Review of Markov Models for Maintenance Optimization in the Context of Offshore Wind

Abstract

The offshore environment poses a number of challenges to wind farm operators. Harsher climatic conditions typically result in lower reliability while challenges in accessibility make maintenance difficult. One of the ways to improve availability is to optimize the Operation and Maintenance (O&M) actions such as scheduled, corrective and proactive maintenance. Many authors have attempted to model or optimize O&M through the use of Markov models. Two examples of Markov models, Hidden Markov Models (HMMs) and Partially Observable Markov Decision Processes (POMDPs) are investigated in this paper. In general, Markov models are a powerful statistical tool, which has been successfully applied for component diagnostics, prognostics and maintenance optimization across a range of industries. This paper discusses the suitability of these models to the offshore wind industry. Existing models which have been created for the wind industry are critically reviewed and discussed. As there is little evidence of widespread application of these models, this paper aims to highlight the key factors required for successful application of Markov models to practical problems. From this, the paper identifies the necessary theoretical and practical gaps that must be resolved in order to gain broad acceptance of Markov models to support O&M decision making in the offshore wind industry.

A Review of Predictive and Prescriptive Offshore Wind Farm Operation and Maintenance

“…Maintenance decision support models often use Markov decision processes [47], which extend Markov models with the possibility of taking actions and receiving rewards. This enables the model to simulate a scenario where outcomes are partly random and partly controlled by a decision maker [48].…”