A Bayesian hierarchical assessment of night shift working for offshore wind farms

Abstract

Abstract This article presents a Bayesian data‐modelling approach to assessing operational efficiency at offshore wind farms. Input data are provided by an operational database provided by a large offshore wind farm which employs an advanced data management system. We explore the combination of datasets making up the database, using them to train a Bayesian hierarchical model which predicts weekly lost production from corrective maintenance and time‐based availability. The approach is used to investigate the effect of technician work shift patterns, specifically addressing a strategy involving night shifts for corrective maintenance which was employed at the site throughout the winter. It was found that, for this particular site, there is an approximate annual increase in time‐based technical availability of 0.64%. We explore the effect of modelling assumptions on cost savings; specifically, we explore variations in failure rate, price of electricity, number of technicians working night shift, extra staff wages, months of the year employing 24/7 working and extra vessel provision. Results vary quite significantly among the scenarios investigated, exemplifying the need to consider the question on a farm‐by‐farm basis.

Keywords

• Bayesian modelling
• decision making
• failure modelling
• night shifts
• offshore wind
• OM