Mathematical Programming Models and Algorithms for Oshore Wind Park Design

Research output: Book/ReportPh.D. thesis – Annual report year: 2018Research

Documents

View graph of relations

Designing an offshore wind park is a complex process, involving several different
expertises, and multiple tasks. In this thesis we developed Mathematical
Programming models and algorithms to help the wind park designers. In particular,
we focused on two optimization problems arising at the design phase of
offshore wind parks, namely the optimal allocation of turbines in a given site
and the connection of turbines through cables. We briefly touched upon the
optimization of offshore jacket foundations as well.
This thesis was motivated and supervised by Vattenfall, a leading company
in wind park development and operation. Thanks to our close collaboration,
the optimization problems have been described and modelled as they arise in
practical applications and they have been tested on real data. Our work proved
to have a huge impact in practice, being able to increase park production and
reduce costs. Having a sound optimization tool to help the designers allows
also for different what-if analyses and scenario evaluations. This is of key value
for Vattenfall, especially when looking at new technologies on the market.
The mathematical optimization models and algorithms developed have been
considered of great interest also by the Operational Research (OR) community,
and resulted in six journal papers. This thesis wants to follow the two-fold
nature of our project, offering interesting material both to wind energy experts
and practitioners, and to OR experts. Therefore we alternate OR journal
papers, with practical examples and impact evaluations.
Finally, we proposed an application of integrating Machine Learning and OR,
where we investigate if a machine, trained on a large number of optimized
solutions, can estimate the value of the optimized solution for new instances.
This research question is of interest for all kinds of optimization problems, and
is here studied on our specific wind farm application.
Original languageEnglish
PublisherDTU Management Engineering
Number of pages310
Publication statusPublished - 2017

Bibliographical note

Ph.D. afhandling

Download statistics

No data available

ID: 158908587