Time series semi-Markov decision process with variable costs for maintenance planning

Research output: Contribution to conferencePaper

Abstract

Deciding when and how to maintain offshore wind turbines is becoming even more complex as the size of wind farms increases, while accessibility is challenging compared to onshore wind farms. Planning future maintenance actions requires the wind farm operator to consider factors such as the current condition of the turbine, the cost of a given maintenance action, revenue generated by the asset, weather factors and vessel availability. Rather than making case-by-case decisions for each turbine, the approach described in this paper allows the wind farm operators to automate the process of short to-medium term maintenance planning through application of a Semi-Markov Decision Process (SMDP). The model proposed here is capable of suggesting the cost-optimal maintenance policy given weather forecast, future vessel costs and availability and the current condition of the turbine. Using the semi-Markov approach, allows the user to implement time varying failure rate. As the model is capable of utilising time-series data, future weather and vessel constraints can be applied depending on the information available to the user at the time, which will be reflected in the optimal policy suggested by the model. The model proposed here facilitates maintenance decision making in wind farms and will lead to cost reduction through more efficient planning. In addition to that, the model can be used to carry out a cost-benefit analysis of using vessels with different properties.

Conference

ConferenceEuropean Safety and Reliability Conference ESREL 2016
CountryUnited Kingdom
CityGlasgow
Period25/09/1629/09/16

Fingerprint

Time series
Planning
Turbines
Costs
Onshore wind farms
Availability
Offshore wind turbines
Cost benefit analysis
Cost reduction
Decision making

Keywords

  • offshore
  • wind
  • maintenance
  • planning
  • markov
  • decision
  • process
  • MDP

Cite this

Dawid, R., McMillan, D., & Revie, M. (2016). Time series semi-Markov decision process with variable costs for maintenance planning. 183. Paper presented at European Safety and Reliability Conference ESREL 2016, Glasgow, United Kingdom.
Dawid, R. ; McMillan, D. ; Revie, M. / Time series semi-Markov decision process with variable costs for maintenance planning. Paper presented at European Safety and Reliability Conference ESREL 2016, Glasgow, United Kingdom.
@conference{44fbf93208ce484ab4beef2f21f8ac50,
title = "Time series semi-Markov decision process with variable costs for maintenance planning",
abstract = "Deciding when and how to maintain offshore wind turbines is becoming even more complex as the size of wind farms increases, while accessibility is challenging compared to onshore wind farms. Planning future maintenance actions requires the wind farm operator to consider factors such as the current condition of the turbine, the cost of a given maintenance action, revenue generated by the asset, weather factors and vessel availability. Rather than making case-by-case decisions for each turbine, the approach described in this paper allows the wind farm operators to automate the process of short to-medium term maintenance planning through application of a Semi-Markov Decision Process (SMDP). The model proposed here is capable of suggesting the cost-optimal maintenance policy given weather forecast, future vessel costs and availability and the current condition of the turbine. Using the semi-Markov approach, allows the user to implement time varying failure rate. As the model is capable of utilising time-series data, future weather and vessel constraints can be applied depending on the information available to the user at the time, which will be reflected in the optimal policy suggested by the model. The model proposed here facilitates maintenance decision making in wind farms and will lead to cost reduction through more efficient planning. In addition to that, the model can be used to carry out a cost-benefit analysis of using vessels with different properties.",
keywords = "offshore, wind, maintenance, planning, markov, decision, process, MDP",
author = "R. Dawid and D. McMillan and M. Revie",
year = "2016",
month = "9",
day = "25",
language = "English",
pages = "183",
note = "European Safety and Reliability Conference ESREL 2016 ; Conference date: 25-09-2016 Through 29-09-2016",

}

Dawid, R, McMillan, D & Revie, M 2016, 'Time series semi-Markov decision process with variable costs for maintenance planning' Paper presented at European Safety and Reliability Conference ESREL 2016, Glasgow, United Kingdom, 25/09/16 - 29/09/16, pp. 183.

Time series semi-Markov decision process with variable costs for maintenance planning. / Dawid, R.; McMillan, D.; Revie, M.

2016. 183 Paper presented at European Safety and Reliability Conference ESREL 2016, Glasgow, United Kingdom.

Research output: Contribution to conferencePaper

TY - CONF

T1 - Time series semi-Markov decision process with variable costs for maintenance planning

AU - Dawid, R.

AU - McMillan, D.

AU - Revie, M.

PY - 2016/9/25

Y1 - 2016/9/25

N2 - Deciding when and how to maintain offshore wind turbines is becoming even more complex as the size of wind farms increases, while accessibility is challenging compared to onshore wind farms. Planning future maintenance actions requires the wind farm operator to consider factors such as the current condition of the turbine, the cost of a given maintenance action, revenue generated by the asset, weather factors and vessel availability. Rather than making case-by-case decisions for each turbine, the approach described in this paper allows the wind farm operators to automate the process of short to-medium term maintenance planning through application of a Semi-Markov Decision Process (SMDP). The model proposed here is capable of suggesting the cost-optimal maintenance policy given weather forecast, future vessel costs and availability and the current condition of the turbine. Using the semi-Markov approach, allows the user to implement time varying failure rate. As the model is capable of utilising time-series data, future weather and vessel constraints can be applied depending on the information available to the user at the time, which will be reflected in the optimal policy suggested by the model. The model proposed here facilitates maintenance decision making in wind farms and will lead to cost reduction through more efficient planning. In addition to that, the model can be used to carry out a cost-benefit analysis of using vessels with different properties.

AB - Deciding when and how to maintain offshore wind turbines is becoming even more complex as the size of wind farms increases, while accessibility is challenging compared to onshore wind farms. Planning future maintenance actions requires the wind farm operator to consider factors such as the current condition of the turbine, the cost of a given maintenance action, revenue generated by the asset, weather factors and vessel availability. Rather than making case-by-case decisions for each turbine, the approach described in this paper allows the wind farm operators to automate the process of short to-medium term maintenance planning through application of a Semi-Markov Decision Process (SMDP). The model proposed here is capable of suggesting the cost-optimal maintenance policy given weather forecast, future vessel costs and availability and the current condition of the turbine. Using the semi-Markov approach, allows the user to implement time varying failure rate. As the model is capable of utilising time-series data, future weather and vessel constraints can be applied depending on the information available to the user at the time, which will be reflected in the optimal policy suggested by the model. The model proposed here facilitates maintenance decision making in wind farms and will lead to cost reduction through more efficient planning. In addition to that, the model can be used to carry out a cost-benefit analysis of using vessels with different properties.

KW - offshore

KW - wind

KW - maintenance

KW - planning

KW - markov

KW - decision

KW - process

KW - MDP

UR - http://esrel2016.org/

M3 - Paper

SP - 183

ER -

Dawid R, McMillan D, Revie M. Time series semi-Markov decision process with variable costs for maintenance planning. 2016. Paper presented at European Safety and Reliability Conference ESREL 2016, Glasgow, United Kingdom.