Graduation assignments Production and Logistics

MSc assignment: decision support for repair and maintenance of offshore wind farms

Introduction

Existing energy production capacity from offshore wind farms amount 1 GW which is sufficient to power 1,2 million households in the Netherlands. Further plans are to add additional 13.5 GW until the year 2030.  Besides the installation costs, maintenance is a main driver of the expected cost of offshore wind farms, for which unexpected damage leads to very expensive and complicated logistics for the wind farm operators and maintainers.  Thus, to reduce the cost of wind energy, reducing uncertainty about maintenance decisions is key.

Status

To support with this challenge, Independent eXperts (IX) has developed software for Strategic Operation and Maintenance Optimisation Simulations (SOMOS), which aids decision making on strategic level (planning horizon 20-30 years). It focusses particularly on maintenance, logistics and infrastructure which need to be decided at the beginning of the wind farm operation.

Goal

IX in cooperation with the University Twente and Joulz (www.joulz.nl), is currently developing a follow up of SOMOS, called WiMOS. WiMOS is a tactical prognostics and decision support system, which shall provide more detail and accurate assessment on the condition and risks of the systems of offshore wind power plants. Furthermore, it shall actively support in making the optimal decisions and planning for the upcoming months. This requires a more detailed planning than SOMOS, using additional information on e.g. condition information of critical components for the operation of the wind farms, weather conditions allowing maintenance (or not), availability of resources (ships, maintenance engineers, tools, spare parts), etc.

The proposed assignment aims at complementing the planning logic within the existing software for offshore wind farms maintenance simulation (SOMOS) into a decision support system for the short to medium term (up to six months).

Scope and tasks

The optimum maintenance scenarios are selected after the evaluation of changing variables such as opportunity cost of intervention, and the risk (probability of occurrence vs impact) of the failure of critical components of the wind farm.

The optimization is run following two approaches. First, enhancing scheduled maintenance to reduced chances of unexpected maintenance outside planned maintenance interventions. This may cover a variety of preventive maintenance concepts, like time-based, use-based, load-based and/or condition-based maintenance. because of the high costs for initiating maintenance (a ship with all required resources should travel to the wind farm), clustering of maintenance activities and opportunistic maintenance tasks are also important issues. Second, by minimizing effects of unexpected (corrective) maintenance by ensuring the required logistics to be available at the time they are needed while minimizing costs. Again, corrective maintenance may be combined with opportunistic maintenance where useful.

The tasks of the selected candidate are:

  1. Complete a gap analysis, and prepare a plan of approach in cooperation with the internship supervisor
  2. Prepare a model for the six-month wind farm operational analysis
  3. Develop optimization algorithm for delivering the lowest cost of energy.
  4. Model validation and testing on one or more cases to be defined with the problem owner
  5. Documenting and reporting

Execution

This assignment is intended as a master thesis for a student of Industrial Engineering and Business Information Systems. The existing code make use of probability for estimation of failures and costs, and Monte Carlo simulation for the evaluation of life scenarios. The code is written in Visual Basic, although in the future a different language is foreseen.

Application

To apply and/or for more information please contact:

Andrea Sanchez  a.sanchez@ixwind.com or Eric Kamphues e.kampuhes@ixwind.com

For more information about IX, please visit our website: ixwind.com