out design, to compare theory and practice at Vuyk and look for new insights. Next to the design process itself, the characteristics of ship concept design tools are investigated in literature, leading to a focus on concept exploration and various methods how to deal multiple objectives. This results in the design requirements for the design tool for machinery space arrangements and a focus on the most relevant requirements to consider in the limited time frame of this thesis, to limit the scope. The best modelling method and resolution approach for the problem are investigated. The Facility Layout Problem and Particle Swarm Optimization algorithm are identified as the best options. The model is defined accordingly. The capabilities of the model for the design tool are tested in a concept design of a jackup vessel. Initial runs of the model led to problems for the Particle Swarm Optimization algorithm, as its gets stuck in local minima while trying to solve the problem. This severely limits the solutions diversity. Several adaptations for the model are developed to overcome this problem. The particle swarm algorithm is adapted with several mutations, allowing it to better escape the local minima thus creating more diverse, better optimized solutions. The model is also improved with an additional grouping constrained and a pathing filter to better account for the identified machinery space design drivers. The resulting tool is applied in a case study on the early concept design phase for a dredge pump room and connected engine room for a trailing suction hopper dredger. The tool is used to evaluate early design choices in space arrangements for a diesel-direct drive of the dredging pump and jet pump, in comparison to alternative diesel-electric solutions with three V16 diesel-generator-sets, four V12 sets or six V8 generator sets. The application of the tool leads to valuable insights for the arrangement options for the machinery spaces. The developed tool is shown to be a useful addition to the early concept design of machiney space arrangement.

Early Stage Ship Design
Engine room
Particle Swarm Optimization
Facility Layout Problem
Machinery Arrangement

http://resolver.tudelft.nl/uuid:6e7a6d44-d293-44f8-a4e6-113423ceec59

Student theses

master thesis

© 2019 Roel van der Bles